JBL MS-8 FAQ

Tendean17 · Feb 20, 2011

DIYMA.. This is my first Thread.

Bikinpunk says "If you love car audio tuning but really hate or just do not know how, then this processor is for you. The MS-8 does in 5 minutes What would take me a couple months of hard work to do. Then throw in the added benefits of great sound at different seats and I've just added many more months onto my manual tune"

I love car audio and I'm not the lazy guy. I've tried hard but still could not make me happy in my car. Maybe i just do not know how to tune it well and there is no comparison that i can do here. There is hardly any installer who has competence in the small town where i live.

I find it difficult to find the posted articles of this processor because they are included in several threads. Require several weeks to collect, read and know about the processor further. Select from the thousands of posts and to make me easy then I just copy and paste the article into almost 100 pages of Word Document. Then i just print the document and i can read it in my spare time .. then I thought "Why not i just share my documentation ?" That’s why i make this thread.

I documented the complete enough and i think is useful also for other friends in this forum. After all the material collected so i separate them into categories that make me easy to read. This classification is not 100% accurate because sometimes what the experts say covering into several categories.

In case someone interest can PM me your email address and i will send to you. If you feel useless .. easy .. just remove from your computer. I live 180 degrees Out of Phase with almost all of you my friends .. so i need maybe 1 or 2 days to send it to you. I do not have the JBL MS-8 but i have planned about it.

Hopefully useful .. Sorry for my english and this is not my review.

Tendean17 · Feb 20, 2011

JBL MS-8 - Categories :

• About JBL MS-8
• Prepare your system design first
• Target Curve for JBL MS-8
• Phase Frequency Response
• Center and Rear Channel
• Aux Input, Signal Level and Preout Voltage
• Measurement and Setup
• System Noise
• Happy Customer
• Feature Request

Here's the Links ...

JBL MS-8 - FAQ ( This Thread )

Tendean17 · Feb 20, 2011

About JBL MS-8

300Z :

MS-8 comes with the pre-amp/amp, a small display, a wireless remote control, a binaural microphone and a setup CD.

MS-8 will have 8 speaker level inputs, 8 line level inputs and an iPod input. The inputs are summed to provide a full-range 2-channel signal. If 6 or fewer inputs are required, then the last 2 can be used for an additional aux input.

There are no digital inputs. Why? Because the real benefit of digital input is "no noise". The downside to providing one is that for 99 percent of users, it's more hassle than it's worth and will cause a great deal of confusion. Not too many people understand that the connector doesn't determine the signal. What I mean is, if we put a toslink input and a user hooks up a toslink output, it will only work is the signal is compatible. DVD-Audio isn't available on a digital output, DVD signals are 48k, PCM is 44.1, home-made digital audio can be anything. The signal from tuners is often only output on the RCAs. For the vast majority of users, there is no benefit and too much opportunity for disappointment. As far as noise goes, our inputs are differential, so the common mode noise rejection is super high. There won't be any noise.

There are 8 input channels, so the 8 speaker level inputs and 8 line level inputs are basically in parallel. You can use any combination.

Once the signals have been combined and un-EQed (for a flat 2-channel signal), the signal is processed with Logic7. That provides signal steering for a center channel (if you have one--if not, no problem) and processing for side and rear channels. L7 works on any 2-channel source and is our version is written for cars rather than live-in rooms, so it sounds MUCH better than any of the encoded formats in a car. The 2-channel downmix of any encoded DVD or DVD-A disc will play back in full surround. If good-old 2-channel is what you want, L7 is defeatable and the channels are fully configurable (there are 8 output channels and they can be pre-amp channels or powered channels--20W x 8 at 4 ohms, 30W at 2 ohms). You can have 3-way front, a center and a sub, 2-way front, rear and a sub...whatever you want to do.

The electronic crossover that's built in is fully configurable. You can assign any channel to be anything and it includes an EZ setup mode and an advanced mode. In EZ setup, you tell each channel the speaker location (front right, for example), then you tell it what speaker is connected (6" full-range). It sets the crossover point. In advanced mode, you tell the channel the location (right front) and then assign a filter type (HP, LP, BP) and then you set the filter frequency (you can assign any value between 20 at 20kHz) and the slope (1st-4th order).

After the crossover setup is completed, you move on to the EQ. You put on the microphones (they look like airline headphones but contain mics instead of speakers) and insert the CD. The display will give you some instructions to sit in the driver's seat and look at the left mirror and press "go". the unit will make a quick sweep of all 8 output channels. Then it will ask you to look forward and will make another sweep. Finally, it'll ask you to look to the right--another quick sweep. You can measure only the driver's seat or up to 4 seats. After the measurements are made (takes about 5 minutes) the unit will calculate the frequency response, level and arrival time for all 8 channels in each seat and crunch some numbers (another 30 seconds or so). It auto-tunes the car with 48 measurements per seat (up to 4 seats). It will output a tuning optimized for the driver, passenger, compromise between driver and passenger and one for the rear seats. If you use a center channel, both front seats will sound the same and the image will be great for rear seat passengers too.

After the auto-tuning is done, it will allow you to change the target curve. You can call up a 31-band EQ tool and make whatever changes you want. Unlike a regular EQ, you don't have to find an RTA and tune the car with the EQ, you just draw the curve you want to hear and press "go" and it does the work in implementing your curve. Then you can switch back and forth between your curve and the automatic one and continue making changes until you're satisfied. The curve you draw will always be adjusted in level so that the maximum number of bits are available to describe the signal (optimized for dynamic range). Once you save the curve, you can access any of the settings optimized for any seat using the remote control and the display.

You can turn Logic7 on and off, adjust the level of the center channel, use a balance control, fader, 3 or 11-band graphic EQ or adjust the level of the bass. THe bass control isn't a gain control for the subwoofer output, it's a filter that works with the crossover and applies the right amount of bass to ALL channels so the illusion of bass up front isn't destroyed when you turn up the bass.

Answers to some likely questions:

1. You don't have to use the unit's volume control. You can use the one in the head-unit if you want to.
2. Maximum input voltage on the RCAs is 2V and 15V on the speaker level inputs. The signal is converted directly into digital after the preamp buffer, so a high signal level is far less important in this device than in conventional ones. The input is fully differential, so there won't be noise. I suggest speaker level connections because they are COMPLETELY isolated from ground.
3. The automatic EQ isn't exactly parametric or graphic. It's a very powerful algorithm that works on the impulse response to adjust both time and frequency response. It's amazing and does in about 30 seconds what I can do with an 80 band parametric EQ, crossover, time alignment and a serious analyzer in about 3 days.
4. The display doesn't have to be mounted. If you don't want iPod control or the ability to adjust after setup, you can unplug the display and use MS-8 as a "black box".
5. The unit is small--about 8.5" x 11" x 2.5"
6. Price will be about $800...TBD
7. The software is updatable via USB and a PC.

It does what all other OEM integration tools do and what every other DSP (EQ, Crossover, Time alignment, 7.1) processors do, but it sounds better, is easier to use, is less expensive and is far more advanced in terms of DSP power. Best of all, it's a tool you can be successful with, rather than a whiz-bang collection of filters and adjustment possibilities that require a PhD in acousitcs to use.

Tendean17 · Feb 20, 2011

t3sn4f2 :

Andy from Harman giving more details on the MS-8 and comparing it with the PXE-H650.
From Audiogroupforum.com

"There are lots of differences. First, MS-8 is more expensive.
Some technical differences are:
1. MS-8 includes power for speakers.
2. MS-8 includes a center channel output and a matrix surround processor (Logic7), which is more about fixing the image for the passengers than about reproducing an audio equivalent to a roller-coaster ride. The Apine doesn't include a center output and doesn't include that image processing.
3. The MS-8's crossover is fully configurable. It'll support any system of 8 channels or fewer, including 7.1, 5.1, 3.1, or the standard car-audio 2-channel bi-amped or tri-amped front stage and a sub. anything is possible, since all the channels can be anything, but crossover setup is manual in MS-8. It's automatic in the Alpine, but it's less configurable. The outputs are fixed.
4. MS-8 includes an auxiliary input and a remote control and display which allows you to make some adjustments after setup and includs a volume control for those pesky OE systems that include dynamic "bass elimination" (many GM).
5. MS-8's subwoofer level control is a shelf that's applied to all the channels through the crossover and the bass management algorithm. It'll preserve the impact in the front of the car AND add bass.
6. The automatic equalizers are completely different. The Alpine uses a 512-tap filter, which also equalizes phase and sets time alignment. It also includes some spatial averaging for multiple microphone placements (6). When you equalize with the Alipine, the first microphone placement sets the time alignment and the rest of the placements are used to smooth the frequency response over most of the car's interior. Multitap filters that operate in real time are a relatively new possibility. In years past, multitap filters in real time were only a hope, since there weren't many microprocessors that could process all that information quickly enough. The benefits of usiing a multitap filter are that they can be very precise and they equalize phase as well as magnitude since they operate on the impulse response measurement. For one tiny point in space, they can also eliminate the sound of plenty of reflections, but their ability to do that accurately diminishes in larger listening areas, since the effects of reflections at high frequencies can be very different even a few inches away from the original microphone position. The other important thing to note about multitap filters is that the 512 "bands" are distributed in a linear fashion rather than logarithmically. That means the resolution is fixed across the audio band. 512 taps gives you roughly 40 Hz resolution. That means you get 2 adjustment bands between 20 and 100Hz and 25 bands between 10k and 20k. Multitap filters, by default place more adjustment possibilities in the high frequencies than in the low frequencies because of the linear distribution of those "bands". That's the only drawback. The Alipine allows you to select from several target curves for adjustment after the automatic setup.

One more note about multitap: They are the shiznit for headphone EQ, because the "listening space" is fixed. With multitap EQ, you can add the reflective properties of a completely different space and transform the listening area to a completely believable representation of a much larger space. With speakers, that isn't possible yet because both of your ears hear both speakers and moving your head helps you determine the location of sounds (just like when your dog cocks his head when he hears a sound he doesn't recognize--we do the same thing, it just doesn't look so ridiculous).

MS-8's EQ is different. We also use a spatial average, but we use a binaural measurement system and 3 mic positions PER LISTENING POSITION. That gives us 6 measurements per seat for each of the 8 channels plus a time alignment adjustment for each seat. Once the setup is done, you can choose an optimization for any seating position and switch between them. For frequency response EQ, we make standard frequency response measurements, eliminate the phase measurement, average the measurements), calculate the phase response of the average, turn the measurement into an impulse response measurement, apply 8 biquads (filters) to the impulse response according to the target curve and the crossover settings using a very complicated and sneaky algorithm that I can't divulge because we're applying for a patent. The result is a VERY powerful EQ that can be implemented on a relatively inexpensive DSP for each channel and leave plenty of space to use the same algorithm on the eletrical signal of the MS-8's input for flattening of the input signal. The distribution of the bands is logarithmic and makes a completely adjustable target curve easy to implement and accurate. Each speaker location is equalized separately and, because of the spatial average, the acoustic sum of the channels matches the target curve. Once setup is complete, you can fine tune the car using a 31-band drawing tool. You draw the curve you want to hear and the MS-8 implements it and allows you to audition your changes vs. no EQ and vs. the automatic implementation of the predefined target.

Both pieces of equipment are technological marvels and they both include input channel summing and signal conditioning, crossover and EQ). MS-8 includes more stuff (center channel, Logic7, amplifiers, a display and remote, equalization memory and multiple seat optimization, center channel output and automatic input configuration--MS-8 will figure out what you've connected to the input regardless of polarity), but it should. It's more expensive.

Which one sounds better? You'll have to be the judge.

One last note: Both of these products are super-important and may help to revive the industry and get new customers interested in making their cars sound great while preserving their factory user interfaces. They have both been long development processes with plenty of invention and innovation, software development hiccups and decisions about which features to implement. Both products will require some new thinking on the parts of installers and salespeople about how one implements great audio. Simple 2-channel audio isn't dead, but these kinds of advancements make better listening experiences possible using a new set of rules.

Kudos to Jason ad his team for beating us to market. The other difference is that MS-8 isn't quite finished yet-but it will be.
__________________
Andy Wehmeyer
Global Product Line Manager, JBL Car Audio
Harman Consumer
______________________________________________________________

t3sn4f2 :
Me - "Hi Andy. If I dont have a center channel, can I use Logic7 mode with a no center setting and does it then do something special to the prcoessing so that its like a phantom center setup, or do I have to go with stereo mode in that case and loose the other benefits of Logic7. Also is the auto EQ only for Logic7 mode or is it availiable for stereo mode too?"

Andy Wehmeyer :
"T3, If you don't identify a channel as a center, Logic7 won't be completely engaged. There will be some ambience processing for the rear and the time alignment will be set for a single listening position. After setup, you can choose between optimizations for each of the 4 seats, but they won't all image simultaneously like they will if you use a center. EQ is completely separate from L7. It works no matter the setup."

t3sn4f2 :

The Main Menu is:
1 Input Selection
2 Audio Controls
3 User Presets (EQ)
4 Calibration/Setup
5 System Settings

The Audio Controls Menu is:
1 System Levels
2 Tone Control
3 Input Levels
4 Logic 7 ON/OFF
5 Controls Defeat/Active
6 Reset to default

The Tone Control will include the user-adjustable EQ.

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
Wow...another super long thread about MS-8.
I think most of the questions about the product have been answered here, thanks to someone pasting responses from the now almost dead carsound forum. Thanks for that.

I agree that there haven't been many room correction algorithms that work well, but the one in MS-8 is a good one. There's a huge difference between room EQ in a room and room EQ in a car. Correcting response in a room with real speakers really requires two different kinds of measurements--a near field measurement of each speaker and a correction scheme for that and a second set of spatially averaged measurements for correcting room modes. In a car, we can combine this into one, since the seating positions are fixed and since the speakers are very close to the listeners (compared to a room). There's so little delay between direct sound and reflections (the time and magnitude) are so close that we hear them all as speakers. In rooms, it's important to differentiate between the two.

Can a person do a better job of tuning than MS-8? I can and I'm sure there are a few others who can too, but MS-8 has been designed to make everyone's car sound better--people like my mom who love music but don't know anything about audio--and for installers to implement. I can do a better job than MS-8 but in order to do it I need lots of bands of parametric EQ (currently I have 176 biquads available for eq and crossover), time alignment, phase shifting parametric all pass filters, separate gain control of all speakers, Logic 7, a mic array and multiplexing mic preamp, an RTA with 1/24 octave resolution, an analyzer than can measure impulse responses and phase, and a couple of weeks.

MS-8 includes all these capabilities and does about 90% of the job in about 10 minutes.
Regarding the debate between Car PC and MS-8: A car PC can include many if these tools all kludged together (except Logic 7). A car PC is like a basket full of groceries and MS-8 is like a great meal.

Why am I building a car PC?
1. Because i'm a glutton for punishment
2. Because the aux adapter I use from USA-SPEC screws with the CAN bus in the car and prevents the steering wheel track up and down button from working properly
3. Because I'm tired of having an iPod controller screen stuck to the top of my dash
4. Because Gary Biggs is building a new car with a MAC Mini as a head unit, asked me for help in figuring out how to configure it and he got me hooked.
5. This is the biggest reason: Our DSP engineers are constantly working on new algorithms to do this and that (top secret) and we need a way to evaluate and car-optimize them as VST plug-ins before they become embedded solutions for home audio

Tendean17 · Feb 20, 2011

donkeypunch22 :
Andy, thanks for taking the time to answer. I needed some inspiration to keep the faith!
Oh, and Andy, if you could, what are "eight biquads of filters"? I know a filter is the same word for an eq, right? But what is a biquad?

