DiyMobileAudio.com Car Stereo Forum banner

1 - 20 of 32 Posts

·
Registered
Joined
·
464 Posts
Discussion Starter #1
If you listened to just one tweeter and midbass driver, such as one side of a comp set, what is the apparent point of origin of the sound? Does it sound like it's coming mostly from in-between the two drivers?

What about for imaging? - I think in that case the midbass driver would dominate and most of the directional cues would seem to come from there.

What I'm getting at - I'm thinking about cheating with an install - instead of putting a comp set in the kicks, put the midbass drivers in the doors and the tweeters in the kicks - the middle point between the drivers would only be moved back toward the seats about an inch. What do you say?
 

·
Premium Member
Joined
·
6,045 Posts
If you listened to just one tweeter and midbass driver, such as one side of a comp set, what is the apparent point of origin of the sound? Does it sound like it's coming mostly from in-between the two drivers?

What about for imaging? - I think in that case the midbass driver would dominate and most of the directional cues would seem to come from there.

What I'm getting at - I'm thinking about cheating with an install - instead of putting a comp set in the kicks, put the midbass drivers in the doors and the tweeters in the kicks - the middle point between the drivers would only be moved back toward the seats about an inch. What do you say?
The origin of sound depends on a couple things.

The first thing is the wavelength of sound. Basically, above 1khz your perception is sound is based on amplitude, and below 1khz it's based on phase. Amplitude is easy to manipulate; a $40 equalizer can do it. Phase is a lot more difficult to manipulate.

This is why a lot of psychoacoustic 'tricks' designed to make the soundstage appear wider rely on manipulation of high frequency response. It's easy to tweak the response, and certain changes can make an audible difference in where the perceived location of sound is.

This is a complex subject. I could easily spend an hour doing a 'brain dump' of how we perceive the location of sound.

To keep it simple, here are a few tips:

  • At high frequencies, frequency response rules everything. This is why the spherical tweeter enclosures work so nice. They improve the off-axis and on-axis response. At high frequency, LOCATION isn't as important as AMPLITUDE. There's no need to obsess about your tweeter location - instead you should obsess about getting them to match. You can put your tweeters in some very oddball locations - just ask Image Dynamics.
  • At midrange frequencies, EVERYTHING matters. Both frequency response and loudspeaker location. There's no shortcuts here. All the DSP in the world won't make up for a crummy location.
  • At low frequencies, you have a lot of leeway for location. What's critical is to get the phase response correct. This is why equalizing path lengths is so effective. It gets the phase of the left and right drivers closer in 'sync'. If you're willing to settle for a system that only sounds good in one location, you can 'cheat' with DSP. Personally, I'm not a fan of DSP because it screws up the reflected sound, but I'll admit it can be effective, particularly if your stuck with sub-standard location.

A lot of the soundstaging stuff was basically figured out in the 90s. Guys like Eldridge and Clark and Holdaway posted about these subjects extensively, particularly on audiogroupforum.com

HTH

 

·
Registered
Joined
·
5,098 Posts
If you listened to just one tweeter and midbass driver, such as one side of a comp set, what is the apparent point of origin of the sound? Does it sound like it's coming mostly from in-between the two drivers?

What about for imaging? - I think in that case the midbass driver would dominate and most of the directional cues would seem to come from there.

What I'm getting at - I'm thinking about cheating with an install - instead of putting a comp set in the kicks, put the midbass drivers in the doors and the tweeters in the kicks - the middle point between the drivers would only be moved back toward the seats about an inch. What do you say?
I have my tweets on the dash and the mids in the door. I will often tune playing only the mid and tweet on one side. I do this to get the two drivers in phase and for tonality. When I have the TA right the point source for the entire sound will be up, at the tweeter level. The rest is already covered by Patrick.

If you're running a 3 way the mid range is driver that needs ideal placement, both for PLD and axis response. Less than ideal locations for mid bass can be covered with TA and the tweets can be largely covered with a L/R eq. The axis you mount the tweets at is more important than the PLD between them. This is all from a one seat setup.

