DiyMobileAudio.com Car Stereo Forum banner

1 - 16 of 16 Posts

·
Registered
Joined
·
1,205 Posts
Discussion Starter #1
I know there has been a lot of going back and forth about John, and we all understand he has not made it to the point of acceptance to the team due to his currently noncompetitive status, but his recent PM to me underscores why I think he is a worthwhile candidate to watch. Of the dozens and dozens of PMs I have gotten about my car, his was the most thoughtful and well-worded of all of them. As a result I went out of my way to answer the questions I thought he was going to ask, and then eventually the questions he DID ask (LOL). It is long and wordy, but it is the subject for one of the 'articles' I was hoping to post to the Team section in due time. I've included both the question and my response. I hope that some of it helps someone along the way, and by posting this I am no way insinuating that I know what the hell I am talking about. Not the expert here, but this is...

High Frequency Reproduction the Way I See It.

Ba dum dum duuuuuuuuhm...!
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #2
The suestion, as asked. Taken without permission from PM. May the Diety of PHPbb3 have mercy on my soul.

"Todd-

I know that reflections in general are a bad thing and should be avoided as much as possible. I know that's why most home audio listening rooms don't have the towers sitting in the corners, but in a car it's nearly impossible to avoid reflections. I have, however, begun to form the opinion that reflections in the treble are the most offensive (and the easiest to avoid in the car). I decided to shoot you a PM since I noticed the tweeter location in your car appeared to try and avoid reflections and wanted to get your opinion on it.

In my car, I have the tweeters in the pillars about 15 degrees off-axis for the opposite listener. I did this because I was aiming for a two-seat setup and wanted to compensate for the proximity of one speaker by putting it more off-axis to the near listener...essentially it was an attempt to acoustically balance the tweeters for both listeners since it cannot be done easily electronically. Initially, I was happy with how everything sounded. I felt like I had a good balance for both seats, the stage was across the dash, and it sounded very "full" and "lively". I ended up ordering a dashmat and noticed a large difference with it in. It seemed to make things less "full" and "lively", but it also allowed the stage to be much more "solid". Since it was easy to measure, I used my RTA to compare the frequency response with the mat in and out. Surprisingly, there wasn't a noticeable difference on the RTA but an obvious difference to the sound (I'm still a bit surprised by this since, as I learn more, I would expect the peaks from the reflections to amplify the FR). I actually didn't car for how the car sounded with the mat in when I was going back and forth, so I left it out. However, I quickly learned that I wanted the mat back in. It seemed that when listening at loud levels for extended periods of time, fatigue set in quickly and I had to turn the volume back down. With the mat in, I lost the "liveliness" but found I could listen for much longer periods. This is where it began to dawn on me that the "liveliness" that made things fun, were reflections creating artifacts that weren't there.

I mentioned noticing the tweeter location in your car, but honestly didn't think much about it until recently. After installing the MS8, I set everything up without the dashmat in. I had pulled it while working around the dash to get my Denford out and didn't think to put it back before tuning. I noticed while listening to 'Whiskey Lullaby' that neither singer was "stable". The didn't really pull to the left or right either, but more appeared to "vibrate" (I don't have a better term for it). I decided to slap the mat back in and...voila, solid vocals! It made a huge difference. My wife happened to be sitting in the car when I did it and she agreed. I did seems to lose some "sparkle" though. So I started thinking about the artifacts and the effect they have. It is why I formed the opinion that they can be a easy "trick" to help a non-ideal install but should ultimately be avoided.

Now, my reading on the forum does not follow this linear experience (I obviously knew amplitude dominated in the treble because I tried to compensate with tweeter aiming), but sometimes the things you read and do don't click immediately. Today I believe I had that 'ah ha' moment with the high frequency information. What I was thinking sounded "lively" and "full" was because of reflections...bad reflections. "Lively" was coming from peaks and nulls, and since we are most sensitive to amplitude above 3khz, these peaks and nulls were creating a very uneven, and very noticeable, frequency response. The "full" sound was from all the reflections across the glass and dash, and while it can easily make your stage seem to go completely across the dash, it also keeps anything from being truly "solid" or fixed where it should be (i.e. it "vibrates").

I do want to keep the MS8, and I believe it CAN do a good job with drivers that don't have reflections...in the treble. So I'm looking to avoid that in my next build. I wanted to see if you tried to avoid the same issues in your build. While the Murano doesn't have sails, I have also been pouring over PB's threads on directivity and I think I can do something else to avoid them. Primarily, use a (better designed) waveguide in the center to reduce reflections. For the sides, I will either put a waveguide in the very corner where the pillar, dash, and windshield meet, or I can use the sphere's that PB suggested.