Andy Wehmeyer :
Let's make this easy and say a biquad is a filter that can be configured to be a high pass of just about any alignment; low pass of just about any alignment, parametric EQ of nearly any frequency, gain and Q; notch, high shelf; low shelf or phase shift. The MS-8 assigns the filter type and values (frequency, Q and gain) based on the measurements it makes and the algorithm (predefined process or set of instructions for making decisions written as code) that determines how the decision will be made. So, for the purposes of this discussion, MS-8 has 8 opportunities per channel to implement something that does part of the job of fixing the channel's response. The details of how it makes the decisions are proprietary, patented and too difficult for me to try to explain.

I'm sure someone will flame me for oversimplifying this--OK, all you DSP programmers...flame away!
______________________________________________________________

quality_sound :
So is the process manual or automatic? I know I've heard that it's got manual controls but are those in addition to the automatic controls or on top of? What I mean is, do the manual controls adjust the settings done by automatic process or is it a separate set of filters?

Andy Wehmeyer :
Some of it is manual and some is automatic. The user enters the crossover frequencies and assigns the channels. Then the user helps MS-8 make its measurements by placing the microphones and pressing "Go". MS-8 adjusts the EQ and, consequently, optimizes the crossovers and slopes for proper acoustic performance. Then, if you want to make adjustments, you get a 31-band EQ. The 31-band EQ is a separate set of filters that you can use to draw whatever curve you want.

Unlike most 31-band graphic EQs, the response tracks the settings precisely. What many users expect is that if they boost all the sliders by 12dB, that the response should be flat, but boosted by 12dB across the spectrum. This is almost never the case, because making the filter Qs narrow enough to do that makes the response look like a comb. Making the filters wider provides more gain than one would expect when adjacent bands are boosted. Also, adjacent band boosts and cuts are rarely executed by conventional EQs as one would expect. The math used in MS-8's 31-band EQ adjusts adjacent bands automatically so that the curve you draw is the curve you get. This is a big deal, by the way.

For those of you who have an EQ laying around, plug it into your sound card. Make it a loop-back. Generate some pink noise and look at the response as you make adjustments. You may not like what you see and it's one of the reasons that tuning with a conventional 31-band EQ and using a 31-band RTA rarely results in great sound.

The whole point of MS-8 and the point at which it differs most from every other processor that's come to market so far is that it's intended to provide a bunch of tools you can use easily to be successful in making your car sound great. It's not intended to be the tool corral at Home Depot, where almost anything is available, but it's up to you to learn how to use it. If we just took the on-chip library from the TI DSP we're using and added a GUI, this product would have been finished three years ago, but it would have been just like every other DSP EQ/Crossover. There would have been a bunch of people who can pronounce "equalizer" and who have heard the terms "Butterworth", Linkwitz-Riley" and "All-pass filter" raving about the resolution of the available adjustments, but the success rate in making cars sound great and, consequently, the sales rate for the product would have been just as dismal as every one of its predecessors.

This industry doesn't need more tool boxes, it needs more carpenters and more folks who are willing to step up to provide real solutions. Giving a guy with no arms a garden hoe and a shovel doesn't get the carrots planted

"Placing the microphone" means putting it on your head and sitting in the driver's seat, turning your head from side to side according to the prompts, and moving to the other seats (if you want optimizations for those seats too). No change there.

Yup. 10 runtime presets. Once the auto EQ has been run, there's no need to do it again. The correction filters for the car don't change when you add your own spin with the 1/3rd octave EQ.
_____________________________________________________________

BMWTUBED :
Admittedly I haven't read all 2119 posts on this subject, but I have done a fair amount of googling related to the 325i install. I mainly find 3-4 of the same press release information from about two years ago. My question is: is there any documented release of information regarding how each of the 8 channels were powered, passive crossovers used, exact tweeks that Gary made after the MS-8 did it's thing, etc? I would guess the system was powered as follows, but would love confirmation if anyone knows:
left tweet / mid - 1st chnl (using passive x-over?)
center tweet / mid - 2nd chnl (using passive x-over?)
right tweet / mid - 3rd chnl (using passive x-over?)
left mid bass - 4th chnl
right mid bass - 5th chnl
rear left tweet / mid - 6th chnl (using passive x-over?)
rear right tweet / mid 7th chnl (using passive x-over?)
subs - 8th chnl

Andy Wehmeyer :
Stock speakers in a BMW winning a major competition
Almost sounds too good to be true
Almost correct. Sides and rear were driven by the same two output channels. JBL PX 300.4 amps on all the stock speakers and a BPX 2200.1 on the subs. Tweeter was added to the sides, rears and center with passive crossovers.

No additional tweaks after the autotune. The car won using the "compromise" optimization for the two front seats

Custom built passive crossovers I assume?
Just a capacitor on the tweeters. I think a 4.7uF.

Tendean17 · Feb 20, 2011

lenkiatleong :
Does anyone understand what Andy meant by "The car won using the "compromise" optimization for the two front seats. ". This comment sets me back one step. I was expecting MS8 to sound as good in term of image and soundstage) as what Mark Levinson has done for Lexus cars where every one in the car get good image and soundstage. Am i wrong to assume this?

Andy Wehmeyer :
The Auto EQ measures the distance (acoustic) to all the speakers connected to each of its outputs. Then it EQs the combinations of channels that are assigned to each spatial location (right front, left front, center, left side, right side, left rear, right rear and sub). If you choose the driver's seat listening position, time alignment is optimized for that seat, just as it is with other systems. If you choose the passenger's seat, it's optimized for the passenger's seat. If you choose "front seats", then the left and right time alignment is defeated, but individual driver time alignment remains (midbass to tweeter, for example).

Obviously, with no center channel, this works like any other time alignment scheme. Only one seat can be optimum. A center image in a stereo system depends on mono information arriving at precisely the same time, at the same level and with the same frequency response from both speakers. If you have a center channel and Logic 7, a great center image no longer depends on this because mono information is steered to the center speaker. If the center speaker has the same bandwidth and the same frequency response as the front right and left, it'll work fine. The driver's seat and the passenger's seat will sound the same.

What if your center speaker won't play below 300 Hz and you have to rely on a phantom center image for sounds below 300 Hz? MS-8 steers information below the center channel high pass filter to the right and left and eventually the subwoofer so nothing is missing. However, the left and right midbass wont be time aligned for either seat.

Is it a big deal? Not really, unless you plan to use only a pair of tweeters as a center channel.

Car audio enthusiasts typically prefer a strong and stable center image and OE consumers are less critical regarding imaging. There's an ongoing fight about where the center image should appear. I believe--after 25 years of building cars for customers--that it should appear in the center of the dashboard (or the hood, if possible) no matter which seat you sit in. Some of the folks who build OE systems believe the center image should be directly in front of each listener. MS-8 is tuned so that the center is in the center. If you like it the other way, it's simple to adjust--just turn the center channel down a bit in the menu.

MS-8 is not going to do everything perfectly--it wont peel apples or make gold out of a rock from your front yard--it isn't designed to be all things to all people. If you believe that you have to have a high-end D/A convertor or you have some other highly technical REQUIREMENT, you may find that another product meets your needs. MS-8 isn't some gold-plated, gold-PCB, million-dollar connector laden class-A multi-channel amplifier with all kinds of super esoteric crap designed to appeal only to enthusiasts. Those kinds of products make regular folks like me afraid to try audio as a hobby. I don't believe that in order to enjoy listening to a great sounding system, one has to be endowed by God with some heightened sensory perception and to be so enthusiastic about the gear that one would read (and take seriously) all the flowery prose intended to make religion out of science.

MS-8 is designed to make a great sounding system available to anyone who wants it at the most reasonable cost possible. There's plenty of high-end thinking packed into the little box and none of the components are compromised in the interest of shaving cost. We also haven't included any esoteric junk that inflates the cost for limited return. This is an exercise in reaching the point of diminishing return in parts cost and providing performance and ease of use that are without equal. If you're willing to stretch a bit in your opinion of what is and isn't required for great performance, MS-8 will be fun and eye-opening for you.

I'm expecting another iteration in the mail on Monday to begin final testing of input use cases--testing of UN-EQ, Logic7, all of the audio controls, etc, while the engineers finish coding the auto EQ for the TI610 DSP. Once that's done, we'll send the design to our validation team for final testing against the product spec-to be sure there are no noises or misbehavior.

This thing has taken so long to develop that there's no way we're going to skimp on validation. I can only imagine the backlash if we release something that hasn't been fully tested and we can't provide an answer to all questions regarding setup or use in various systems.

If the DIY spirit is about doing things yourself with parts you find at radio shack at great cost in money and time and you're a hardcore DIYer, MS-8 isn't for you. This thing is designed to be a tool that nearly anyone can use to be successful in building and setting up a great sounding system or improving an existing but not-so-great-sounding one without having to spend thousands of dollars on esoteric drivers, fiberglass kick panels, additional amplifiers (unless you require a car that plays much louder than a premium factory system), ridiculous interconnects and speaker cables, etc, etc, etc. If the DIY spirit is about finding tools and gear that allow you not to be dependent on a retail installer who claims that all of that is necessary to have a great sounding car, then MS-8 is what you need and will be worth the wait.

Sorry it's taking so long. I hope it'll be worth the wait for all of you. It will be for me. I have 50 people at work asking me every day when they'll be able to get one
______________________________________________________________

Andy Wehmeyer :
So...I'm back in LA for a week of meetings and spent some time in the MS-8 lab today. Next week, we should have the first fully-integrated software build and we'll start the rest of the debugging. The various modules have been debugged separately.

It sounds great and it works great too.

Here are a few cool features that might go unnoticed:

1. The subwoofer level control is a shelf filter that's applied to all the channels through the crossover instead of the usual gain control that most everyone else uses as a sub level control. The benefit is that once you get the bass sounding like it comes from the front (oh yeah, that's automatic) you can adjust the subwoofer level all you want and it never sounds boomy and never gives away its location. We were listening in the lab and we moved the subwoofer all over the place, ran the auto EQ each time and no matter the location, it always disappeared into the sound of the rest of the speakers. No amount of sub level adjustment with the shelf filter made the location any more obvious.

2. The 31-band EQ precisely tracks the curve you draw with the "sliders". With most graphic EQs, adjusting two adjacent bands in the same direction (say +6dB) will give you quite a bit more boost at a frequency in between the two sliders. If the filter Qs are narrow, you'll get two peaks at the proper amplitude, but with a hole in the middle. The EQ in MS-8 adjusts all the adjacent bands automatically to precisely match the curve you draw. This is a big deal and no other car audio EQ that I know of includes this "math". If you want to see what happens with a regular EQ, hook one up to your analyzer. If you use a PC and sound card, just connect it in a loop, play some pink noise and adjust a bunch of the sliders--you'll be shocked. If you find one that does what MS-8s EQ does, I'll be shocked.

3. The UN-EQ not only flattens the response, but it also removes any channel delay present in the OE system.

I've explained these features in previous posts, but they've basically been little additions to my wish list and today I saw them all working--and working perfectly. Ahhh...like Christmas for a little kid

If you want to use a laptop-based analyzer to check MS-8's work and to help tune the graphic EQ according to your taste, believe me, it's much simpler to have the laptop running only the analyzer. That way, you don't have to switch back and forth between windows.

The biggest difference and the one that will make MS-8 a better tool for making cars sound great is that all of that work that you're used to doing with crossovers, a bunch of EQ channels and delay settings are done automatically. The setup process is a serial process because everything is optimized once you finish the measurements. Setting crossovers is a matter of driver safety and potential output level, but you just make suggestions with the settings you input. The crossover points you choose are the ones that get implemented, but the response shapes are dictated by the combination of the EQ, which is applied before those filters, and based on an average of 6 measurements and the filter settings you choose.

All of the trial and error that IS current car tuning with conventional tools is eliminated and replaced by a process that knows the measurements and, through an algorithm (a list of instructions that aid the machine in making decisions), determines which solution set is most appropriate for fixing the acoustic problems. It WILL do a better job than any manual process at arriving the best possible optimization of the system, according to accepted science, the preset target frequency response and the requirement that the car has an image where the center is in the center of the dash, the stage is in front of the listener and the bass sounds like it comes from the front. Of course, that science doesn't take your preference into account and that's why we provide the tools that you'll need and want to make changes according to your preference.

It cannot fix a really ****ty system, but it can do a better job of making even a ****ty system sound better than another process. The use cases that it will support are based on reasonable and realistic system design. The minimum supported system is a simple stereo system--2-channels. A stereo system can be up to quad-amped, but the subwoofer output is a mono signal (2 output channels are available). I fyou insist on stereo bass, then you can tri-amp the stereo system. If you use rear speakers, subtract 2 channels from the channels available to do the rest. If you use a center channel, subtract one more. With MS-8, there's little benefit in separate channels for tweeters. The ease of crossover adjustability in tuning is moot--since you're not tuning and time alignment is unnecessary up there. MS-8 sets the levels and the response shapes via the EQ. We provide the ability to use separate channels for those who want that, but the addition of a center channel or rear or side speakers is a FAR better upgrade.

When we started designing this thing, I told the engineers that the auto-tune was fine, but that i wanted a back door to the target response and all of the filters because I believed my process, performed manually, would provide better results. Now, I know that was BS. It DOES provide my results because it's my process built in a little box that thinks much faster than I can and can sort through all the possible optimizations in a few seconds. It isn't a compromise

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
OK. So here's the update.

We've confirmed that the automatic acoustic EQ works with all supported use cases--and it sounds great. We've tweaked the target to provide a little more bass than I like, but the user menu includes a control to reduce and increase it.

We've added a "bypass input setup" to the setup menu, so if you have an aftermarket radio, you can bypass the Un-EQ part of the algorithm if you have a full-range output from the radio.

We've also discovered, in all of this testing, that CD transports and pickups are sometimes good and sometimes not so good. The algorithm that detects left and right and picks the impulse peaks to determine polarity and factory time alignment works PERFECTLY, so long as the CD player clocks at 44.1k. That might seem like a "given", but in reality, it isn't. We've discovered inexpensive and expensive aftermarket decks that don't clock at that rate and even some high-end OE systems that are off by enough to wreak havoc with signal correlation at high frequencies.

We're hot on the trail of a fix for that and should have some working code that eliminates that problem in short order.

What we're most interested in is a simple installation procedure that doesn't require any user troubleshooting, and in the interest of that objective, we've eliminated the audio-sensing turn-on, and have opted for a REM IN terminal that should be connected to the head unit's REM OUT connection. If that doesn't exist, then REM IN must be connected to the car's ACC terminal. MS-8's REM out terminal should turn on ALL the other devices in the car except the radio. For OE systems that don't include a REM wire for amp turn-on (some CAN-BUS and MOST systems), connecting MS-8 to ACC will suffice. We'll include a turn-off delay that's programmable to help eliminate any audible pops.

At this point, it's full-steam ahead.
______________________________________________________________

matt1212 :
Andy, Just curious why is it that when using an aftermarket head unit (Pioneer F90BT) with a front rear and subwoofer output its better to use just the front input signal? I would have thought that it would have been easier for the logic 7 to know all three?
Also, I'd assume this means that the signal from the head unit should be completely unprocessed, as in the full 20hz-20khz signal? (No crossovers, bass boosters, etc...)

Andy Wehmeyer :
Logic 7 generates center, side and rear from the 2-channel signal. Many of the other 5.1 algorithms are designed for rooms where the TV is the vocal point, dialogue is most often reproduced and where the center speaker is similar in size to the left and right, and finally where the seating positions are more similar to the standard equalateral triangle. It doesn't need to know the center and side information from discreet outputs. If you'll watch or listen to DVDs, just select the 2-channel downmix in the disc setup or the setup menu for your DVD player. In that downmix, the canter and surround information have been "encoded" into stereo and L7 will play back in full surround. Generating the sub signal for the car should always be done by MS-8. LFE channels for rooms aren't optimal for cars. Plus, just hooking up front channels is simpler.