Along these lines, Patrick, what would you consider great pathlength differences? I've seen 30cm/12in listed as good, what would be great?
I'd say get it as low as possible then use a bit of TA. Under 8-10" is what I have read about as being passable.
 

·
Registered
Joined
·
2,849 Posts
Ok. Deciding whether to mod my seats to move back even farther than I already have. My midbasses in the kicks have a PLD of 9.75", and my intended midrange location has a PLD of 10.5". Seats are already moved back 1.75" farther than stock, farther then that and I'll need to extend the pedals and steering. I have time alignment, just trying to get it as physically correct as possible before using electronics.
 

·
Premium Member
Joined
·
6,045 Posts
Along these lines, Patrick, what would you consider great pathlength differences? I've seen 30cm/12in listed as good, what would be great?
I obsessed over imaging in the car to an incredible degree. This probably culminated about three years back, with the "perfect soundstage" thread.

After doing that project, I realized there's just not a whole lot of imaging in the damn recordings.

Ever since then, I'd say my priorities look like this:

  • Good articulation. An articulate design allows you to 'see' into the recording in a way that 95% of speakers don't. For instance, with an articulate loudspeaker the human voice sounds natural and lifelike, and you can perceive the 'texture' of percussion.
  • Dynamics. There are a lot of small speakers that are articulate, but lack dynamics.
  • non-fatiguing. I find that hi-fi systems tend to become fatiguing. For instance, I've heard a lot of overly detailed 'hifi' systems which sound 'exciting' for the first few minutes, but an abundance of hyped-up detail makes them fatiguing quickly.

Basically, imaging is a fun trick on that occasional recording that has soundstaging 'in the mix.' But for 95% of my listening I get more excited about deciphering the lyrics to a song I've loved for years, or hearing the singer's voice as it sounded in the studio.

The crossover on this Unity sucks - it's literally one capacitor - but compare how voices sound on it, versus a conventional loudspeaker:
http://www.youtube.com/watch?v=P-oAUZsK6es&feature=related (unity horn)
http://www.youtube.com/watch?v=SAatTvOlMIM&feature=mfu_in_order&list=UL (typical loudspeaker)

That's why most of my projects tend to use drivers that are *extremely* close to each other, typically less than ten centimeters, and I don't use steep filters because they wreck your phase response.

In the videos, note how the voices coming from the Unity horn *almost* sound like a human being speaking in a room, while the typical loudspeaker sounds exactly like what we're accustomed to - a loudspeaker. And that JBL is very nice - it's hardly a bad design, but it's phase isn't in the same league as the Unity. (I took the JBL apart, and it uses steep slopes that will screw up your phase, and the drivers aren't as tightly packed as the Unity either. The JBL woofers are spaced over 20cm apart.)

I hope it doesn't sound like I'm smashing the JBL. It's a fine loudspeaker and I am happy with it. But it sounds like a loudspeaker. The Unity horn, by contrast, *almost* sounds like someone speaking. The downside of that particular horn is that it's EXTREMELY directional. In the video you'll notice that in some locations the voices are very natural, but in others, it sounds like the recording is in a barrel. That's due to the extreme 'directionality' of the horn. (I'm working on that, stay tuned...)

IMHO, a lot of people in car stereo are NOT satisfied with their systems. (Just look at my sig.) And I believe 95% of the car stereos out there make some fundamental mistakes. For instance, a lot of two-way component sets use a low crossover point, and a tweeter that's located 20cm away from the woofer, or even further. That can never work. No amount of DSP in the world can take the output of two drivers spaced 20cm apart and assemble them into a cohesive whole when the frequencies are so short. This is the reason that many people throw in the towel, and go back to stock systems, then find that the stock systems are more musically satisfying than battling with 'tuning' an expensive component system. I seriously find it heartbreaking to hear how many people drop $1000 on a car stereo, only to sell it in the classifieds because it was ultimately disappointing. And I'm not saying that low xover points are impossible; I use them all the time. But the drivers need to be *extremely* close together to make it work. Basically, the distance between your drivers is greatly influenced by the size of the wavelength. And low xover points are particularly tricky because our ears are more sensitive in the midrange than at high frequencies.

