• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Deep dive: Questions after reading Toole and Geddes for stereo setup.

johnp98

Active Member
Joined
Jan 29, 2021
Messages
133
Likes
201
After reading the books by Toole (Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms) and Geddes (Premium Home Theater : Design & Construction) I have a few residual questions. It is very likely that I did not comprehended 100% of what I read and that these leftover thoughts and questions have been answered but I was looking for further direction and to start a conversation about peoples thoughts on such issues.


I am looking at most things through the lens of stereo music listening and soundstage and specifically focusing on setting up the L&R speakers (as I firmly believe that the best way to set up the L&R channels is for stereo music listening, and then if one wants to integrate that into a 5.1 or home theater setup they will be already optimized and in the right position). So I am wondering specifically about setting up the L&R channels and specifically imaging / soundstage and creating a wide sweet spot.


Here is what I would like to discuss:



1) Speaker positioning from the front wall and soundstage depth:


There was not too much discussion in terms of ideal positioning from the front wall outside of knowing about speaker boundary interference and adjacent boundary effects and how it will affect the bass response. It seems that bass response is the main consideration with positioning the speaker in/flush/in front/away from the front wall. Toole says “Adding absorption to the front wall, behind the loudspeakers, reportedly improved image localization and reduced coloration... Memo for Listening room recommendations: add sound absorbing material to front wall”

It seems the general audiophile advice is to pull speakers away from the front wall to help create a deeper soundstage etc yet I did not see this explicitly mentioned by Toole or Geddes. They seem to be ok with positioning the L&R speakers close to the front wall as long as one pays attention to bass response and possibly absorbs.

What I am wondering is if there truly is a benefit in soundstage depth with pulling the speakers away from the front wall?

Or do people think it is a psychoacoustic effect by having the front wall further away from the source (not saying that it is not a very real psychological effect, but merely one that cannot be seen on the frequency response, nor reproducible with blind tests)?

I wonder if this explains it: “Generations of listeners have noted the obvious differences in directional and spatial impressions created by sounds panned to the real left and right loudspeakers and those panned to intermediate positions, including center. The difference is that the extreme left and right locations are created by monophonic signals, delivered to single loudspeakers, whereas the intermediate image locations result from “stereo” signals, delivered to both loudspeakers simultaneously, with amplitude biases and/or delays appropriate to define the direction. The common impression is that the left and right panned sounds appear to originate in the loudspeakers themselves, whereas the intermediate images appear to originate further back, in a more spacious setting, and sometimes elevated. Instead of a soundstage extending across a line between the loudspeakers, the center images tend to drift back-ward.” - Toole

As then if the sound is hard panned to the L or R then it comes further forward (from the speaker), and then if intermediate positions then it appears higher up and further back and hence creates the impression of depth? Do people think that explains what we observe / hear?

What are peoples thoughts on the optimal distance from the front wall and imaging dept?



2) Speaker positioning from the side wall and side reflections:


Despite popular belief, it seems that the side wall reflections are important and desired. Specifically those that can come from wide angles (60°) as Toole writes “Start to think in terms of “preference,” “spaciousness,” “low interaural cross-correlation (IACC),” and “lateral reflections” as positively correlated with each other…. For maximum “preference” from the Ando (1977) data, it seems that reflections from about 30° to 90° are most effective. When IACC is measured, a broad minimum is seen around 60°, corresponding to a maximum in the preference ratings. Preference, therefore, is associated with low interaural cross-correlation….IACC exhibits a broad minimum around 60°.”

So should this goal of 60° determine the side wall distance goals? I think a lot of people would have a hard time getting 60° reflections from the near side wall but would be getting these angles from the far side wall (see image… as near wall reflection would be 30°+16° = 46° and far wall would be 97°-30° = 67°) .

1613854731021.png


Should we be setting up our speakers to ensure ~60° angle of the near wall reflections? If we do not have enough lateral room for this should this call into question the orientation of our rooms (having the width of the room be larger than the length if it allows for 60 angle reflection)?



3) Precedence Effect and side reflections

I am sure I just missed something somewhere when it comes to the precedence effect. But my understanding is that the primary sound will be heard first and that all reflected sounds will be combined into the first arriving sound. My understanding is that this is commonly misunderstood to mean that the delayed signals are ignored which is certainly not the case, as they are combined with the original signal, but that the original signal has all these other delayed signals added into the interpretation of it.