As far as the additional processing in the head unit goes, it's pretty important to defeat the surround processing, if possible. MS-8 can't downmix perfectly if it's fed center and sides. If you run setup with bass and treble boosted, MS-8 will flatten the signal

Hey there, Hairy Night Creature,
Yes, that's right. If you don't install the display, the unit remembers all the settings when it's unplugged. It's like having stored only one favorite. In that case and if you've done a bunch of bass boosing in any of the EQs or level controls, I suggest setting the MS-8 volume below 0dB to accommodate that boost by a few dB and using the factory control. Obviously, the factory volume control's loudness and dynamic EQ won't be eliminated, but those settings are usually chosen to "correct" for some ambient characteristics (road noise, etc) or a simple equal loudness contour and they often improve the performance anyway

There's a lot going on behind the scenes that isn't apparent. the center frequencies on the 31-band EQ aren't adjustable, but it's really a drawing tool. Ms-8 will implement any curve you draw. Since the channels are already matched by the autotune, this is much easier and you don't need an RTA. As far as clipping and distortion are concerned, this unit doesn't operate any differently than any other unit. If you clip the head unit, you'll hear clipping. If you clip the DSP, you'll hear clipping. How would you suggest that we monitor a music signal from the head unit to determine if it's clipping? The unit does that during setup, but it can only do that because the setup CD includes a known MLS signal that MS-8 correlates to determine the impulse response. It has no idea what the music content should be.

Of course there's short circuit protection on the speaker outputs.

This thing will make your car sound great and that's what it's supposed to do. It was never intended to be an iPhone or a studio mixing console.

Think of this as a great backyard BBQ without the white tablecloth and water "with gas" that you'd find at Tavern on the Green

Tendean17 · Feb 20, 2011

Mic10is :
Andy what are the available XO points and slopes? Are the Xo points fully selectable or are they preset at like 1/3octaves? Which slopes are available and what type are they (linkwitz, butterworth etc...)
Do you think in a future upgrade it may be possible to allow all functions to be tweakable after Autotune?
also, what would you say is an ideal speaker size for center and rears in a Logic 7 set up?

AdamS :
Crossover Points are in specified in exact Hz. There are min and max frequency values in some places to prevent errors.
Slopes are 6,12,18, and 24 dB/octave
We use Linkwitz-Riley wherever applicable, otherwise Butterworth
______________________________________________________________

Mic10is :
Andy what are the available XO points and slopes? Are the Xo points fully selectable or are they preset at like 1/3octaves? Which slopes are available and what type are they (linkwitz, butterworth etc...)
Do you think in a future upgrade it may be possible to allow all functions to be tweakable after Autotune?
also, what would you say is an ideal speaker size for center and rears in a Logic 7 set up?

Andy Wehmeyer :
Adam has already answered the question about crossovers.

I think once you try thie product, you'll find that all of that other stuff dosen't need to be tweaked after autotune. One of the reasons that we're all after all of this control of every parameter all the time is because stuff is never quite right. Trying a thousand iterations of crossovers and EQ takes us a long time, but it doesn't take MS-8 very long to try thousands of solutions. The slopes and frequencies get adjusted when each channel receives it's share of the EQ. Furthermore, MS-8 builds the system as you enter the info. Tuning with crossovers is overrated and I find that there are basic settings that work for almost all systems, so long as they're EQed properly. I can't tell you how many championship-winning IASCA cars I've helped out with where the single biggest problem was crossover selection based on a bunch of wishful thinking. Front stages crossed at 40Hz while the sub was crossed at 60 or 70, midbass that plays from 40-60Hz, etc. The best way is to use steep slopes, minimize interaction and cross speakers ABOVE their resonance. So long as you follow those rules, you'll be happy with the results.

MS-8 may make this so easy that some of the mythology that currently makes tuning an "art" may be dispelled. I'd love that and so many more cars will sound great as a result.

At the begining of this program, I told the engineers that I wanted a back door so I could tune stuff myself. I no longer want that back door, I just want an MS-8 of my own so I can put away the RTA and the mics and have fun listening.

Fortunately, I get serial #1. Adam already has one
______________________________________________________________

Andy Wehmeyer :
Well...that's a good question. That one uses a DIRAC autotune, which is a fairly long FIR filter for each output channel. In my opinion, it's kinda like taking a Sikorsky SkyCrane to the grocery store. I've never heard it, but considering the amount of processing in there, there's no excuse for it not to sound great already. What improvements are you hoping for?

If you'll help me out a little bit and tell me how much power the marketing materials say the 9-channel amplifier provides and tell me the year of your car, I'll look up the connections in the chassis wiring book and see what can be done.

There is one potential problem with using the output of the factory amplifier--the FIR filter equalizes phase and frequency response. Logic7 in MS-8 relies on flat phase to steer front and rear. Depending on what the FIR filters actually do for each channel, it could cause some strange steering.

Send me the info (power and year of the car) and let's see if the DIRAC amp is MOST or how the head unit hooks up to it. I think some of my DSP buddies at work have some documentation too. I'll check with them.

Here's my take on the car tuning thing:

An algorithm can only be successful in meeting its objective if the person describing or implementing the process knows what he's doing and can communicate the process properly to the person writing the algorithm. I know how to tune cars and Ulrich and Adam know audio and math. That's why MS-8 works.

The performance of these kinds of things has far less to do with the components (DSPs, D/As, speakers, filter types and amplifiers) than it does with the target curve and the method by which the machine arrives at the target. I'm sure that MS-8's autotune won't make everyone happy, but as Technic says, having some additional tuning tools will help to satisfy the tweakers
______________________________________________________________

Se7en :
Andy, I have a question regarding the time allignment functionality of the MS8.
I have been a Tact Audio user for about 11 years now and this system also features automatic room equalization and time alignment correction. One of the realities/negatives about the Tact correction system is that it is not "self aware" of it's own processing time (which increases/decreases based on frequency and XO slope selection) and does "not" make appropriate adjustments to it's TA corrections. This means that often out of the box adjustments can be mis-alligned.
A number of Tact users have published useful charts to help users perform manual TA adjustments to the auto allignment based on XO frequency and slope between the subs and mains (in a home system).
I was wondering how the MS8 deals with this especially given the number of channels supported. Thanks!

Andy Wehmeyer :
Great question and a big problem if you sit far from the subs and speakers. In a car, we sit very close. Delay based on distance in a car is a very few degrees of phase at most sub/midbass crossover points (unless the box is in the back of a Ford Excursion and you're 4.5 feet tall). Using a really steep slope (4th order is sufficient) will minimize the interaction of the sub and midbass enough to make a good transition in the car possible. Since we're not using FIR filters, the computation time is relatively short for EQ and doesn't vary from one channel to the next regardless of how complex the filter is. We have compensated for the difference in processing time for the sub and the other channels.

The real difficulty in the processor is in the bass management section, where steered low frequencies are added to other channels below the crossover frequency of the channel to which that bass was originally steered. Thanks to Ulrich and Adam, we have an ingenious arrangement of additional variable all-pass filters that match the phase of the steered bass to the unsteered bass.

While that's not the answer, precisely, to your question, it is an cool part of the MS-8

Oh, one additional thing that should be mentioned. In any peak-picking algorithm designed to acoustically locate speakers or align signals, the amount of high-frequency in the content affects the algorithm's ability to locate the speaker, especially if a stndard impulse response measurement will be used. In MS-8, this applies to D-EQ and Auto-EQ. Once again, Adam and Ulrich have figured out a cure for that too.

The one problematic configuration is a 2-way speaker using a passive crossover where the tweeter is much closer to the listener than the mid. My tip for that scenario is to cover the tweeters during the first of the 4 sets of sweeps for each seating position. That will cause the MS-8 to align based on the output of the midrange, which is important because we use ITDs to localize sounds below 1kHz, and then the EQ will take care of the ILD at high frequencies. Works like a charm

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
We haven't done any listening tests in the same car, same system with two differnt head units. From a technical perspective, I don't think it matters much, so long as the dynamic EQ from the OE system (if such a thing exists) in the car provides acceptable improvements

MS-8 matches all the channels and tries to maintain plenty of room to avoid digital clipping. If you use MS-8's amp for the center, it'll adjust right and left to match it. same for rears. I suggest using similar power for center right and left. If you use the MS-8's amps for the rear, sent the gains on your amps down a bit, run setup and then adjust the amp gains up by precisely the same amount until you get a balance between front and rear that you like. Same goes for the sub. If the sub "level control" in MS-8 doesn't provide enough bass for you, adjust the amp gain up after you run setup.

The spatial average helps to smooth the high frequency measurement, which is always inaccurate with a single mic placement. If you mod the mics for an inner-ear implementation, you'd need a new mic correction filter and honestly, I don't think it'll matter much. The difference would probably be a little more high-frequency content allowed from the rear speakers

Lycan is right.

In fact, I'd go so far as to say that a mono (L+R) center does more to screw up a good image than to enhance it, unless there's some additional trickery involved in the processing, like crosstalk cancellation, but that isn't available in car audio products.

For 2-channel listening, a matrix processor (PLII or L7) is the ticket, although getting used to listening to music the way it was recorded takes a little while. We're so used to listening to different pathlengths that many of us like the sense of space that those phase anomalies provide.

I did a demo for a well know record producer several months ago of a car with a strong phantom center image and his response was "it sounds mono". the demo was of a track he mixed. We then went upstaris and I showed him exactly how much distinct left and right were actually in the track and he scratched his head.

Tendean17 · Feb 20, 2011

Andy Wehmeyer :

Andy Wehmeyer said:
OK, now that my butt has recovered from 15 hours in the same chair yesterday, here goes.

Every few milliseconds, Logic 7 computes a steering angle by determining a couple of things about the stereo signal. Left, right and center steering is determined by level differences between left and right signals. Mono information--the information that's common to left and right--is steered to the center. Don't confuse mono as L+R for what's actually going on here. Although this isn't how it's done, you could think of the mono info as (L+R)-lL-Rl. It's all of the sound minus the sound that appears only in the left or the right. That mono information is also attenuated in the left and the right. The amount of that information that appears in left and right is an important part of making this thing sound great in cars. If the information is completely removed from left and right, then I don't think it works very well in cars. The images are spot on, but they're really small and don't sound believable, to me. If the mono information isn't attenuated at all, then the stage is narrowed a bit and the images are big and kind of nebulous. Left of center and right of center aren't very accurate. Attenuating the mono signal in left and right by 6dB works great and that's what MS-8's L7 does.

The front-to-rear steering angle is computed by determining the phase relationship between left and right. -180 degrees steers to the rear. -90 degrees steers about halfway between front and rear. The level differences between left and right determine rear left and right steering.

Finally, when information is steered front, one of the rear and side channels is polarity-swapped. This helps to anchor front steered sounds in the front, and that's why sides and rears shouldn't make a bunch of midbass. That out-of-phase condition in the back works for imaging, but not for midbass. Fortunately, we all want the midbass to sound like it comes from the front, so it's just easier to make sure that it does.

Side and rear are also delayed by several milliseconds, but rear is delayed more.

Very low frequencies are not steered. They're simple stereo.

Here's a set of drawings that sort of illustrates this. The first one illustrates what you can expect from a L+R center.

And here's a simplification of what happens with L7:

So, when you implement this in a car, here are some pointers:

1. Sides and rears don't have to make bass. I use a 3" and a tweeter in the back and a 6" component system in the sides, but they're crossed over at 100Hz, 4th order. The 6" speakers in the sides are overkill, but that's what fit in the doors.

2. It's helpful if the side and rear tweeters are at about ear level. Don't mount them in the bottom of the doors.

3. If you won't put 4 speakers in the back and you'll use only 2 instead, MS-8 will process them as sides. That's no problem and the difference between 5.1 and 7.1 is really hard to hear. It's nice to have rears and sides if you'll have rear seat passengers. Logic 7 in OE systems mixes some front-steered information into the sides, so the rear seat passengers have their own stage at the back of the front seats. Making those adjustments is car and speaker-location dependent and it's seriously difficult to do without screwing everything else up. MS-8 will provide a single stage, located at or in front of the dashboard for all seating positions. I prefer that and it happens automatically and easily when the front seats are right.

4. Front steering works perfectly for all frequencies that the center channel will play and pathlengths don't matter much. For midbass frequencies that the left and right will play but the center channel won't, pathlengths are critical. A phantom center has to be generated for those sounds. If your center channel is a 3" and you have big-ass 8" speakers for right and left mounted in your doors, there's gonna be trouble.

5. The subwoofer level control in MS-8 isn't a level control for the subwoofer output channels. It's a shelving filter that boosts bass in ALL channels below 60Hz, but never above 160 Hz. When you boost or cut, the slope of the filter is adjusted. This helps to maintain the illusion of bass up front by sending the appropriate amount of midbass to the front speakers and bass to the subwoofers. It also maintains the crossover point. Here are a couple of graphs that illustrate this and why it's better.

The first one is a conventional control. It's pretty obvious why this moves the image of the bass to the back and sounds boomy. There's too much interaction between sub and midbass. This is why people claim that "underlapping" the bass is important and why so many people try to get big-assed midbass drivers in the front of the car and cross the sub at 50Hz. With MS-8, that's totally unnecessary, and the evidence of that is in the second graph.

I use a pair of 6" speakers driven by about 40 watts in the doors, another one in the center channel and a 500 watt amp on a pair of cheesy 10" woofers in an IB in the rear deck. The bass is seriously anchored to the front and the midbass is great. No hassles, no constant tweaking and I can boost bass by as much as 10dB on top of the target curve (which is a total of about 20dB) without the image shifting to the rear. I think my crossover point is about 80Hz, 4th order.

BTW, this feature is also included in the MS-Amps using a wireless rotary control that can be paired to any number of bass and full range amps in the system.

So, for front right left and center, you'd be better off with three 5" speakers (R,L, and C) than with 8s in the doors and a 3 in the center. If all you can get in the center is a 3", then try to move the midbass to the kick panels. If you can't do either, the car will still sound great, but the image for center-steered midbass sounds will be larger than it should be and will be biased a little bit to the side on which you sit. Not such a big deal.

Tendean17 · Feb 20, 2011

BMWTUBED :
Isn't this more or less what the MS-8 Biggs BMW had? Except the 8s are under the front seats and the center is a 4" (I think)

Andy Wehmeyer :
Yup, and it works fine, with the same caveat about midbass placement. was good enough to win IASCA at SBN, but the under-seat placement is a little better than in the doors. We crossed the MB and the 4s at 200Hz, 4th order
______________________________________________________________

Andy Wehmeyer :
The 2-channel downmix available on most DVDs is designed to encode the discreet channels into the 2-channel mix in a way that matrix processors can decode. See the diagram above for a clue about how those signals are mixed in. If you were in your living room at home, then there would be no reason to do any of this. The encoded stuff is designed for listening in rooms an designed mostly for movies. The principal objective (as it seems to me) is to anchor dialogue to the screen and provide 3-D effects for movies--flyovers, frog noises all around, dinosaur steps that shake the house, etc. From my perspective, LFE and bass management are fancy terms for "crossover and summing circuit". While they may be complicated to explain and they sound cool, bass is easy and it's even easier in cars.

Some music DVDs are mixed in multichannel so that they place te listener in the audience and some are mixed so that they place the listener on the stage. You can probably guess what my perspective on that is. In any case, discreet algorithms are often panned by music lovers as "non-musical" sounding and I tend to agree, unless the system includes a huge center channel and subs placed all over the room and are tuned differently for music than for movies. Furthermore, I've never in all my life heard a multichannel movie demo that i cared about. I've never heard an explosion or a dinosaur and have no idea what they're supposed to sound like.