A side benefit of all of this, ironically, is that the imaging improves quite a bit if the phase response is excellent. As you can imagine, it's hard to create a believable soundstage when the output of two drivers is out-of-sync in the time domain.

 

·
Registered
Joined
·
17 Posts
Just wanted to say thanks Pat for continuing to provide some of the most helpful responses here, complete with good info - translations (for the technical stuff) - and examples. Thanks!!
 

·
Registered
Joined
·
464 Posts
Discussion Starter #8


.

To keep it simple, here are a few tips:

  • At midrange frequencies, EVERYTHING matters. Both frequency response and loudspeaker location. There's no shortcuts here. All the DSP in the world won't make up for a crummy location.
  • At low frequencies, you have a lot of leeway for location. What's critical is to get the phase response correct. This is why equalizing path lengths is so effective. It gets the phase of the left and right drivers closer in 'sync'.

]


- This is the conclusion I came to 2 hours after starting the thread. The midbass drivers of a two-way must go in the kicks for best results. I can't cheat physics. Thank you.

What is PLD/Path Length difference - I think I know what it is, but have never seen so many refer to it this way - can someone post a link to an a reference or article about it?

And what is the figure expressed in inches - is "PLD" 10 inches" mean your speakers are 10 inches wider apart than you are from them? I would think this distance expressed as a percentage would be more useful.
 

·
Premium Member
Joined
·
6,045 Posts
Just wanted to say thanks Pat for continuing to provide some of the most helpful responses here, complete with good info - translations (for the technical stuff) - and examples. Thanks!!
Thank you! I appreciate that. I went to the Consumer Electronics Show, and it confirmed my suspicions. I'd say that 90% of the systems I heard were little better than what you could find at Best Buy. There were plenty of $20,000 systems that actually sounded worse.

A lot of this depends on your own personal prefences and priorities. The more I learn to listen critically, the less I care about soundstaging and the more I care about reproducing the midrange naturally.

But everyone is different. To quote J Gordon Holt:

"Listeners with identical hearing acuity and identical standards of judgment will usually be highly critical of different aspects of a system's performance. Thus, expert A may be terribly, terribly critical of what happens in the high treble range, expert B may be hypercritical of bass, and expert C may have a Thing about middle-range smoothness or "coloration.""

The whole article is great - and what he said 49 years ago is still true today:

Why Hi-Fi Experts Disagree | Stereophile.com
 

·
Premium Member
Joined
·
6,045 Posts
- This is the conclusion I came to 2 hours after starting the thread. The midbass drivers of a two-way must go in the kicks for best results. I can't cheat physics. Thank you.

What is PLD/Path Length difference - I think I know what it is, but have never seen so many refer to it this way - can someone post a link to an a reference or article about it?

And what is the figure expressed in inches - is "PLD" 10 inches" mean your speakers are 10 inches wider apart than you are from them? I would think this distance expressed as a percentage would be more useful.
Are you sure the midbasses have to go in the kicks?

I'm not saying it's a BAD location. But consider this:

At low frequencies, it's *phase* that tells us where things are located. For instance, if both of your midbasses were located directly in front of you, that would be the worst possible location. The phase of the left and the right would be virtually identical, and the soundstage would have no width.

Now put the midbasses in a triangle, like in your living room, and you get nice width. That's because there's a phase difference between when the sound arrives at your left ear, and your right.

But here's something wacky - what if the midbasses were right NEXT to you? As in literally to your left and to your right?


As crazy as this sounds, it can work. It's because of the way we hear, and because our perception of location at low frequencies is determined by phase.

I've tried it, I believe Durwood is still doing it, and the old Buick Grand National did it too.