I could keep going, but I think I've typed enough to generate some discussion. I would like to also state that I don't think reflections anywhere else in the musical spectrum are worth trying to avoid in the car (that opinion can change however). I say this because as frequency decreases we're actually less sensitive to peaks and nulls, and it also becomes harder to fight because the larger drivers we use have wavelengths that are longer and omnidirectional (and therefore make reflections nearly impossible to avoid). I am also interested in your opinion on this since I think you may be doing something with purposely using drivers that beam on dash*.

* I could be wrong about this and only draw it from you using 6.5" JL's as midranges and they seem to mate with TINY tweeters in the sails (which is why I asked about your crossover point).

I look forward to hearing from you and I hope my reasons and thoughts are well-founded on the subject.

John"
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #3
John- I wrote much of this in response to your original post, and see by your following PM that you are well past most of this. I am including it anyway as it will be forming the basis of a short article for the Team section of the forum. I will address the balance of your PM at the end.

------------------------------------------


My tweeter setup isn't very unconventional; they are placed in the sail panels as they are the widest point in the car. Though the midrange sets stage width, the tweeter can be used to set stage boundaries. As you've noted, however, they can also be used to screw up stage boundaries as well.

In my car you got the chance to hear that (I think) on a live track; the whole stage seems to open up. If you listen carefully, the sounds that appear to leave the car aren't reflective phenomena; they are individually identifiable to the maximum with of the stage. This is achieved by balancing the tweeter's power response, just as you have mentioned. The ratio of direct to reflected sound reaching the listener must be high enough that you clearly hear the direct sound as being LOUDER than the reflected sound (because we localize high frequency sound by IID; intensity differences). Next, you want the direct and reflected sounds to arrive very closely matched in TIME. If they are not closely matched in time, the human brain will INTEGRATE them into one sound due to the Haas effect, but that will result in the sounds SMEARING and losing focus. Finally, we have to balance the left and right IIDs with each other, and provide both ears with a smooth, balanced frequency response to help reduce 'pulling' of the soundstage. The ear will que in to the speaker playing a note the loudest if the note was destined for any location on the stage besides hard left and hard right. We hear this in the sibilance of a singer's voice jumping away from their mouth and coming from the far (or near) tweeter. This is helped by the 'my soundstage ate my windshield' thread, what I call 'differential equalization'. We maintain the same SUM while changing the ADDENDS.

Briefly, let's take 5Khz for example.
At 5K, the overall loudness is where we decide we want it. 5Khz blends appropriately with 4K and 6.3K. Looking at the EQ, 5K is at +2 in the left tweeter and +2 in the right tweeter. Every time Natalie Merchant breathes at the mic, the sound comes from the left tweeter. We need differential EQ at 5K to help with this. The SUM of the L and R EQ at 5Khz is 4 (as in 2+2). The sum needs to remain the same, but the addends must change to balance the problem we have at 5K. We reduce the left by one and add one to the right. 1+3=4. Not enough, so we do it again. 0+4=4, and now Natalie's mic no longer jumps about. We kept the sum the same, and changed the addends to compensate. The same amount of 5Khz sound is in the cabin, but it is coming from different locations on the soundstage.

(continued in the following post)
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #4
So now we understand what we need to do to place the tweeter in the soundstage, and we understand how to fix simple problems with an imbalance of the soundstage. Let's put it into practice.

I first placed my tweeter as CLOSE to the glass as I could get it firing straight up. I tried this at a number of different locations and arrived at the sail panel as the best choice for my car. This places the direct and reflected (remember, ever speaker in the system combines with every reflection in the system to create an acoustic mirror driver) sounds so close together that not only are they heard nearly simultaneously, but they actually CREATE AN ARRAY. (This, by the way, is something that I tried to leverage on the midrange installation). If I had left the tweeters in this configuration (firing straight up next to the glass) what you get is a HUGE soundstage. The tweeter and glass are reflecting and bouncing sound everywhere. The near side, far side and front glass are all exposed to the off-axis response of the tweeter, and the on-axis response is firing towards the headliner. The direct:reflected ratio clearly favors the reflected energy. There is no ability to focus the response, however. Nothing appears as being tight and focused in the soundstage. Vocals and sibilance flail wildly around the soundstage depending on frequency and amplitude of the reflected sounds, and there is so much comb filtering of the reflected energy that the response is peaky at best. So we know we can't stay here even though the effect is interesting. When we really get to listening we can hear how it's just not right.