MS-8 is designed to make cars sound great, but not necessarily to preserve all the whiz-bang technology that marketing materials for multichannel processors and processes tout. The primary benefit of Logic 7 in cars, considering the overwhelming use case of 2-channel material over discreet multichannel recordings is in center channel processing to overcome the installation headaches of kick panel mounting necessitated by off-center seating. For cars that have a center channel location this is a big deal for a pair of listeners in the front seats. Even for single seat listening, I prefer a dedicated center channel. A secondary and more subtle benefit is the front rear steering and the sense of ambience it provides. This depends on the recording, but I've never wanted to turn L7 off, no matter what I'm listening to. Of course, none of the multichannel stuff means much if there's no room correction, and that's the primary sonic benefit of MS-8. Finally, the fact that this thing can be hooked up to nearly any system is important too.

Logic 7 is not intended to synthesize a room independent of what's contained in the recording. Room synthesis is a separate project and that's why there's a mac Mini in my car--hint, hint...

Compared to what else is out there for making cars sound great easily, MS-8 has no peer, unless you're a seriously accomplished tuner with some tools that aren't available in over-the-counter car audio prodicts
______________________________________________________________

Technic :
So other than the 31 band EQ for manual adjustment, the MS-8 auto-EQ curve should be fundamentally the same used as the OEM?

Andy Wehmeyer :
No. The BMW doesn't autotune. The curve is different. The bass contol is completely different. The tone controls are different. MS-8 allows you to save favorites and optimize for different seating positions. The MS-8 includes preamp outputs, a display and a remote control. The MS-8 provides a center output that supports a tweeter. The MS-8 will do 7.1 and the 3-series BMW is a 5.1 system. The MS-8 is fully configurable and will support 48 different speaker system configurations.

I'm not sure what kind of answer you're hoping for. If you want me to tell you that MS-8 will provide better sound than the L7 OE system, I can't that depends on your requirements. If you want me to tell you whether it will sound better than the DIRAC system, I can't do that either. I've never heard one
______________________________________________________________

Technic :
The question was more towards what differences are between the Logic7 processing and implementation in the MS-8 and the BMW OEM version, not what the MS-8 can do other than that.
And so far there are at least 4 distinct differences.
Thanks.

Andy Wehmeyer :
These are the main differences.
1. Bass Management is not done inside of MS8's L7. L7 doesn't generate a subwoofer channel. It is done with the MS8 Allpass / Bass Management system alluded to earlier by Andy.
2. MS8 handles 45+ configurations. Whether the system is 2.x, 3.x, 4.x, 5.x, 6.x, 7.x, etc, Logic 7 will be optimized.
The rest is pretty subtle
______________________________________________________________

Andy Wehmeyer :
The EQ, time alignment and user adjustable features work whether L7 is on or off. Yo may find that those features provide such a benefit that the additional spatial effects provided with your discreet multichannel recordings pale by comparison. I can assure you that if you can select a 2-channel downmix, the playback will be better than a discreet solution that doesn't include great EQ and time alignment

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
The guitar and the vocal don't travel through the PCB nor the solder. A simple alternating current passes through them and yes, it'll make it through to the other side just fine. The point is that it's necessary to change the signal to make up for the deficiencies in the acoustic system. Think of it as believability rather than accuracy.

How about this. Let's say your wife or girlfriend sends you to the store to get 10 cans of Green Giant creamed corn. You're walking home from the store with the corn and as you round the corner at the stoplight, someone jumps out from behind a bush, hits you over the head and tries to take 5 of those cans. You fight him off, but 5 of the cans are dented and the labels are scratched up. Knowing that your significant other won't tolerate anything but the most pristine of cans, you walk the rest of the way home lamenting the damage. Just as you turn to enter your driveway, there's a guy who can, with the wave of a magic wand, restore the cans to like new condition. In order to get him to do it, you'll have to renounce your belief that to repair the cans and the labels is somehow an anathema and admit, for all to hear, that a repaired can is, in fact, equal to an unaltered can. You reluctantly agree, fall on your knees and recite, "A can is a can and will always be a can so long as it's a can." The can wizard smiles and with a wave of his wand, the cans are as good as new.

You enter the house, and your wife (or girlfriend) asks, "Hey, what was all that commotion I heard outside?"

You reply, "Nothing, honey. I was concerned that the cans of corn might have been degraded during my walk home. Here, check them out."

To which she replies, "Wow, these are the most beautiful cans of Green Giant creamed corn I've ever seen. I don't remember the picture of the corn being quite so vivid and the logo really stands out. Hey, check that out. You can almost make out the shape and size of the Giant's giant wand. Oh my God. You're so handsome. Dinner can wait." She leads you upstairs to the bedroom and makes love to you like you were both 19 again.

Hmmm...I'll take some...EQ, that is, and a bunch of other signal processing too, especially if it makes listening more enjoyable
______________________________________________________________

Andy Wehmeyer :
I'm not getting into a super lengthy discussion of the differences between using TA and equalizing pathlengths. That's been discussed and hammered on ad infinitum on this form and on others. From my experience, the difference isn't worth the hassle and the kick panel locations in many cars cause as much difficulty (frequency response and the car's drivability) as they eliminate (near side bias).

If you want to solve the problem once and for all and don't want to make fabrication your life's work, use the existing center channel location (if your car has one), buy an MS-8 and leave the right and left speakers in the doors. If you don't have an existing center channel location, then you'll have to choose. MS-8 makes the choice a little easier by providing good auto EQ and TA and it takes 5 minutes to tune.

I've seen some 5-minute kick panel jobs and none of them sounded very good--they were precisely aimed with a laser, though
______________________________________________________________

BMWTUBED :
Andy,
I'm planning something very similar to the 3 series you and Biggs did.
What are your thoughts on using a good mid/sub in the underseat locations and foregoing the sub in the rear? I was thinking of the Morel ADMW9 running
30hz - 200hz. Any real down side to this other than the low end limitation?

Andy Wehmeyer :
Keep the subs that are down there. I use a pair of the same 10" models in an IB in my rear deck and they rock. A couple hundred watts each and you should be fine (so long as you're not a basshead). If you decide that you need more bass, consider adding a sub in the back. Just changing out the speakers under the seats isn't going to be a night and day difference
______________________________________________________________

MattyKHZ :
OEM and SQ is what I am after which I believe is what the Audison Bit One was providing once all the bugs had been ironed out.

Reading the whole of this thread it seems that Logic 7 and centre channel usage is what the MS-8 is all about. Unless I am wrong.

I am wanting to get the best out of my Denon amps and MB Quart speakers in a 3 way active set up from my OEM HU in R56 Mini Cooper S. As the MS-8 will probably not be over here this summer with US probably getting the first allocation I would have to look elsewhere and Audison seem to offer everything I need and be available.

Andy Wehmeyer :
I'd say that L7 and its center channel processing are a big bonis, but the fundamental advantage of MS-8 is the signal summing and conditioning and the auto EQ. The BitOne is a great piece of gear too. If you're a good tuner, you'll be able to make the car sound good. If you're not, the MS-8 is the tool for you.

Basic signal summing and un-EQ for the factory signal is one VERY small part of the chain. There's nothing very important about getting a flat signal from the head unit unless you're also running a process that requires a flat, in-phase signal. Furthermore, simple analog summing is an insuffieicnt solution. The output signals from many factory systems include delay, which will cause comb filtering in the summed signal. Is that a big deal? Depends on how low the first part of the "comb" is. If it's in the midrange, then it may be a big deal. MS-8 fixes that by eliminating the relative delay.

Matt,
Obviously, I'd like you to have an MS-8 and you'll get an email this week from our office in Amsterdam, but I also understand that delivery timing is a big part of your decision. Over the course of the last 3 or 4 years I've helped several people here and on the carsound forum choose the best alternative because they couldn't wait any longer. If you can't wait any longer, I suggest choosing an alternative based on whether you have an RTA (or have access to one) and whether you're good at tuning cars. If you are, the BitOne is the right way to go. If you're not and you can't wait any longer, get the Alpine.

Tendean17 · Feb 20, 2011

2DEEP2 :
Andy spoke like a true Product Manager, knows where his product is...
But someone can offer a short sell on product before it even hits the street from the OEM... Hmm.
You know China has some great engineers, they can copy just about anything
It would not be the first time a Chinese firm committed patent infringement.
Aren’t those fake Focals from China?

Andy Wehmeyer :
Not possible with this product. The DSP in this thing is a custom part built only for Harman. Additionally, some of the other components are seriously long lead items (120 days) and we've gobbled up the entire world's existing inventory. I can't say that there will never be a counterfeit MS-8, but I can say with certainty that a counterfeit unit won't be available first.
______________________________________________________________

ibanzil :
When using imprint there was a way to open up the files audyssey used when calibrating the system. Will there be any way to see what the MS-8 did as far as adjustments to TA, eq, and such?

Andy Wehmeyer :
Nope, at least not yet. Originally we were thinking about making that part of the code update downloader, but in the interest of getting the project finished, we scrapped that. I thought it would be cool to have a way for users to upload the files to JBL.com so we could provide a view of what had been done through some online tool and so we'd be able to help troubleshoot if necessary and collect data about cars. All of that is a pretty big piece of work and I haven't had time to make a business case for the investment in time that would require.

Troubleshooting is turning out to be really easy, so far. For the 10 or 12 cars that have MS-8's currently, there have only been a couple of questions like, "Hmmm...the midrange sounds strange", but a 5 minute conversation about the speaker system, a suggestion or two and an another round of acoustic calibration (a 5 minute process) has fixed all of those quickly.

For example, one of our guys has a Dodge Durango witih the OE Alpine system. He had plans for an additional sub, since the factory sub is a little 8" in the back somewhere. He decided that to get started, he'd just put in the MS-8 and do the sub later. He called and said the midrange sounded strange and wasn't super happy with it and I asked if the tweeters in the front were actually tweeters or if they were those standard Dodge 2" high frequency speakers. Who knows? Anyway, the grille was too big for a tweeter and the mids are in the bottom of the doors. He had the crossover set at about 4k, so we moved it down to about 1k to take the mid out of the response before it started to beam and before the first big peak in the response (happens with all cone drivers). 10 minutes later everything was fixed and he's completely happy without adding an additional sub. Must have been 2" drivers...

I guess my point is that if there are so few people who can be successful in making their car sound great using basic audio tools and an RTA, what's the benefit in a print out of some complex series of impulse responses, except to take the techno-babble to a much higher level and expose our I.P to the public? If you want to know what MS-8 did and you have a PC based analyzer, there is a way to connect the analyzer to the input and output of the MS-8 to make a measurement of each of the output channels. In order to understand completely what's going on, especially for any system that includes "fixing" an OE head unit, each part of the process would have to be separated and displayed. When you set up MS-8, it'll be apparent that the process is a series of steps that have to be performed in a specific order. Correction filters for each step aren't saved in memory, but the correction for the end of each separate series is.

If you're technically astute enough to decipher the data, then you're technically astute enough to figure out how to make those measurments with any analyzer that does FFT and IFFT. A careful read of the manual and some careful thinking about the setup process will make that obvious.

Here's a challenge. The first person who figures out how to measure this using an audio analyzer wins a free pair of JBL 660 GTi components--on me. Send the documentation to me, be sure to send the data as screen shots or text files with an explanation of what you think is going on and tell me the equipment you used to make the measurements. Adam and I will look over it and determine whether you're a winner. I'd love to give away a set of these speakers and will give some serious benefit of the doubt in order to give them away.

What I mean is, get close enough and we'll send the speakers. I'm not looking for a loophole to prevent me from sending $2000 dollars worth of speakers--I'm not your health insurance company.

Oh, one more thing. Submissions have to be kept strictly confidential.
______________________________________________________________

Andy Wehmeyer :
Show me what the unit does to the signal for each of the output channels you're using and include what happens to the input signal too and you win a pair of the world's best car speakers. Think of this as a science experiment. You have to show your work or explain it and it has to be presented privately and not as a group effort for all to see on this forum.
Have a GTG in someone's back yard, figure it out together and appoint someone to be the presenter.
I'm not giving hints.
Oh, don't bother with Logic7. That'll be too difficult to decipher with the tools you'll have. Plus, I think I've explained it sufficiently here.
______________________________________________________________

kaigoss69 :
Andy, for a 3-series BMW (e90) with OE Logic 7, what is the recommended configuration if I am only interested in driver seat optimization? - Would you do exactly what you did with the 2006 325i or based on what you know now, would you make any changes?

Andy Wehmeyer :
As far as the speaker system goes, I'd definitely add a tweeter to the center channel. If you need the car to play loudly, you'll want to add amps, but if not, add a sub and use theunder-seat 8" for midbass. It'll work great.

Tendean17 · Feb 20, 2011

michaelsil1 :
Does the MS-8 save the settings if the power is disconnected?

AdamS :
It saves when
1. setup finishes
2. remote in is disabled (ignition off)

If you lose power and get it back, it will go back to 1 or 2, whichever is the newest
______________________________________________________________

rain27 :
Is there any way to reset the MS-8 to original factory settings without going through the display menu system??
Pressing the reset button on the unit itself does not reset all settings.
I was wondering if there was any other way????

AdamS :
There's one other way, but it hasn't been made public yet.
If you re-download the firmware, it is almost identical to resetting all settings.
The Reset button on the unit is the same as Remote In disable followed by Remote In enable
______________________________________________________________

kaigoss69 :
The BMW OEM L7 systems are known for the "rear fading issue". I have experienced this myself. The music will sound normal and then suddenly, it would shift certain portions of the song towards the rear speakers. It would sound weird, of course, an not natural as the fading would go back and forth. For me, this happened mostly when listening to FM radio and not so much when using CDs or the iPod. So is this the same issue (phase) and what exactly is causing it? Will it sound similar with the MS-8? How much out of phase information is in today's music and is it dependent on the source?

Andy Wehmeyer :
This is why L7 has an "off" position. It's a matrix processor. It does what it does and with most music it sounds great. All kinds of phase shift happens in FM broadcast when reception isn't good. Additionally, FM is about as close to mono as a signal can get and the blend function in many tuners which is designed to make poor reception sound like good reception further monos the signal.

FWIW, and as DS-21 has stated, all of the matrix surround processors steer according to phase. The difference between L7 and the others is that L7 does a much better job of keeping left and right rear separate. If you like it, use it. If you don't like it, don't use it. If you have a center channel and want to use L7 steering for the front (phase has no effect on front steering), then don't install rears or move the fader to the front.

Listening to just the rears or the fronts only with L7 is an unintended use and will sound strange, of course.

One day at CES, some of the EMEA judges got in my car (it was the demo one year) and after listening for nearly 2 hours, one of them jumped out and exclaimed, "I did it. I found something that breaks the code! Listen!" He put on some preposterous recording of a person speaking while pink noise rotated around the car. Of course, with no matrix, the pink noise won't rotate, but the combination of rotating steady state noise and a transient signal like someone speaking break the code. My response? "Dude, get out of here. It took you two hours to figure that out? You're the only one after four days who has cared that the code doesn't work with all songs and all recordings. If I were you, I'd throw that disc away.It won't work in any system. I also breaks stereo, since the noise is specifically encoded to rotate with a matrix."

I have a friend who works at an Italian car audio magazine who refers to those people as the "Car Audio Taliban". Seems fitting.

Tendean17 · Feb 20, 2011

Prepare your system design first

Andy Wehmeyer :

OK, I'll try to clarify and to overcome my commitment phobia. MS-8 is designed to make a competent system sound great. It can't fix profound errors in system design. Speaker placement is still important, but not as important as it is with basic 2-channel systems.