In the midrange, you CANNOT do this, because in the midrange, phase *and* amplitude matter. You *might* be able to get away with it in the midrange if you did a couple of tricks (hint - you can trick your ear into perceiving that sounds are located in front of you by manipulating the frequency response to compensate for the shape of your ear).

Anyways, I don't necessarily agree that midbasses have to be in the kicks. The kicks are a great place to put a midrange, and they're not a terrible place to put a tweeter, but you can put the midbasses in some bizarre locations if you understand the psychoacoustics involved. To get away with this, you have to be very careful with the slopes and the xover points of the midbass, and that enclosure better be solid as a rock, because the slightest rattle or harmonic distortion will give away it's location. There is also a powerful psychoacoustic advantage to not 'seeing' the midbasses. I find that 'seeing' the speakers has a huge effect on my perception of the soundstage. Yet another advantage of having the midbasses out of sight.

Check out the Opsodis paper for more info on this, or the Autosound 2000 technical briefs, or any of the papers on binaural recording
 

·
Registered
Joined
·
204 Posts
Patrick, what about a 2 way with tweeters cross high. Like 5k give or take and then mounting the mid and tweeter apart?

I have my mids in the doors as far away as possible, the kicks are just to small for my mids as
a/ In right hand drive and the accelerator pedals intrudes.
B/ Tight airspace
but when i put them in the kicks in towels to test different location there isnt much difference in centre image, left and right but what is different is if i put then very slightly on axis as opposed to flat on the door there is a big difference in sound. Better.
Difference between doors and kicks is about 7" if measuring from the centre of the cone in each area.
When i try tweeters in the kicks it sounds beautiful and very coherent but the stage is way less stable. In the pillars its stable but much easier to source the sound.
No Time alignment at hand, fully analogue, basic 5 band parametric EQ on the Arc DXE.

Iv yet to try mids in doors but this time aimed a bit on axis, tweeters in the kicks on axis and an extra set of tweeters in the pillars on a steep slope at around 7k attenuated to stablise the stage.

Good thread.
 

·
Registered
Joined
·
464 Posts
Discussion Starter #12
Oh, I know you guys are are talking about PLD now - the distance between tweeter and mid drivers on one side.

Patrick, it's true that crossovers cause problems of their own, as you well know. That's why some guys like a fullrange driver. But the fullrange introduces new problems of it's own - it's hard to reproduce a wide band of frequencies from a single dynamic loudspeaker driver. Specifically, treble distortion usually goes up with wideband drivers compared to a convention two-way with a dome tweeter.

It may also depend on what type of sound the user likes also, what their reference for great sound is. Some people like a "natural" sound like unamplified live instruments can give. A wideband driver might be better for that. I, on the other hand, like the sound of high end headphones, and this is the sound I strive for in the car, but with a front sound stage and subbass. I don't really consider this sound natural at all though, it's just my preference.
 

·
Premium Member
Joined
·
6,045 Posts
Patrick, what about a 2 way with tweeters cross high. Like 5k give or take and then mounting the mid and tweeter apart?

I have my mids in the doors as far away as possible, the kicks are just to small for my mids as
a/ In right hand drive and the accelerator pedals intrudes.
B/ Tight airspace
but when i put them in the kicks in towels to test different location there isnt much difference in centre image, left and right but what is different is if i put then very slightly on axis as opposed to flat on the door there is a big difference in sound. Better.
Difference between doors and kicks is about 7" if measuring from the centre of the cone in each area.
When i try tweeters in the kicks it sounds beautiful and very coherent but the stage is way less stable. In the pillars its stable but much easier to source the sound.
No Time alignment at hand, fully analogue, basic 5 band parametric EQ on the Arc DXE.

Iv yet to try mids in doors but this time aimed a bit on axis, tweeters in the kicks on axis and an extra set of tweeters in the pillars on a steep slope at around 7k attenuated to stablise the stage.

Good thread.
I guess I'd look at it like this - there are three options. One is easy, one is medium, one is very difficult.