Next, I needed to balance the tweeter's power response to reduce the reflected energy as much as possible. I began rotating the tweeter toward the opposite listener. This reduces the reflected energy and increases the direct energy reaching my ears. The more I rotate the speaker towards the other side of the car, the less reflected energy I get from the near side glass, but the MORE I get from the far side glass. So I rotate it to a spot directly in front and above the closest listener. The tweeters fire like this: /--------\ in front of me. Now the stage is more focused, but the driver furthest from me is louder than the driver nearest me. No good. So now I start rotating the speaker towards the back of the car while maintaining the 'toe in' that I have set in the previous step. The near speaker gets louder as the far speaker stays about the same relative volume (my axis with the far speaker doesn't change as quickly as I am at a less acute angle to it. Not so with the closer speaker!). I continue rotating and retuning until I have the best balance of direct and reflected sound, and the balance left to right is as close to even as possible for the majority of the frequency range. As a bonus, I am firing directly at one of the softest absorptive surfaces in the car besides Ray Hower's head (the headliner) further reducing reflection of the most spectrally rich energy (the center lobe of the tweeter)

(Continued in following post)
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #5
OK, so now I have a rough location and aiming for the tweeter. Now I need to assess the soundstage.
I notice that the right side of the stage is wider than the close side; a common problem. Up to this point my crossovers were set the same for both speakers; I chose 4.5 Khz by ear as that is just below where my midranges started to roll off. Remember that you are at different angles to the midranges as well, so their power response also differs between them.This I reduced by placing the midrange very close to the reflective barrier to create a close-set acoustic mirror; a vertically stacked midrange array. Back to the tweeter, I need to pull the right stage in a little and try to open the left stage up a little. I chose to reduce the xover point on the left stage, placing a little more midrange in it, while raising the crossover point on the right stage. What?! Since the tweeter is up high and in a good line of sight, doing this (with the midrange-to-tweeter spacing in my car) tends to close up the right stage and open up the left stage. What we are looking to do is balance the left and right stage around the center of the stage. This allows the judge to more precisely locate center by either the MECA method OR the 'right' method. Either way, center of the car should become center of the overall soundstage. By placing more spectral information in the left tweeter, the soundstage appears to come from the wider tweeter location instead of the narrower midrange location. Clearly, this will not work if you are crossing at 10K and only using a 'supertweeter' methodology. The radiation pattern of the tweeter also narrows the higher you play it (or cross it over), so I am reducing reflection on the left and increasing it on the right...by the tiniest amount.

This is also why I don't do that unless the midrange and tweeter are RIGHT next to each other. There is too much power (tuning capability) in being able to bend the soundstage a little bit with your tweeters.

I really like the use of the Ring Radiator tweeters. They tend to have a smoothly falling off axis response, and a strong on-axis response; for our purposes ideal and highly directional. They also tend to have a very open and airy, detailed response. My tweeters cross over at 3.6k on the left and 5K on the right. My stage is well centered and level side to side.
My tweeters are located in the sail panel under the A-pillar, and fire into the headliner just above and in front of the near-side listener.
I reduce the near side tweeter by 1dB as opposed to the far side tweeter for balance. I also reduce the midrange and midbass levels on the left side of the car. 1dB for the midrange and 2dB for the midbass.
----------------

(Continued in following post)
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #6
Back to your specific questions...

It is my opinion that the 'liveliness' that you associate with the reflections is due in large part to the time factor as much as (if not more than) the frequency response factor. This would not show up readily on the RTA as the RTA averages time information into its power response measurement. It would however show up in impulse response measurements. Unfortunately for your situation, the MS-8 tunes partially by impulse response (room correction). I have never tried to make that measurement with an MS-8 in Bypass, so the results are uncertain. The vibrating, as you put it, (shimmering is another term I have heard used to describle it) may very well be an artifact of the room correcting function of the MS-8 in play. It uses a 'moving window' function to reduce processing overhead when applying high frequency FIR filtering (as best I can tell...I have not read anything to confirm it, but it appears to be a measure often taken as an optimization from my research). The best way to manage that is of course to make the MS-8's job as easy as possible. The less correction it has to do, the fewer artifacts are audible.

You may have read about my 'tweeter lens' experiments in my previous build. This is precisely the reason that I was following that route, albeit with rabbithole reasoning. My designs were simplistic, but based on the Bang and Olufson BEO patents. Have a look at the BeoLab5 and it will make sense. The lens created the directivity without the obnoxious horns normally used for that purpose. The result can be quite pleasing and overall rather convincing, but my measurement equpiment didn't have the resolution and I didn't have the time to devote to re-engineering their design. It was hit-or-miss engineering on my part, and though it was mostly hit, it wasn't up to snuff. I fear that is the primary problem with waveguide design from a DIY perspective; managing the tradeoffs is a time-consuming and daunting challenge for a problem that can be solved with simpler technology. Watching PB's threads has reinforced that thought as well, as he appears to take a swat-the-beehive approch to his otherwise methodical theorizing. I've yet to see a design that was the result of a mathematical approach based on the problem (a parabolic Unity horn isn't parabolic just because it is 'generally oblong' in shape...! The parabola should be a mathematically correct three dimensional model of the function, not something carved out of a buck with one eye closed.) I am certain there are gains to be had with constant directivity designs - that is more than clear from my experimentation and from the efforts of others (Mark E's Bronco dashboard went through numerous revisions as an extended waveguide to the horns in the car - and he had the help of Linkwitz to accomplish it. I don't have those resources to leverage ot the time to invest, so I fell to 'good enough').