1. MS-8 can autotune just about anything, but it can't put back what isn't there. There's a lot of boost and cut available, but if it runs out of boost, it runs out of bits and that will cause digital clipping. That sounds like ugly speaker distortion. You can help MS-8 do its job if you hear ugly-sounding distortion by making sure you don't have any huge holes in your response. This is why Adam is suggesting a bacis level matching of your system before the autotune. Some other installation mistakes that may cause MS-8 to run out of boost are:

a. Choosing bandwidths that your speakers are unable to play. Don't try to cross an 8" that's mounted off axis in the door at 3k, for example.

b. A big suck-out caused by incorrect driver phase. Phase reversal may be a function of distance between a midbass and a mid or a mid and a tweeter. It can also be the result of improper wiring. If you hear distortion and it goes away when you change your crossover point dramatically, this may be the cause.

c. Choosing too narrow a bandwidth for your midbass and having the gain on the amplifier too low.

You can determine if the problem is that MS-8 has run out of boost by turning processing off. IF you turn it off and the distortion goes away, check these three things first. If you can determine which speaker makes the distortion, turn the gain on that amp up slightly and recalibrate. If that doesn't fix it, but moving the crossover point does, then try the new crossover point or try reversing the phase of the right and left channels of only that specific amplifier.

We've only seen this a couple of times and it isn't an MS-8 defect, it's just that the system has too big of a problem for MS-8 to fix on its own.

2. I don't think I've been noncomitttal on the center channel bandwidth issue. THe center channel should be the same bandwidth as the front right and left. In some cars this is possible and in some it isn't. IF you can only get a small speaker in the center, MS-8 will steer center midbass to right and left. If this is the case, then the location of the front midbass becomes more important. IF you can get your center down to 120Hz, that's great. If not, put in what you can and listen to it. The compromise will only be in matching the listening experience for both seats. the seat for which you choose optimization in the seating position menu will sound great.

3. The center doesn't play L+R. It plays only what is common to left and right. That's like Left plus right minus the absolute value of left minus right. A mono signal that's the ssame in left and right will steer to the center and will be attenuated by 6dB in the right and left. A left only signal will NOT appear in the center and a right only signal will NOT appear in the center.

4. Signals that are the same in the left and the right, but are out of phase between left and right will be steered to the rear by logic7. This is designed to help reproduce the ambience contained in many live recordings and the ambience that's faked using reverb or convolution reverb in many studio recordings. It works well, but some recording engineers reverse phase in studio recordings to create some other "effects". MS-8 doesn't know the intent of the engineer (and i don't either), so MS-8 will do what it does. If you listen to a multi-channel recording down mixed to 2 channels, MS-8 will recreate the rear using only the two channel mix.

5. Rear speakers should be full range, above 100Hz. The high frequencies are helpful in recreating ambience, so they should play to 10kHz. I recommend that the don't play much of anything below 100Hz, because Logic7 reverses the phase of one side of the rear during front-steered sounds to anchor the image in the front. When sounds steer rear, the phase of the sides and rears is corrected to make the steering more appraent. THe reverse phase for front steered images mucks up the impaact a bit, so cross the sides and rears over at 100 Hz or higher with a steep slope. This reversal happens very quickly each time L7 samples the music. I don't remember the interval but it's short--too quick for you to hear. One thing that L7 cannot do is steer a transient signal rear, if a steady state signal is steered front at the same time. I have a recording that makes this easy to hear. Mono pink noise that steers center and a voice that rotates around the car doesn't work. Fortunately, there are very few recordings that are made this way. L7 doesn't work on every recording and recordings that have little stereo or reverse-phase info won't place much info in the back of the car. Live recordings and direct to 2-track will be awesome.

6. I've never used MS-8 with horns. horns often require lots of EQ to sound right. Thay also require very little power. Be careful with the level of the sweeps, but I don't seeany reason why they wouldn't work.
______________________________________________________________

Bikinpunk :
Oh, and to you dashmat guys… you can throw that sucker away.
I don’t use one. It was hard for me to get rid of smearing with my bitone tune.
The ms-8 does this perfectly. No ‘walking’ stages, no smearing of voices. Everything is focused and focused well
______________________________________________________________

Bikinpunk :
Yea, right after you go through the system setup, toward the end of that, there's the part where you can test the channels. If you look at video #1 and go to about 5:56, you can see where you're at the "output diagnostics" screen. You can toggle the left/right side on & off. You'll then be able to level match your speakers from this screen and you'll have the protection of the ms8 crossovers.

You can do this by ear. You can put a mic in the car and do it with an RTA, use an SPL meter, or whatever you want. The other, a bit tougher, method to level match would be to use your drivers' specs, look at the sensivity and if they're all within a couple dB you could just level match at your amps' outputs (ie: set the output voltage to roughly the same). If they're not, then you'll have to do some calculations to give you approximate voltage needed to get xdB output per driver, then set your amps to that. It won't be dead on because the drivers' sensitivity will vary from mfg spec, but it should be close enough.

Another method, if you didn't want to do this at the setup would simply be to move your cd player's inputs (or any source inputs such as ipod) and hook up directly to each amp. Play pink noise, set the gain about where you want it, and then continue. If you do this, though, MAKE SURE that you're careful not to blow a tweeter since you're bypassing the ms-8 crossover. In fact, I'd really urge you to use the ms-8 to level match drivers. I mean, REALLY, encourage you. Not only for driver protection, but because it's just easier.

Hope that helps.

Tendean17 · Feb 20, 2011

Kkant :
When going active on the tweeters, on/off thumps are sent full-range to them. I've blown a few tweeters that way. If running them active, best to put a cap in series: a cap that is much lower than your xover freq, so it doesn't really affect the tweeter's active slope. Like say a 47 uF cap.

Andy Wehmeyer :
Yes, this is right. You shold ALWAYS protect your tweeters from damage with an additional cap. I've used 20uF, but 47uF will work too
______________________________________________________________

Andy Wehmeyer :
A passive network between midrange and tweeter is sufficient, but for midbass, i'd suggest an additional amplifier with an active filter. For this setup, I'd simply add a 4-channel amplifier and use half of it for the mid/tweeter and the other half for the midbass. If you're using a center channel, the time alignment for left and right is nearly inconsequential, since the left right and center steering nearly eliminates the need for the phantom center, which is what TA provides.

The amplifier channels are designed to drive 2 ohms, but theyt are not bridgeable

Certainly crossing the mids over lower than 1k is fine--the idea is to use the drivers in their piston ranges (where dispersion is wide). I also think that the rest of your system plan is good. So long as the 5" you plan to put in the center has useful output down to around 100Hz, it'll be great. The bass management system will take care of the rest.

As far as ventilation goes, if you'll use the 20 watt outputs, I suggest mounting this thing under the seat or somewhere that some air will be able to flow. The unit size is about what you've described--I don't have the info in front of me.
______________________________________________________________

cajunner :
you have to acknowledge that while the RTA is showing amplitude, there's likely some artifacts very near the same intensity but coming in at a small delay, such that by comparison with the original clock one might be able to determine what's reflection and what's decay....

Andy Wehmeyer :
This is a simple gated measurement compared with a non gated measurement. The gated measurement will give you the anechoic response (no reflections). The non gated measurement will give you the total. difference is reflections. This is how MLSSA and all the other analyzers that do this work. However, low frequency resolution depends on the size of the room. In a car, it wouldn't be possible to accurately differentiate between the two below the midrange. Equalizing above that point will be very effective in one tiny spot. None of this is necessary for good sound and is mostly an exercise in technology for technology's sake. Comb filtering isn't nearly as audible or objectionable as people think and that's what the phase dfferences between diect and reflected sound create.

Occasionally, if the stars are aligned, there can be some huge peaks caused by constructive interference, but without a head-tracker and a convolver, they are damn difficult to remove--moving the speaker is the best fix in those cases.
______________________________________________________________

SSSnake :
Care to elaborate on the 24 channels (seems like a LOT).

Andy Wehmeyer :
Sure--
3-way in the front doors...tweeter, 3" mid and dual-coil 6"
2-way center...tweeter and dual coil 6"
2-way in rear doors...tweeter and dual coil 6"
2-way in the rear deck...tweeter and 3"
2 10" woofers in the rear deck (IB).

Hmmm...that's 23 channels. I guess I have an extra unused channel somewhere.
Everything but the subs are driven by 20-watt ICs
______________________________________________________________

cajunner :
that's what I was thinking, pick a super high efficiency, ragged response driver and let the auto-eq fix that right up, no more worries about peaks and dips when the MS-8 has got your back...
no more choosing between lower sensitivity, higher cost product that is designed at high cost for smooth RTA, slap in a couple of Whip or Eminence 100db mids and you're sounding like Skaaning on the flip fo' show...

Andy Wehmeyer :
Of course, this is a valid method, and you'll likely see much more of this from the OEs as the popularity of electric cars and hybrids increase. It's much cheaper in terms of money and current to use high efficiency drivers than more powerful amps, but optimizing drivers for high efficiency always means reducing the useable bandwidth. You'll see more 3-ways and more small midrange drivers playing narrower bands of frequencies. All of this will be good for imaging too and will make cars sound better. The attendant help will come in the form of crossovers that keep drivers playing in ranges where their directivity is low (sound radiates in all directions).

Fortunately, a driver has to be really crappy to have a ragged response in the narrower passband where it's most efficient. If you choose this high efficiency route, MS-8 will take care of most of the ugly peaks and dips, but beware of the big peak that you see in the on axis response of all moving coil direct radiating loudspeakers (the kind that have cones and domes). That high frequency peak isn't attenuated as much as the "non-peaky" response off axis. That's makes EQ less effective, because removing the peak in the on-axis response will help to create a bigger hole off axis--basically, the reflected sound won't sound like the direct sound. In a car, that's a bigger problem than in rooms. Crossovers should be chosen to take that peak out of band. This is a big issue for a 6" 2-way, but no issue for a 3-way that includes a little midrange driver.

Wide bandwidth speakers are the OPPOSITE of the kind of driver that you'd need to implement this kind of design. Be prepared to cross woofers over a little higher and use low pass frequencies for the rest of your speakers that are a little lower than you're used to.

Hmmm...this sounds like fun. I think I need to find a guinea pig at work and build this car with the cheapest drivers I can find.

Tendean17 · Feb 20, 2011

kkant :
I didn't know that this was a fundamental property of all drivers. Could you elaborate? What causes it? And what characteristic of the driver is the frequency of this peak a function of? IOW, can I predict the frequency based on the diameter of the cone, for example?

Andy Wehmeyer :
Yeah, it is. It's a function of cone distortion and you can see it in the graphs above. These all appear to be really good drivers and someone has either paid careful attention to this or these measurements are made through a low pass filter.

The worst of the peaks and the one that doesn't go away off axis is caused by the center of the cone becoming decoupled from the rest of the cone. This property is exploited in the use of whizzer cones, where a small cone is attached to the center to shape and extend the response of that peak. The reason it doesn't go away off axis, is because the actual radiator that emits the sound is a small part (smaller diameter) of the cone that has much lower directivity than the whole cone. The stiffer the cone, the higher the frequency and the higher the Q of the peak. If the cone is more compliant (like a polypropylene cone) the lower the frequency of the peak and the lower the Q of the peak (wider peak). The size and weight of the coil and former affect this too. A heavy coil will cause the center of the speaker to decouple at a lower frequency and a larger diameter coil will increase the size of the secondary radiator. This is one of the reaons that sometimes you'll find really high end 6" midranges have small coils. It's to tame this high frequency problem. The trade off is thermal power handling. So, is a big coil better than a small one? Not necessarily--depends on the intended application.

This is why so many inexpensive car audio speakers use PP cones. It's easy to make the speaker sound good, especially in cars, because the floppy plastic cone has a low-Q rise at high frequencies. It's easy to align a tweeter with the high frequency response of the woofer for a generally rising response. In a car, where we hear the power response of the speaker (average of the output at all angles) that rising response sounds pretty flat and it only requires one crossover component--a cap on the tweeter.

Flat response requires that we cross the speaker over lower--before the peak, but that requires a bigger tweeter or a small midrange and more crossover components.

If any of you remember the Infinity Prelude MTS speaker (a home speaker), this is a really good example of one way to get flat response from super-stiff cones. That speaker uses a bunch of 5" woofers a 2-1/2" mid and a tweeter in a sort of expanded MTM arrrangement. The super stiff cones push the peak to high frequencies and the steep crossover slopes eliminate it from the response. The result is really flat response and near perfect directivity index for rooms, but it's expensive because the crossovers have many components, the cones are ridiculously expensive and the additional midrange is required. The marketers decided that we had to use those same cones in car audio drivers but no matter how we tried, we couldn't make it work very well for big 2-way speakers. That series of Kappa speakers didn't sound great and weren't big sellers--the marketers learned a lesson.

There are many things that can be done to tame this peak and a stiff cone isn't always the best cone. Many pro-sound drivers have concentric rings molded into the cone (see older JBL pro speakers). Those rings are designed to create compliance in the cones to cause the cone to distort in a way that smooths the response at high frequencies--important when you'll use a 15" speaker to make 5-or 600 Hz. This can also be accomplished by varying cone thickness along the profile.

Exponential curves for speaker profiles are also designed to tame this peak, when compared to flat-sided cone-shaped cones. Dustcap materials and shapes can help to tame the peak also.

Woven cones tend to distort according to the pattern of the weave, and adjusting the weave or the resin used to stiffen the fabric can help to flatten response too. Incidentally, that resin (type and amount) along with the material and the coarseness of the weave is what is used to adjust the compliance of the speaker's spider.

All of this is why simple cone-material stories are often a bunch of marketing hooey, especially when they just indicate that the cone material is stiff. Polypropylene isn't stiff. It's the opposite of stiff and it's chosen specifically for that property (also because it can be made to last a long time in adverse heat and humidity conditions). Laminates can be made to be stiff, but something can also be laminated in between the outside layers to make the cone stuffer, alter the profile or provide damping for resonance. If you read about the bodies of many high-end cars--I think Lexus does this--the body is sometimes a laminate of two layers of thin metal with some damping material in between. This is designed to eliminate resonance and quiet the cabin.

Finally, a laser interferometer can be really helpful in designing good speakers. We use one. Basically, the designer can take a prototype into the chamber and make a simple frequency response measurement. That will show peaks and dips. Then, he takes the speaker to the interferometer, excites it at the problem frequency and the machine makes a movie of the movement of the cone at that frequency as a wireframe and a heat map. Then, appropriate measures can be taken to stiffen or weaken the cone to eliminate the peak or dip. I have some of these movies from the development of the 660GTi speakers along with some made for some really high-end competing speakers. The movies are huge and I have no idea how I could post them so you guys could see them. The interferometer works on boxes too. You can watch the box flex and add a brace at the appropriate spot to stop the flexing.

So the moral of the story here is that a light cone driven by a strong motor won't make much bass, but it will be efficient. If that cone is really stiff, it may have flat response in the passband, but it may be ragged at the top of its frequency range. If the cone isn't so stiff, it may have peaks and dips lower in the range of frequencies it plays, but the top end may be smoother.

If you want to find out the quality of a driver and determine whether it's the right one for your system, look at the frequency response curve. It's the only way to tell. Reading about materials in marketing documents isn't helpful (and, in fact, it can be seriously misleading), unless you know what those materials may require in additional stuff. A speaker designer can look at a driver and tell in a few minutes whether the appropriate measures have been taken to provide flat response, but even he won't be able to tell whether the measure have worked without a response curve.

I know, I know...ask for a quarter, get a dollar...

Tendean17 · Feb 20, 2011

DS-21 :
Audyssey claims, though, that their taps don't have a linear distribution. According to their website, the MultEQ XT has 8x the resolution in the bass than above the bass.
What that means, honestly, no idea here. I'll defer to you guys.

Andy Wehmeyer :
I'm not a DSP engineer, so I'm going to keep this on a really basic level so I don't get Lycan's hot poker of truth in the a$$ again.