The easy option is find a good location for the midrange, and be sure there's no crossover anywhere even *close* to the midrange frequencies. For instance, if you look at the Fletcher Munson curves, you'll see that we're extraordinarily sensitive from 1500hz to 6000hz. Ironically, that's right where we tend to stick our crossover point! Silly no? So getting one driver to cover that whole range can work wonders. Check out the reviews on the Dayton dome midrange, and you'll hear a lot of people wax poetic about it. It's a great driver, but I think most of the reason people love it is that it's optimized to cover this incredibly critical pair of octaves. If you can, there are a lot of good reasons to try and cross over as low as 1khz, or better yet, 500hz. This is because our perception of location transitions in that frequency band. So if you have a crossover point at 1500hz to 6000hz, you're playing with fire because our ear is very sensitive. And if you have a crossover point from 500hz to 1500hz, you're playing with fire because our perception of location depends on both amplitude and phase, but only at a narrow band - about one octave. Above and below that octave, it's location is determined by amplitude and phase, respectively.

Okay, that was the easy solution.

I'd say the 'medium difficulty' solution would be an array of two drivers. Basically, your standard component set. I'm really not a big fan of this solution, unless you can figure out a way to get them extremely close together. (Like David Smith did in the Snell solution.) If you can't get them very VERY close together, I'd be inclined to go with the easy solution.

The VERY difficult solution is to 'synthesize' a full range driver using an array. IMHO, the ultimate solution. Unity horns do this, Quad electrostatics do this, very carefully designed arrays do this. These designs basically work like a giant coaxial loudspeaker, and they take advantage of the fact that you have some leeway on loudspeaker location as frequencies go lower and lower and lower.

For instance, I'd say that it's a terrible idea to have a woofer and tweeter separated by 30cm with a crossover point of 2khz. 2khz is 17cm long. The length of the wave is critical. But if you can change that by a factor of ten, it's not such a big deal. A 3cm gap between woofer and tweeter makes everything work better. And these concepts work at all frequency ranges. A 300cm gap between a midbass and midrange crossed over at 200hz would be silly. But reduce that gap to 30cm, and everything gets better.


Does that make sense? You can 'synthesize' a full range driver by paying attention to the wavelength and crossover point.
It's like that giant loudspeaker in 'Back to the Future', only sliced and diced into various bands. The important parts are at the center, where the wavelengths are very short. At the edges, we have a lot of flexibility because the wavelengths are so long.

I do not believe it's possible to create a believable illusion that the band is in front of you with two drivers radiating from two separate points in space. To really get the illusion right, you need to 'trick' the brain into thinking you're hearing one speaker. Even the 'easy' solution noted at the top of this post does this though; it leverages the fact that we're more sensitive at some frequencies than others, and the fact that you can create a believable illusion 80% of the time if you simply get the midrange right.

Get the midrange wrong and your stereo will not sound believable, IMHO

 

·
Registered
Joined
·
162 Posts
PLD stands for Pathlength difference.
Pathlength: means we measure the distance from the speaker to your head or ear.
Difference: means you subtract the difference between the two speakers measured.
In most cases people are measuring pairs of midrange drivers.


We measure the distance from the headrest to the drivers side kick panel. 55.5 inches

Next we measure the distance to the passengers side kick panel.
65.5 inches

The pathlength difference for kick panels is 66.5" - 55.5" =10"

Hope this helps. :)
 

·
Registered
Joined
·
204 Posts
Some good options there.
I have tried the Midbass in the doors, an L3SE midrange in the kicks and tweeters in various positions. Crossing the Midbass LP over 250hz made the midbass locatable and dragged the stage right (drivers side).
Having the midrange cover 200 to 7k in the kicks was great but stage was still low in the kicks with tweeters but i need to aim more i think.
Im just having difficulty blending 3 ways with no TA. Going 2 way give a much more coherent sound.
I think futher testing is needed. ;)
 

·
Premium Member
Joined
·
6,045 Posts
Oh, I know you guys are are talking about PLD now - the distance between tweeter and mid drivers on one side.