If you really want to know where the problem lies, tape a towel to the BOTTOM of your dash under the steering wheel. Push the body of the towel back over the pedals to create a padded cave. Place the tweeter back into that space on both sides of the car pointing at something soft and distant. Run a tune on the MS-8. It will be freaky-deaky good, though it may result in a stage below the horizon as well as the possibiltiy of numerous other negative side effects. This is the best you will likely ever get for HF response, but other factors will suffer.

(Continued in following post)
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #7
The MS-8 IS great. There is no doubt. But I think to set the car up initially you are better off using a manually tuned processor. Since you can't see what the 8 is doing to (or for) you, it is difficult to identify the source of a problem and FIX it. I tried (pretty hard) to use it to set my car up. I gave up and put the 701 in. though the 701 runs out of steam at 90 percent of perfect, the MS-8 runs out sooner...and there is no way to tell when you get there what will be required to get the rest of the way home. I wasn't content with this, and it took me months of practice to learn how to out-tune the MS-8.

Have you had luck getting a centered, balanced stage using a center channel that didn't sound hollow and lifeless? I never could, and I don't know why. The headbobbing that worked so well to get a phantom center seemed to do nothing to move my center image away from the steering wheel...

On closing, note that I am by no means an expert on any of this. It is at best a pieced-together mental model of all of the things I have read on the subject, and probably has holes large enough to drive a semi truck trough sideways. It is, however, enough to have driven me to the results you heard and I think that's pretty darned good. Always room to improve the mental model and the reults in the car though! I hope the time you invested in reading this was worth it.

-Todd
 

·
Moderator
Joined
·
18,607 Posts
Man, there's a lot here I'd like to discuss but I don't have the time now. This is the perfect place to bring up narrowing a tweeters output on one side of the car via absorptive material on the reflective side of the tweeter and letting it play freely in to space on the interior side of the cabin.
 

·
Registered
Joined
·
1,205 Posts
Discussion Starter #9
Pretend it's EMSQ without the haterz and go to town, brotha!
LOL

In the meantime, any discussion that can be brought to bear on the subject is invited! After all, 'Scientific Car Audio - Truth In Sound Quality', right?
 

·
Mobile Audio SQ
Joined
·
3,367 Posts
Erin, isnt that what Gary did in the Benz? He put "covers" or half domes over the tops of his dash speakers to keep them from having reflections off the glass IIRC.
 

·
I'm baaaack!!
Joined
·
4,397 Posts
Erin, isnt that what Gary did in the Benz? He put "covers" or half domes over the tops of his dash speakers to keep them from having reflections off the glass IIRC.
Yep. He and I talked about the different versions of "shroud" that he has used while that car was evolving. It's sad that I won't get to judge it again.
 

·
Moderator
Joined
·
18,607 Posts
Erin, isnt that what Gary did in the Benz? He put "covers" or half domes over the tops of his dash speakers to keep them from having reflections off the glass IIRC.
Problem I had with what he did is the material wasn't enough to absorb anything below probably 8khz. The shroud wasn't long enough to control dispersion in the lower end and id bet by the time they were effective, narrowing was occurring In fact, he probably created some issues. I'd like to see measurements.
 

·
Mobile Audio SQ
Joined
·
3,367 Posts
What about using a tritrax(?)horn type setup where the side that was facing the reflective surface was shrouded or protected while the inner wall of the horn was covered in absorbing material. Set the tweeter far enough back into the horn to give it some dispersion and a good amount of space to work with going forward.
 

·
Mobile Audio SQ
Joined
·
3,367 Posts
Not working in my head. Plot it in Hornresp? :D
Sort of, more like a deep JBL GTi housing where it's a rounded off square horn housing. The tweeter would be recessed about 1" into the housing in a tube or recessed housing and the horn itself is treated. No idea of what would happen just an idea
 

·
Premium Member
Joined
·
3,854 Posts
What about using a tritrax(?)horn type setup where the side that was facing the reflective surface was shrouded or protected while the inner wall of the horn was covered in absorbing material. Set the tweeter far enough back into the horn to give it some dispersion and a good amount of space to work with going forward.


ID NX30 mounted in a one off pair of Tractrix horns:D:D
Chris Owens said it was one of the top 3 best sounding cars at IASCA Finals that year....she was in Novice that year-finished 4th
 
1 - 16 of 16 Posts
Top