There is a fairly recent development that makes it pretty easy to divide those taps across frequency bands, perform separate convolutions and then connect the bands together again. If I'm not mistaken, it's often referred to as "Fast Convolution". We're working on another processor that would use this technique for an even simpler device. There's nothing inherently wrong with using multitap FIRs, so long as compromises are managed in a way that keeps cost and latency in line. To be blunt about it, the method doesn't matter so much as the result. For all DSP, the algorithm is much more important than the device on which it runs and on which tools are used by the algorithm. The IC is the hardware, the filters are the tools and the set of rules (the algorithm) by which the device decides how those tools are used are the real meat of any DSP-based solution. The best way to determine the appropriateness of the solution is to analyze the result of using it. If the result falls short, then an analysis of the algorithm, the the tools and the choice of hardware can be useful in designing something else or changing the design.

The day that someone tells me that the pins on a TI 7xx have to be gold plated in order for the product to be considered high end might be the day that I look for work in another industry.

Hmmm...maybe I should be careful about what I write...
______________________________________________________________

Andy Wehmeyer :
No fuss created. Thanks for the clarification. I'm happy that someone can provide better info when I'm wrong. Sometimes the long tail of information that never goes away is unfortunate because we can never be smarter than the stupidest thing we've ever written...
______________________________________________________________

rain27 :
Andy, What is your opinion on the lowest frequency a tweeter should play at when installed far from the midbass in a 2way set up? There is some debate that a tweeter should not be played lower than 4k if the mid is low in the door and the tweeter is in the sail panel, a pillar, or dash.
When setting the xover on the MS-8, I'd like to what limits to be mindful of.

Andy Wehmeyer :
Depends on how big the midbass driver is. If it's a 6", then 2kHz or even lower is helpful in trying to keep the low pass for the midrange as low as possible so it won't be used where the dispersion is narrow. That'll give you better similarity between the on-axis and off axis frequency response. If it's a standard 1" tweeter driven by lots of power, then that's not a wise choice because you'll probably blow it up. Choosing the crossover is a matter of driver safety first, directivity and a match to the next smallest driver second, and trying to shape the response third. IF you have an EQ, then that's the better tool for response shaping if moving the crossover around compromises the first to criteria.

I think the winner of the debate mentioned above should be the person who has most vehemently argued against the suggestion.
______________________________________________________________

Andy Wehmeyer :
When you use MS-8, it may be helpful to adjust your perception of how best to use a crossover. In conventional systems, we often use the crossover to get the smoothest response possible before applying EQ. If you have an RTA, this is probably the first part of the tuning process. The most important part of choosing the crossover is to make sure that you set high pass filters so that the drivers are safe from too much bass. Then, it's a matter of choosing the precise frequencies and slopes to get the best blend between the speakers according to the microphone while setting the levels of the adjacent bands.

With MS-8, you don't have to do all of that. Just set the crossovers for steep slopes to minimize driver interaction and choose frequncies that keep the drivers safe.

For example, a couple of days ago we put one in a 2000 BMW 3-series with the old amp that has analog filters built into the amp. All we did was add a simple capacitor to the tweeters in the front and rear and connected the tweeters and small mids to the MS-8. We used separate channels for the front midbass drivers and a pair of chanels for the subs. Worked great. No need for some complicated passives between the mid and the tweeter.

MS-8 won't fix a stupid system, though. If you use an 8" midbass in the door and a 1" tweeter in the top of the door, it won't fix the problem you'll have at the crossover point due to directivity issues, but neither will a huge collection of tunable filters or overlapping crossovers, no matter how cool the PC interface may be. Bottom line is if you build a competent system, MS-8 will make it sound better than 99 percent of the other processors available whether they have an autotune or whether you or a real acoustician adjust the EQ.
______________________________________________________________

michaelsil1 :
I have another question.
I read in the manual that JBL highly recommends using 24db slopes would I run into any issues using (Phasing, SQ) 12db slopes?

AdamS :
I'll give you the beginning of the answer and then you can tell me if your eyes are glazed over or if you want to know more...
LR or Butterworth filters in the acoustic domain don't add well for 2nd order. In fact, they cause a null. This is easy to fix. Either the lowpass or the highpass signal needs to be flipped.
So let's suppose you have L/R (2-way), cutoff of 80Hz, 1K
and a center cutoff of 200Hz, 1-way
Then for L/R....
0-80 negative polarity
80 - 1K positive polarity
1K and up negative polarity
This in and of itself would be fine.
But now add in your center.
0-200 negative polarity
200 and up positive polarity
Now let's map these together.
0-80 L and R negative, C negative
80 to 200, L and R positive, C negative (this isn't good)
200 to 1K, L and R positive, C positive
1K and up all positive
So basically in this example from 80 to 200 you get out of phase stuff.
If you have a 2 way system or if your speakers are all the same, this will be OK. We take the Lowest L/R driver as the positive reference and flip the rest as needed to avoid nulls.
Lesson here, unless you really have a need for 2nd order, try not to do it. It's the only way I really know how to make things sound poorly, with or without MS8.

Tendean17 · Feb 20, 2011

Bikinpunk :
Alright, I did some more driver swapping tonight in my buddie’s car (he’s a member here so he can chime in if he sees this).

We tried wavecor tweeters, his PPI tweeters, and a set of h-audio trinity’s running full range from 300hz +. We swapped the drivers out, and tried them in different angles, re-running the calibration each time.
Every time, there was a different end result. With the PPI’s, the system was just too harsh and the imaging was crap. The trinities were aimed on-axis for their testing and sounded great. Imaging was much better and center was dead center. The wavecors were crossed at 3000hz and sounded great as well. I won’t get into the things we noticed. The point is that each change of the system resulted in a different system response. I’m not talking ‘critical listening’ differences; I’m talking driver characteristic differences and glaring response differences. Simply swapping tweeters and re-running the setup resulted in a better center image.

What I’m trying to say here is simple:
Installation does still play a role in the final output even with the ms-8. Something I think many people who are using this unit may not realize. (ie: I hooked it up to my system and it didn't sound awesome like you said it would). The ms8 will not, and does not make every system “sound the same”. I thought it would. I experimented and found that to not be the case. It’s up to you to realize your car’s full potential: don’t rely on a processor to do it all.

It only goes to show that install plays a huge role, no matter what processor you use. The ms8 simply makes it easier on you to hear the differences. You can’t have a shoddy install and expect the ms8 to be the only answer to fixing your problems. You’ll still need to play around with angle and location to see what works best.

Take advantage of the ms8: use it to experiment. Try things.
This whole idea of “yea, but you can’t tinker anymore” is complete BS. You can still tinker. I’d argue that it’s now much easier for someone to tinker and realize how the differences in install and drivers affect your system’s response. Because of the ms8, I’ve been able to have more fun with trying new drivers. Back with my old processors, it was much harder because you really had to re-tune everything. I can’t guarantee that the ms8 is going to give me the optimal tune everytime, but I think it sure takes a lot of the guesswork out. So much so, that I’m now switching things, running the setup, listening, taking notes, and switching again.

To the people who own the ms8: realize and harness its potential! You have a great tool at your disposal. Don’t be complacent… take advantage of this toy to learn and experiment.

- Erin
______________________________________________________________

Andy Wehmeyer :
OF course it can't fix everything. However, if you're finding huge differences when you switch out drivers, then the differences are most likely due to frequency response differences between drivers and sensitivity differences, which may cause you to adjust the levels of the sweeps. If the "image is crap", then the sweeps are too loud. Even the crappiest speakers should provide a decent image, so long as they cover most of the audible bandwidth. The differences between one driver and another shouldn't be big differences in frequency response, as this is what MS-8 fixes. Non-linear distortion, which includes bottoming, odd order distortion caused by coil out of the gap, and even order distortion caused by suspension non-linearities or dramatically different flux profiles above and below the gap or other stuff may be audible and will be the difference betweenone speaker and another. IF the frequency responses of the speakers vary so much that the differences are outside of the 40-or-so dB that MS-8 has to use in tuning, then those differences mnay be audible as well. VERY high-Q response anomalies may not be EQed either. So, yes, some differences will exist. What makes MS-8 cool is that it eliminates many of the inconsequential differences.

Erin, I didn't misunderstand your post. You're right on the money. Like you, I just don't want people thinking, "well, MS-8 didn't fixmy problem, so it must be defective" if the problem is one that can't be fixed by EQ and level matching.
______________________________________________________________

Andy Wehmeyer :
Yeah, you can't just throw some drivers in the car in a shoe box or two, hook up MS-8 and expect to get great results. One other thing to remember, especially for DIYers, is that speaker sensitivity makes a big difference in where input sensitivity controls have to be set. I keep writing that setting sensitivity to about 2V onall the amps is righ, but if you choose some midbass that's 93dB at a watt and a meter, a midrange with a giant peak at 2k with a sensitivity rating of 90dB and a tweeter with sensitivity of 82dB, then the gains will have to be set appropriately.

The car contributes more to the system's frequency response than the speakers do, and the car doesn't change when you replace drivers. Basically, MS-8 has to scale everything to set levels. If your midrange has a 20dB peak at 2k and that's the tallest peak, everything else will be scaled as a result. The window is big, but it's an algorithm. It makes decisions based on a bunch of rules and those rules are written assuming a reasonable system. So long as the system isn't a basket case, MS-8 works great. If your system is a basket case, then it won't. The difference between a basket case with an autotune and a basket case with a toolbox and a tuner, is that the tuner can look at some huge problem and go after a fix that isn't EQ-related, like choosing a different driver, a different location for the driver or a differnt crossover frequency. Of course, MS-8 can't do that. IF it could, we might have named it "Midas".

IF you're starting from scratch, just build a simple system and put the speakers in the stock locations, especially if you'll have a center channel. If you already have some custom thing going on, you may have to make some adjustments in the way you've managed compromises in the past to get great results with MS-8. Things like line drivers, additional EQs in the signal path, fiber optic signals, mono rear fill, 4-way all-active systems and balanced line connections are all things that may raise compatibility issues. That doesn't make them worng or make MS-8 wrong, it just means they may not work well together.

I always try to be straight with you guys and one thing I can say for sure is that many of these things that are added as "features" to traditional gear often include outlandish claims of performance enhancement or some other benefit in the interest of distinguishing just another amp from other just-another-amps or just-another-speakers. Most of it has some technical merit, but the merit may never be experienced because the benefit appears outside the audio band, because the benefit overcomes a technical problem that isn't audible (and certainly isn't annoying), or because the benefit is BS but can't be adequately disproven because it's designed to appeal to one's emotions or to dupe you into believing that there is merit by making the signal louder and claiming to make it "better", "more transparent" or my favorite: "effortless". Anthropormorphization is damn difficult to overcome once people are bought in to the hogwash.

Tendean17 · Feb 20, 2011

The Target Curve for JBL MS-8

Andy Wehmeyer :
The target curve for MS-8 is remarkably similar to the frequency response of a speaker with an ideal directivity index for listening in small rooms. So, if that means that it's designed for a reflective environment, then so be it. Horns don't eliminate reflections, but they do eliminate SOME high frequency reflections from boundariesadjacent to the horn's mouth.

MS-8 is definitely designed to flatten the midrange and the transition from bass to midrange is designed to locate the bass in the front of the car. If you like tons of midbass, you can always put that back with the 3rd octave EQ.

One thing is for certain, if the sweeps are too loud, what you've described as no midbass will certainly be the case. If the first set of sweeps are too loud, there won't be a center image. The first set of sweeps set the channel delays. If the second set of sweeps are too loud, the system won't EQ properly. So, if you have a good center image, that indicates that the first set of sweeps (without the subwoofer) were at an appropriate level. If you have a good center image and a bad EQ, then the subwoofer sweep is too loud. This is a common mistake, because the sweeps shouldn't be loud at all.

The attack that you like so much from horns is, in fact, usually a result of big, high-Q peaks in the midrange and better dynamic performance from the compression driver. There's nothing in MS-8 that will diminish the dynamic range--there's no compression used in the algorithm. MS-8 will tame those peaks, but depending on thier distribution, it may group some of them together and tame them with a single lower-Q filter. This may be what you don't like. Additionally, if you're not using additional tweeters with the horns, MS-8 won't provide the huge boost at high frequencies that is necessary to have highs in the car without additional tweeters.

MS-8 was definitely NOT designed with horns in mind, since the vast majority of cars don't use them. Many horn implementations require a different solution to the standard compromises that are part of building good sounding cars. Depending on your setup and yor tuning abilities, a horn-based system may be better served by a manual EQ.
______________________________________________________________

Andy Wehmeyer :
The additional low pass filter at 90Hz that you use is a workaround for the fact that increasing the sub level increases the acoustic crossover frequency between the subs and the midbass. You've discovered this through hours of listening and tinkering. It's a one-time fix that you've implemented as a compromise. The shelf filter is also proprietary, and only available in MS-8 (although someone else could write the math to make it happen in another processor), and it boosts below 60Hz but NEVER above 160. The slope of the shelf is adjusted as the level is adjusted. It's similar in performance to your additional filter, if the filter were adjusted automatically when you adjusted the level to maintain your crossover frequency and overall response shape

The target response is fixed. The autotune EQs to match the target. If you don't like the target, you adjust the sound using the 31-band EQ. It's not exactly done the way Mike explained it, but the end result is the same.

There's no need to plug in some high-end microphone. The reason that high end microphones still exist for measurement (and I'd be willing to bet they won't exist for much longer) is that before computerized test equipment, the accuracy was only as good as the flatness and bandwidth of the microphone. Now that analysis is done as software, the hardware isn't so much of an issue, provided the flaws are linear. The microphones that come with MS-8 are good panasonic-type electrets, which are the ones that are used in many other measurement microphones. MS-8 includes a mic correction filter, that makes them measure flat, despite their being mounted in a pair of headphones cases. Plugging in some high end mic is unnecessary and would result in poorer performance. The only advantage might be the ability to make the measurements at ridiculously high levels, but that's just not necessary.

Yes, you can tweak the crossovers if you want, but you MUST run the EQ again after making those adjustments. Fortunately, that only takes a few minutes. You'll find that there's less need to set preposterously low crossovers for midbass, midrange and tweeters with MS-8.

I'll provide an ETA when the rest of the debugging is underway and when I have a date that we're sure not to miss

The EQ and the rest of the processing in MS-8 can't fix driver distortion and fixing non-linear distortion with DSP is still very uncommon in consumer devices (although many folks are working on it). Certainly, great Eq CAN make the difference between a good speaker and a great speaker much more difficult to hear, but there will always be room for great speakers--especially if they are to be driven with a lot of power. In my estimation, that's the first and best reason to replace speakers--so you can add more power. The opportunity for substantial improvement in most playback systems by replacing drivers with like drivers is coming to an end. So long as the system plays loudly enough, replacing drivers will soon be like using high-end wire...

BTW, I've done a Richard-Clark-Amp-Challenge thing with speakers in our sound room. The very small difference in the sound (similar sized woofers and tweeters, mounted on a flat baffle in the same way) once the speakers were EQed so they had the same response, wasn't easily perceived my most listeners--until we played them loudly.
______________________________________________________________

Tendean17 :
I think JBL MS-8 use the same target like this ..
http://www.diymobileaudio.com/forum/diyma-sq-forum-technical-advanced/11579-flat-response.html

Andy Wehmeyer said:
OK. Sorry for the delay. As a matter of review and to redirect this thread a bit, I've pulled the following from an earlier post:

So, there are a couple of basic ways around this reflection thing in cars. One is to try to minimize them and the other is to "spread the chaos around". Minimizing them is damn difficult if you want to have a car that can be driven. Spreading the chaos around simply means that we try to make all the reflections have similar frequency response so we can EQ them and the sound of the speakers using the same filters. Using the speakers where the dispersion is wide is the way to do this.