Patrick, it's true that crossovers cause problems of their own, as you well know. That's why some guys like a fullrange driver. But the fullrange introduces new problems of it's own - it's hard to reproduce a wide band of frequencies from a single dynamic loudspeaker driver. Specifically, treble distortion usually goes up with wideband drivers compared to a convention two-way with a dome tweeter.
Watch this video, and tell me -
Can you hear the distortion?

FF85WK - YouTube

That woofer is under $40. The distortion on it is likely terrible, as it has virtually no excursion. IMHO, distortion isn't as audible as people like to believe. More importantly, it's not *offensive*. IE, there are tons of horribly distorted loudspeakers which sound pretty darn nice.

I think the reason that the $40 Fostex 'works' is because it satisfies the 'easy' criteria that I outlined in the previous post. It's anything but hitech, but it gets the midrange right.

It may also depend on what type of sound the user likes also, what their reference for great sound is. Some people like a "natural" sound like unamplified live instruments can give. A wideband driver might be better for that. I, on the other hand, like the sound of high end headphones, and this is the sound I strive for in the car, but with a front sound stage and subbass. I don't really consider this sound natural at all though, it's just my preference.
 

·
Registered
Joined
·
464 Posts
Discussion Starter #17
Patrick, I'm just doing a simple two way component speaker set front and rear in an old 96' Firebird. I have about 5 sets new old stock DEI Vipers and PPI DCX comps. Subwoofer box in the rear well with comps chambered, flanking a 15" sub. Rear comps will get a rear surround signal via Sony head unit with Dolby Prologic. I'm not sure which set I will go with yet, I have too many.

My old rule of thumb with two-way comps was to mount the tweeter close to the mid. I don't remember why, LOL! I'm old, old school, and I don't think about car audio that much anymore. I have forgotten some things.
 

·
Premium Member
Joined
·
6,045 Posts
Patrick, I'm just doing a simple two way component speaker set front and rear in an old 96' Firebird. I have about 5 sets new old stock DEI Vipers and PPI DCX comps. Subwoofer box in the rear well with comps chambered, flanking a 15" sub. Rear comps will get a rear surround signal via Sony head unit with Dolby Prologic. I'm not sure which set I will go with yet, I have too many.

My old rule of thumb with two-way comps was to mount the tweeter close to the mid. I don't remember why, LOL! I'm old, old school, and I don't think about car audio that much anymore. I have forgotten some things.
Does it have a neodymium tweet?

If so I'd just build a simple bridge across the woofer and do a DIY coaxial.

Like this:

 

·
Registered
Joined
·
464 Posts
Discussion Starter #19
PLD stands for Pathlength difference.
Pathlength: means we measure the distance from the speaker to your head or ear.
Difference: means you subtract the difference between the two speakers measured.
In most cases people are measuring pairs of midrange drivers.


We measure the distance from the headrest to the drivers side kick panel. 55.5 inches

Next we measure the distance to the passengers side kick panel.
65.5 inches

The pathlength difference for kick panels is 66.5" - 55.5" =10"

Hope this helps. :)

- Oh, I don't like that at all! That might be fine for determining placement when speakers are on the extreme sides of the car, such as kicks, doors, or sails, or pillars, but if you are talking about dash speakers or horns under the dash or perhaps even floor speakers, everything changes... what is more important and more accurate is the ratio of the distance of the left and right speakers, to your distance from their axis. There is a threshold of how close the speakers need to be to each other to get a good solid phantom center from a stereo pair, while maintaining a reasonably wide sound stage. Things can change dramatically when you start moving speakers closer together.
 

·
Premium Member
Joined
·
11,422 Posts
It's not phase in the midbass that causes a collapse of the stage when a speaker is directly in front of you and one is to the side.

It's the interaural time difference...or complete lack of ITD that causes the width to collapse.
 
1 - 20 of 32 Posts
Top