Using horns and big diaphragms for "pattern control" is the way to minimize the reflections from adjacent boundaries, but those adjacent boundaries are responsible mainly for altering the frequency response and not the image. Image narrowing and causing the car to sound like a car is mostly caused by reflections of right channel information from boundaries near our left ear and reflections of left channel information from surfaces on our right. This crosstalk partially destroys stereo, as each of our ears hears both speakers and the arrival times are within 20mS. This is worse in cars than in rooms, because the surfaces are closer. That means the intensity (loudness) of the reflections is greater and our perception that they and the initial sound are the same is greater. This is what makes a car sound like a car rather than a much larger space. This condition is what "Ambisonics" tries to fix.

I've tried both of these routes to great sound and I find that the "spread the chaos around" method is the most straightforward and because the conditions in the car make this much easier. The cheesy drawing below makes this a little more obvious, if the car is compared to a room.

The first rule of equalization is that you can't put energy back into a null caused by destructive interference in the acoustic system. You can remove energy from a peak caused by constructve interference.

The second rule is that high-Q (narrow) peaks and dips are less audible than low-Q peaks and dips. When we listen to music, our ability to perceive high-Q peaks and dips is further decreased because the chances that a particular sound will happen at precisely the right frequency, thereby energizing the high-Q peak or dip is reduced. In essence, this means that there may be no need to smooth the high frequency part of the comb, but the low frequency part of the comb, particularly the first (lowest-frequency null) is a problem. Plus, many listeners like the way the top of the comb sounds, because it adds a sense of width by screwing up the phase between the speakers and drawing our attention to the locations. This is very hard to hear in the frequencies that tweeters play, but a bit easier, lower in the midrange.

Patrick's method is different than my method and both work. Patrick's goal is to control both frequency response and directivity (through narrow passbands and pattern control--(horns) and using speakers in the range where they beam) and the method I suggest is the opposite. The reason that they both work is because, as I've suggested before, we don't hear the reflections as separate events. Patrick attempts to eliminate adjacent boundary reflections, and I suggest simply EQing them along with the direct sound.

It's important to note, that neither method will eliminate the reflections from boundaries close to our ears, and those reflections are the ones that make a car sound small--those reflections contribute crosstalk (left sounds at our right ear and right sounds at our left ear).

OK, so if you're following me, the basics are that we have to have a right and left speaker system that covers the band of frequencies from the crossover point between them and the sub and 20kHz without huge holes in either the on-axis response or the off-axis response. Of course, the off axis response will exhibit gradual attenuation as frequency increases, but it shouldn't have big peaks or dips near any crossover point. Below is a nearly ideal frequency response and directivity index:

Obviously, the red curve is slightly better than the blue one. For the speaker in the graph above, we use a waveguide on the tweeter to match the directivity of the tweeter to that of the woofer at the crossover. A similar curve (a directivity index that's reasonably flat with no peaks or dips) can be achieved by using multiple drivers in the regions where they DON'T beam. This is why I suggest a 3-way system in the front doors.

The frequency response plot below is one for a 2-way component system that doesn't include any measures to match directivity between the drivers. The crossover point between the mid and tweeter is too high (the mid starts to beam off axis). There's no directivity index plotted here, but it would include a peak in the region where the off-axis response shows a dip.

OK, so now that we've placed a great deal of importance on getting the speaker system right to begin with, here's an example of why that's all good system design practice, but in the end, doesn't matter nearly as much as it might seem. It is important to be sure that the entire bandwidth is covered competently, because we can't put energy into a system if there's no source for that energy--if your speaker system has a big hole at the crossover, you can't get it back.

Below is the near-field respone of the OE speaker in the door of a Mini Cooper. The measurement was taken about a half inch form the grille. Despite the dip at 1k, the response isn't too bad. BTW, I didn't investigate the cause of the dip

If I were building this system, I'd add a 3" mid and cross the 6" at 1kHz. The OE system doesn't include a low pass on the 6" and includes a small tweeter in the top of the door. Anyway, if I close the door and place the microphone at the driver's listening position, this is what the response looks like:

Nice, huh? What's the difference? Reflections and a tweeter. Good luck trying to eliminate all of that--no matter what you do, you'll never get back to the previous graph. The stuff below 1kHz can't be eliminated by anything other than dramatically changing the shape of the interior surfaces and the size of the vehicle. Between 1kHz and about 10kHz, adjacent boundaries contribute to the mess. Above about 10kHz, adding foam and carpet here and there might tame this somewhat, but the benefit isn't worth the effort.

Fortunately (or unfortunately) the response above is what we hear and it can be adequately equalized. Below is the target response I suggest:

the big problems at lower frequencies have to be fixed. It won't be possible to put energy back into the biggest hole, so the peaks will have to be reduced to match.

At this point, looking at a lower-resolution graph may be helpful--we don't want to be distracted by all of that high-Q stuff at high frequencies.

. Here are the steps, including selection of a high pass filter frequency:

Of course, this can't be done with a 31-band graphic EQ and that's why those aren't useful. It also can't be done by moving speakers, changing the aiming, swapping the speaker for something better, adjusting the crossover point between the mid and tweeter slightly, adding damping material to the doors, or anything else. EQ is the only thing that will help.

This is why I suggest the method I suggest.

TA should be set first, especially if you can align the tweeter and mid separately. After TA, set your EQ according to the process above. Once the left and right have been EQed so thay have the same response and the arrival times match, add the sub. EQ the right with the sub. Then turn the right off and EQ the left with the sub. Then, turn on both channels and listen. If you need to make further adjustments, be sure to EQ right and left equally--you're just shaping the response now and you can't equalize frequency reponse problems that appear in the measured response of both channels, measured at the same time (see the lengthy discussion of comb filtering in the previous post). In fact, what you see on the RTA won't even look like what you hear, so don't pay much attention to the RTA of both channels playing at the same time for any frequency above about 150Hz (the region where the sub contributes).

Also, moving the microophone around a head-sized area will help you determine whether to fix the high-Q peaks at higher frequencies. If they change or go away as you move the mic, don't worry too much about them unless they annoy you when you listen to music. Then, pick the annoying one and get rid of it.

BTW, we have a new 2-channel auto EQ algorithm that does precisely what I've suggested and it makes even this Mini Cooper with the factory speakers sound GREAT.

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
Oh, and for wat it's worth, the amount you turn your head will affect the high frequency EQ slightly. If you turn your head really far, there will be more high frequency masking (indirect path to the opposite ear and more off axis to the ear closest to the speaker being measured). There will be NO difference between the tuning below 1k. It's really not that sensitive to the head turning thing.

The most important things during calibration are the right crossover between bass and midrange and the levels of the sweeps.
______________________________________________________________

Trigg007 :
I'm not overly concerned with response the curve options. However, what you claim is not clearly stated in the manual. It'd be nice to confirm though...
My primary concern is about the crossover options. What types (Butterworth/Linkwitz), available slopes & crossover range/setting options. I spent 20 or so minutes on the phone with JBL tech support in an attempt to have my questions answered to no avail. All they could offer was wording similar to the manual. No disrespect to JBL, but I'd like to know before making my purchase decision. All the tech person could offer was a "call back in 2 weeks". Maybe Andy can "chime" in

AdamS :
1st order: Butterworth
2nd order: LR
3rd order: Butterworth
4th order: LR

LR doesn't exist for 1st and 3rd as it is a squared Butterworth.
Crossover frequency resolution is in resolution of Hz.
if you have a sub
subsonic : 20-50 Hz
sub/front: 20-200 Hz
if you don't have a sub
sub/front: 20-100 Hz
others
Front Lo/Mid: 100Hz-10Khz
Front Mid/Hi: 200Hz-10KHz
Center highpass: 50Hz-10KHz
Center Lo/Hi: 50Hz-10KHz
Side highpass: 50Hz-10KHz
Rear highpass: 50Hz-10KHz

There are restrictions... ie complementary lowpass and highpass must be at the same frequency. You can't equalize the system with crossovers by forcing bumps or dips. Use the 31-band for that.
______________________________________________________________

trigg007 :
great...what about a 3-way set-up?

AdamS :
It's included. Here's an example. Suppose you have a left (3-way), right (3-way) and sub.
subsonic : 20-50 Hz
sub/front: 20-200 Hz (between sub and front)
Front Lo/Mid: 100Hz-10Khz (between low and mid)
Front Mid/Hi: 200Hz-10KHz (between mid and high)
Visually (from a low frequency to high frequency perspective):
-Sub-Low-Mid-High
There are crossovers at each of the -.

Tendean17 · Feb 20, 2011

trigg007 :
I hope you're right; can anyone confirm?
Also, is there anything else available (literature) with more detail in what you can/can't do with the MS8. I've read both the MS8 manual and the Bit1.1 manual. To me MS8 seems like "set it & forget it" versus the BIT.1 being more of a nightmare to tune for those of us who just want to enjoy the music. Don't get me wrong, according to the manuals, the Bit1.1 appears to have substantially more processing power, or atleast the manual is considerably more informative (IMO). What are the crossover options? Can the time correction be tweaked/changed versus the autotune feature, etc...

Andy Wehmeyer :

Andy Wehmeyer said:
Yes, you can't change the target function but you can retune with a second 31-band EQ. Since the channels will already have been matched and tuned separately with the auto EQ, there's no need for separate right and left channel EQs for retuning. You can simply draw the adjustments you want to make and those adjustments will be applied equally to all channels. It works great.

I just remembered I have some documentation on the EQ 31-band EQ that you may find interesting.

Most graphic EQs don't give you what you may think you're getting. depending on the Q values that are chosen for the filters, the sum of adjacent band adjustments may be flat or they may have little dips in between the boosts. Many of them have qs that are wide enough not to produce those little peaks and valleys, but then the sum of the adjacent filters produces far more boost or cut that you might expect. Here's an example below.

In this EQ, boosting one band provides an expected response:

Boosting 2 adjacent bands deviates a little from theexpected response--if you look closely, you'll see that there's actually more boost than 6dB.

If we continue boosting adjacent bands, this is what you get:

In this example, boosting a bunch of bands gives you twice the boost when the results are compared to the slider controls.

In the example below, clearly what you get is not what you may have intended:

So, when you choose an EQ and try to use it, it's helpful to know something about the design. This detail isn't always provided and it can make tuning a real pain in the butt. A graphic EQ can be designed with higher Q values to make the adjustment bands narrower, but there's always some error.

MS-8's 31-band graphic works differently. When you adjust bands, the EQ adjusts the adjacent bands automatically to match the response to your intended response. In the three examples below, the little green circles represent the slider controls. the red line represents the output of the EQ and the automatic adjustment of the adjacent bands is represented in the top graphs.

The three examples below are response shapes that are very difficult, if not impossible, to make with any graphic EQ I've ever used--digital or analog. I can't even imagine the amount of time it might take to implement these with an analog EQ.

The wiggly red line at the bottom is just noise. Doesn't reeally appear in the output.Granted, there's some wiggle at the top, but it probably won't be audible.

In this one, there's a little wiggle between 200 and 500 Hz, but it's less than 1dB.

So...with the autotune that "fixes" the car, and this EQ which is a pretty accurate drawing tool, you ought to be able to make the car sound great easily without a bunch of hassles.

Tendean17 · Feb 20, 2011

AdamS :
Andy's data is a little old. We reduced the minor ripple > ~12 KHz even further.
http://www.diymobileaudio.com/forum/1044261-post2659.html
______________________________________________________________

quality_sound :
Exactly. The MS-8 is going to his it's target curve every time. It might change slightly, but it's better to run it once and then draw in YOUR target curve and let the MS-8 make it's changes and be done with it. With the MS-8 all you really need to know is the final curve you want.

Andy Wehmeyer :
Right. The curve that MS-8 hits is one I've used for 25 years to make most customers happy. the beauty of the deal is that the hard work of eliminating peaks and dips, getting TA right and balancing the crossover vs. EQ is done for you. Then, all you have to do is fiddle with an easy tool designed to make fine tuning easy, rather than using that tool (31-band) do do the hard stuff. 31-bands isn't sufficient for the hard stuff.
______________________________________________________________

Andy Wehmeyer :
There are two ways to eliminate the peaks and dips that are caused by reflections--comb filtering--from the measurement: spatial averaging and spectral averaging. We use both. The measurements are averaged spatially and then the response is averaged spectrally when the EQ filters are applied. The differences between the different head angles are small, but if you turn your head so far that both speakers (weeters mostly) are far off axis in measurements 3 and 4, you'll have more high frequency in the resulting curve. A big difference in the placement of the mic for all 4 measurements (tall person vs. very short person) will result in different listening experiences. We remove comb filtering from the measurement because we don't hear it and EQing to remove the high frequency part of the comb is unnecessary and results in bad sound.

Through the course of developing MS-8, some of the DSP guys have developed a spectral average that performs almost as well as the spatial average (95% or so), and that alogrithm will appear in a much simpler product one of these days. It only requires a single measurement and is designed for a very specific (but very popular) use case.
______________________________________________________________

Andy Wehmeyer :
The auto EQ is separate left and right for front, sides and rear, separate for center and sub also. Then, the 31-band changes the curve for all, since they're already matched.
______________________________________________________________

cajunner :
something I'm still fuzzy on, is what frequencies are targeted for manipulation by the auto tune, and whether the addition and subtraction of magnitude-related filters, is also then treated to time alignment for proper phase, or vice versa. I believe this is how MS-8 does it, but the Pioneer might not, relying on filters only?

also, if you use an equalizer function that picks more frequencies in the bottom and middle of the spectrum, (you've only got 500 or so bands of adjustment, so out of 20,000 different frequencies...) will it lead to better control of the frequencies that control width(ITD), and leave the height manipulation(IID) underserved?

so, depending on whether your install needs help more in one area than another due to relative speaker placement, one might be more useful than the other?

Andy Wehmeyer :
MS-8 uses 8 filters that can be any shape (not bound by the typical frequency, magnitude and Q values common to standard parametric EQ) and the entire frequency range is divided into 8 sections. This is applied to a spatially averaged and smoothed curve. Rather than the resolution of the EQ being defined by the number of filters, it's defined, basically, by the spatial averaging and the smoothing. For typical FIR (512, 1024, etc) multitap filters, the resolution is defined by the number of taps (512, 1024, etc) While that might sound like really high resolution, the taps are often distributed over the frequency range in a linear fashion rather that logarithmically. That means that half of the filters are applied between 10K and 20K and half below. That's a problem because a 512 tap filter only provides 40Hz resolution and you need better resolution in the bass and midbass. Without a spatial or spectral average, however, the FIR filter just fills in dips and attenuates peaks. This is great for speaker EQ, but not so great for room EQ in a really reflective environment, EQs that operate this way often attempt to fill in high-Q nulls caused by reflections. This is a no-no and it's one of the reasons that building a great room EQ isn't as easy as just adding resolution to a standard multi-tap arrangement.

There are tricky ways around this and the tricks all have different names. One way is to divide the spectrum into several bands and apply FIR filters of different lengths to the various sections. A low-power 256 tap filter applied to frequencies below 200Hz would give better than 1Hz resolution for the bass, for example. Then other filter lengths could be used for other ranges and then the signal could be combined before the output.

I have no idea what method Pioneer uses for their autotune. The Audyssey algorithm has some form of redistribution of the taps for better resolution at the low frequencies. The quality of the EQ's performance isn't necessarily related to the KIND of filter that's used. Rather, it's related to HOW the filters are used. All have their benefits and drawbacks

Tendean17 · Feb 20, 2011

Phase and Frequency Response

Andy Wehmeyer :
OK, the sound we hear in cars and rooms includes reflections, and those reflections also create comb filtering,
which change the frequency response and change the image, just like the two independent speakers in the screenshots above. We can't time align reflections to get rid of the constructive and destructive interference. For the rest of this, remember that reflections from nearby surfaces have short PLDs and from far-away surfaces have longer PLDs when compared to the initial sound fro the speakers which cuts a direct path to our ears.

For an even more detailed picture of what happens, it's important to consider speaker directivity. Adjacent boundaries will produce reflections for speakers at frequencies where the dispersion is wide, but not where the dispersion is narrow. The angle of incidence equals the angle of reflection, and that's anpther thing to take into account when considering how all of this interacts. We don't have to be concerned with reflections until the sum of all of them reaches our ears, because that's the only place where we hear their sum.

So, there are a couple of basic ways around this reflection thing in cars. One is to try to minimize them and the other is to "spread the chaos around". Minimizing them is damn difficult if you want to have a car that can be driven. Spreading the chaos around simply means that we try to make all the reflections have similar frequency response so we can EQ them and the sound of the speakers using the same filters. Using the speakers where the dispersion is wide is the way to do this.

Using horns and big diaphragms for "pattern control" is the way to minimize the reflections from adjacent boundaries, but those adjacent boundaries are responsible mainly for altering the frequency response and not the image. Image narrowing and causing the car to sound like a car is mostly caused by reflections of right channel information from boundaries near our left ear and reflections of left channel information from surfaces on our right. This crosstalk partially destroys stereo, as each of our ears hears both speakers and the arrival times are within 20mS. This is worse in cars than in rooms, because the surfaces are closer. That means the intensity (loudness) of the reflections is greater and our perception that they and the initial sound are the same is greater. This is what makes a car sound like a car rather than a much larger space. This condition is what "Ambisonics" tries to fix.

I've tried both of these routes to great sound and I find that the "spread the chaos around" method is the most straightforward and because the conditions in the car make this much easier. The cheesy drawing below makes this a little more obvious, if the car is compared to a room.

The first rule of equalization is that you can't put energy back into a null caused by destructive interference in the acoustic system. You can remove energy from a peak caused by constructve interference.

The second rule is that high-Q (narrow) peaks and dips are less audible than low-Q peaks and dips. When we listen to music, our ability to perceive high-Q peaks and dips is further decreased because the chances that a particular sound will happen at precisely the right frequency, thereby energizing the high-Q peak or dip is reduced. In essence, this means that there may be no need to smooth the high frequency part of the comb, but the low frequency part of the comb, particularly the first (lowest-frequency null) is a problem. Plus, many listeners like the way the top of the comb sounds, because it adds a sense of width by screwing up the phase between the speakers and drawing our attention to the locations. This is very hard to hear in the frequencies that tweeters play, but a bit easier, lower in the midrange.

FWIW, this delay between the channels is used by music producers to provide a sense of spaciousness or a separation of the sound in right and left channels.

I use impulse response measurements for setting delays. It's important to turn off low pass filters during the measurements, if you do it that way. No need to time align to the sub.

I usually use 4th order slopes to minimize interaction between the drivers, which are seldom mounted close to each other and to protect them from too much bass.
______________________________________________________________

Andy Wehmeyer :
The out of phase information will steer rear. If you have no rear speakers, it'll disappear.
L7 works great with stereo recordings and recordings designed to reproduce a space. If the mixologist used a bunch of processing because he fancies himself an artiste and wanted to create some cool effect, then that cool effect will be made even cooler by L7.
If you want to test this for yourself, download audacity and make a pink noise track. Then, reverse the phase of one channel in the track, burn it onto a CD or load it onto your iPod and play it back. The noise will steer rear.
Then, record your own voice (or someone else's) and switch the phase of one channel. The voice will appear in the back. You can build a track where the voice will rotate around the car. You have enough info to do that.
Or, if someone has a place i can upload a track like that, I'd be happy to.
______________________________________________________________

Andy Wehmeyer :
MS-8 hates one speaker out of phase, especially one of the front midranges or midbass.
In about 90 percent of cases where people have contacted me about bad sounding MS-8s, especially "no midbass", "center image with no focus", or "near-side bias" this has been the problem.
One of the installation schools and plenty of people who claim that EQ is the devil tout the advantages to running one midrange out of phase. They claim that this improves the center image. It's BS. In most cars with door mounted midbass or midrange, the difference in distance from each of the doors to your ears puts those two speakers out of phase at about 200Hz and at multiples of 200Hz. 200Hz is right in the middle of the vocal range, and singers are often recorded in mono, so they seem to be located center stage. Putting one midrange out of phase when the speakers are mounted in the doors puts 200Hz back in phase, but moves the first out of phase frequency to double the original frequency and multiples thereof.
The reason that MS-8 hates this is because it uses time alignment and signal steering to overcome the out of phase problem for center steered sounds. Because there's a center channel, equal pathlengths aren't a big deal because the mono information doesn't have to be created completely by the acoustic sum of the left and right speakers. It's sent to the center speaker and attenuated by 6dB in the left and right. An out of phase door mid won't KILL the center, but it will screw it up. Out of phase speakers don't sound the same with MS-8 as they do in a simple stereo system, so they're harder to detect when you listen for imaging, but they have a huge detrimental effect on frequency response.
In a L7 system. left of center and right of center images are created by the acoustic sum of left and center and right and center. A sound that should appear in between left and center is attenuated by 6dB in the left and in the center, so that the sum appears between them at the proper level. If either speaker is out of phase, it screws this up too.

Tendean17 · Feb 20, 2011

Andy Wehmeyer :
Phase shift only matters in the acoustic response that you hear. At the crossover frequency, it's sufficient to choose slopes that provide a smooth transition in that region. If there's a deep and narrow dip in the response when you measure it with n RTA, then you should choose a different slope or reverse the polarity of one of the speakers (you'll also have to reverse the polarity of the same speaker in the opposite channel)

Right, phase in the passband doesn't matter, so long as no other speakers are playing the same signals. As I wrote above, phase at the crossover matters only if it affects the frequency response and an acoustic measurement is the only way to tell. Don't sweat the WinISD plot for the woofer. Build the box, put it in the car and EQ it.

The 50 degree difference won't present a problem. IF the yellow trace is the high-frequency driver, it's also important to consider that the output level from that driver ought to be reduced in level, so the sum of the two will be mostly the output form the woofer.
______________________________________________________________

Andy Wehmeyer :
Well, all of that is true and it matters most when we're summing electrical signals. The electrical signal only matters for acoustic purposes since it modifies the acoustic signal. The phase that Adam is referring to is measured on the electrical output.

The phase between two drivers that must combine acoustically to provide the appropriate response at the listening position also depends on the distance between those two drivers and the listening position. If you have an RTA, you can sort this out so long as you have some way to adjust the phase---hmmm...don't know of any aftermarket processors that allow this to be done manually. So, my suggetsion is to use steeper slopes which minimize the need for all of this work, because the range where the drivers have to be in proper phase alignment is minimized--less overlap or a smaller transition band--call it what you want.

For speakers mounted close together on a flat baffle (like a home speaker), the rule about electrical filters that Adam has explained loosely translates to acoustic responses because of the position and the proximity of the drivers and the fact that they can often be filtered in a range where their response is flat. In cars, this is rarely the case and the benefits of mounting the tweeters high in the car and the woofers low in the car where they fit outweigh the crossover phase crapshoot mitigation that might be achievable by mounting the speakers closer together.

Use a 4th order setting between speakers and lower orders on the top or bottom of the entire band you're reproducing. There's no need for a 4th order subsonic filter unless your subs are small. I suggest a 2nd order there. If you're using rear speakers, set them at 100Hz 4th order for other reasons that I can explain later. That's a L7 deal and my suggestion will preserve midbass impact and the illusion of bass up front.
______________________________________________________________

Andy Wehmeyer :
The out of phase voice should steer to the rear. That's what Logic7 does.

MS-8 isn't going to get rid of the bass you feel, but it will get rid of bass that's more than 9dB above the rest. This is why I keep saying to turn the bass amp down during calibration and then turn it back up. MS-8 is designed to make it sound like the bass comes from the front, if possible. That can't happen if the bass is 20dB above the rest of the system. If you don't care about the placement of the bass, but you want a ton of it, then take my suggestion. Turn the gain down, run calibration, turn the gain up.
______________________________________________________________

Andy Wehmeyer :
Out of phase information in the recording will be steered to the rear. In acoustic or direct to 2-track recordings or recordings designed to recreate a live event, this information will mostly be room reflections. Steering those to the rear works pretty well in terms of creating good ambience. The algorithm is similar to PL2 in the steering of signals to the rear, but L7 has some other goodies that make it possible to resolve rear left and right (something not done very well by other algorithms) and another feature that helps to plant front steered images in front. The difference between rears and sides is an additional few milliseconds of delay.
______________________________________________________________

Andy Wehmeyer :
I just checked out the owner's manual for the DCX 730. In a very basic 2-channel system, the EQ is pretty similar. The 730 allows 7 bands of parametric EQ per channel. MS-8 uses 8 per channel. The difference is that the 730 allows you to use standard frequency, gain and Q adjustments. The MS-8 doesn't allow you to manually adjust the parameters of the filters, but the 8 parametric bands can be ANY shape. That means that MS-8 isn't bound by the traditional frequency, Q and gain parameters.

MS-8 uses an impulse response measurement to determine the location of the first big peak in that measurement. Then, it backs up until the response is 12dB down and sets the delays for that point. This better approximates the arrival time of speakers that don't provide much high frequency response compared to just picking the top of the peak. The amount of high frequency in the measurement determines the slope of the line, and more high frequency makes the peak more pronounced. Any of you who have tried to determine arrival times with an impulse response measurement have undoubtedly seen this, especially for subs. The impulse looks like a big hump. The initial arrival of sound from low-passed speakers is NOT the top of the hump. Delaying the rest of the system to meet the top of the hump of the subwoofer is unnecessary and just chews up processing power. In fact, it isn't necessary to consider the arrival time of the sub in a car. However, if your sub amp includes some additional DSP that introduces a delay, adding delay to the rest of the speakers manually can be beneficial.

Since both processors operate at 48k, the resolution of the delay adjustments is the same--that's determined by the number of samples per second. A one sample delay at 48k is .238 inches.

MS-8 follows a set of rules to tune the car. If you tune with the 730 (or any other manual processor) you're still bound by some rules (consider the frequency, Q and gain example above), but not by others. If your system includes something that MS-8 can't deal with, then tuning manually may give you better results. Putting one midrange out of phase to "fix" the center image is one of those things. MS-8 HATES that and will sound terrible if you do this. There won't be any midbass, but the delay adjustment will provide a decent center image, anyway. If you don't like MS-8's target curve and the 31-band EQ doesn't provide enough range to change it to something you do like, then manual tuning may make you happier.
______________________________________________________________

Andy Wehmeyer :

Andy Wehmeyer said:
OK. Sorry for the delay. I don't use True RTA, but they're all pretty much the same and include the same tools.

We'll start with a few basics here and most of this isn't my opinion. Rather, it's stuff that can either be verified and has been written about by various people who get paid for this stuff because they're scientists. Some of them are people with whom I work and some is available at the LIBRARY.

Anyway, cars are full of reflections and there are all kinds of opinions about what to do with them. Since the dimensions of the car are small, the car doesn't create a reverberant field--all reflections arrive within 20mS of the initial sound.

Second, when the same sound emanates from a particular source and is also reflected from a nearby source, the difference in arrival time and frequency responses of that sound create comb filtering. At some frequency and at multiples of that frequency, based on the difference in distance from the two sources to the listener, there will be destructive interference. The lowest-frequency null is the widest and each subsequent null (as frequency increases) will be narrower than the one that preceeded it. Hereafter, we'll refer to that width as Q--low-Q is wide and high-Q is narrow.

Here's a screen shot from an RTA to illustrate this. To simulate this, I've set the RTA to show the sum of left and right, and I've delayed the right by 1mS which is about equal to a foot--1.132', to be more exact.

In the next screenshot, I've increase the delay to 2mS, and you can see that the first null happens at a lower frequency.

In the third screenshot, I've reversed the phase of one channel and it's plain to see that I've moved the null back up, but I've killed the bass response. Switching the phase has put the sound from the two channels back in phase at frequencies where it was out of phase. It's also put the sound from the two channels out of phase at other frequencies.

Also remember, that in order for two sources, in this case a right and left speaker, to produce a phantom image in between the two speakers, they must both be in phase. If you sit equidistant from a pair of speakers that are in phase, there's an image. If you swap the polarity of one, there's no image, less bass and you hear the speaker locations. If you're not equidistant, and the two speakers are in phase, then you'll hear a phantom image at frequencies where they are in phase, and no image for frequencies where they are out of phase. In the case of the third screenshot, --a pair of speakers located where the PLD is about 2.5' and one speaker's polarity is swapped, there won't be an image at low frequencies and the bass will suck, there won't be an image at 500 Hz and at multiples of that frequency. The comb filtering causes this image "problem" as well as the obvious frequency response problem.

For right and left speakers mounted in the doors, the first and lowest Q null happens in the vocal region. Swapping the phase of one channel can dramatically improve the image for a vocalist, but screws up the bass and screws up the image at higher frequencies.

In the fourth screenshot, I've simulated time alignment by putting the channels back in phase and delaying the other one by 2mS so they match. See, flat response.

OK, now that we're clear on this, let's think a little more carefully. When we minimize pathlengths, we simply move the null to a higher frequency. Shorter PLDs place the null higher and higher until when the PLDs match, there's no null. This exercise also shifts the frequency where we begin to hear the speaker locations to higher and higher frequencies. Hmmm...this means that for a sound that includes midrange and high frequencies that should appear in the center, if the PLDs are close, we'll hear MOST of it in the center and we'll hear the high frequencies in the right and left speakers. When we listen to lots of events all playing at the same time, some of it will appear in the center and some on the right and left.

We're all so used to listening to small PLDs that we hear those errors as stage width, even though it doesn't exist in the recording. It's kinda nice, because small PLDs don't screw up anything for the fundamental sounds in the recording, but do screw things up for harmonics. That condition adds this sense of space --or width.

OK, the sound we hear in cars and rooms includes reflections, and those reflections also create comb filtering, which change the frequency response and change the image. For the rest of this, remember that reflections from nearby surfaces have short PLDs and from far-away surfaces have longer PLDs.

For an even more detailed picture of what happens, it's important to consider speaker directivity. Adjacent boundaries will produce reflections for speakers at frequencies where the dispersion is wide, but not where the dispersion is narrow. The angle of incidence equals the angle of reflection, and that's anpther thing to take into account when considering how all of this interacts. We don't have to be concerned with reflections until the sum of all of them reaches our ears, because that's the only place where we hear their sum.

Finally, in all of this talk of reflections, we have to consider that the sound from the right speaker will be reflected off the left side of the car and sound from the left speaker will be reflected off the right side of the car. This crosstalk partially destroys stereo, as each of our ears hears both speakers and the arrival times are within 20mS. This is worse in cars than in rooms, because the surfaces are closer. That means the intensity (loudness) of the reflections is greater and our perception that they and the initial sound are the same is greater. This is what makes a car sound like a car rather than a much larger space.

The first rule of equalization is that you can't put energy back into a null caused by destructive interference in the acoustic system. You can remove energy from a peak caused by constructve interference.

The second rule is that high-Q (narrow) peaks and dips are less audible than low-Q peaks and dips. When we listen to music, our ability to perceive high-Q peaks and dips is further decreased because the chances that a particular sound happen at precisely the right frequency, thereby energizing the high-Q peak or dip is reduced. In essence, this means that there may be no need to smooth the TOP of the comb, but the bottom of the comb is a problem. Plus, many listeners like the way tht top of the comb sounds, because it adds a sense of width.

FWIW, this delay between the channels is used by music producers to provide a sense of spaciousness or a separation of the sound in right and left channels.

OK. I'm clicking "sumbit reply" here and I'll continue in another post--this too much work to lose by pressing the wrong key.

JBL MS-8 FAQ

Top Contributors this Month

Recommended Communities