Toole mentions that delays greater than 30-40ms can then be interpreted as separate sources and thus everything under falls into the precedence effect does it not? So if the signals are being combined to that first signal, then does the distance of reflections matter? Does it matter if the side walls are mere inches vs feet vs meters away? Should it not all be interpreted under the primary signal? If that is the case, then would not all reflections in small rooms (under 30ms), be of no consequence when it comes to imaging? Once again, I am sure I am missing something here, but I like to think I understood that the delayed signal is combined with the original (and not ignored), but that the original signal trumps the delayed signal in terms of localization etc. But now it has me questioning the effects of all reflections and if they truly are damaging to soundstage.



4) Tooles low interaural cross-correlation (IACC) for imaging vs Geddes time intensity trading for image stability?

It seems like there are two opposing forces at play when it comes to toeing in of the speakers.

If you want truly low IACC and to isolate the L&R signals to the L&R ears (as used to the extreme with Ambiophonics or placing a mattress in between the speakers and up to your face) then one would not want to toe in the speakers to an extreme (at maximum aiming directly at the listening position, but probably further outward as ones off axis response allows). This should help with the soundstage and imaging. So I wonder if absorbing the far wall reflections (even though they are at a wider angle and thus less IACC, would help with soundstage would it not (although maybe at the cost of envelopment / spaciousness)

If you want a more stable “sweet spot” then using time intensity trading and aggressively toeing in the speakers as recommended by Geddes sure makes a lot of sense, and is easily experienced when I have personally done it. But is this improved sweet spot and image stability coming at the cost of a larger soundstage? I have not been able to definitively test this as I am not able to do blind tests myself, but I wonder what people impression is.



5) Finding the sweet spot:


I have never found an objective way to find the sweet spot for speakers until I read this from Toole:

“On the matter of the sound quality of the center phantom image in stereo, I recommend a simple experiment: Arrange for monophonic pink noise to be delivered to both loudspeakers. When seated in the symmetrical sweet spot, this should create a well-defi ned center image midway between the loudspeakers. If it does not, something is seriously wrong. If it does, consider what you hear as you lean very slightly to the left and to the right of the symmetrical axis. The timbre of the noise changes and more obviously the closer you sit to the loudspeakers. In fact, it is possible to fi nd the exact sweet spot by simply listening to when the sound is dullest. Moving even slightly left or right of the sweet spot causes the sound to get audibly brighter; there is more treble. It is much more exact to find the sweet spot by listening to the timbre change than by trying to judge when the center image is precisely localized in the center position. There is nothing faulty with the equipment or setup; this is simply stereo as it is—flawed…. Figure 9.7d shows the difference between the curves, revealing the result of acoustical interference. This can be confirmed by a simple calculation. The time differential between the ears for a sound source at 30° away from the frontal axis is about 0.27 ms for an average head. A destructive acoustical interference will occur at the frequency at which this is one-half of a period: 1.85 kHz. It won’t be a perfect cancellation because of a tiny propagation loss and a signifi cant diffraction effect. The wavelength is just over 7 in. (178 mm), which, because it is similar in dimension to the head, will experience a substantial head-shadowing effect at the ear opposite to the sound source. There will be an interaural amplitude difference of the order of 6 dB in this frequency range”

1613854746891.png


Does this seem like the best way to find and judge the sweet spot in ones room?

In the past I have also used repeating short clips from songs recorded in QSound (I have taken a few seconds from the openings to a song in Amused to Death and made a loop on perpetual repat and then inverted the R&L channels to test each speaker and its positioning). Has anyone else tried something similar or what do people think is the best way to test imaging (outside of grossly assessing soundstage from well-known songs)?

If anyone wants to chat and discuss any of the above topics that would be much appreciated!

Thanks!
 

test1223

Addicted to Fun and Learning
Joined
Jan 10, 2020
Messages
503
Likes
511
Hey,

great topic.

1. Front wall distance:
As you mentioned you have to separate the bass and everything above. For everything above the middle reflection of the front wall helps to establish a phantom center. And higher distance to the front wall can cause a deeper image. Early front wall reflections are not good for higher frequencies but are very good for the bass since it increases the directivity in the lower frequencies.
A good distance is about 1.5m.

2. Side wall reflections of 60° are good if they aren't to early, about 2m distance to the side walls or more. They provide envelopment. On a time scale one peak, the direct sound, should arrive and then no reflections should occur after a few milli seconds the "good" front wall and 60° side wall reflections should arrive. All later reflections should be ideally diffuse.

3. You have room information in the recording and add room information with your room. My guess is that the information of the recording are best transported if you have no early reflections. You only need later diffuse reflections to get the envelopment. But with no front wall reflection you have a more unstable center image. And with no side wall reflections you have less envelopment. And it is very hard to get a diffuse sound field in a listening room.

4. Interesting thoughts I think you are right.

5. Pink noise has some advantages for testing. And with some songs you can also test. I am not aware of a better method. Some impulse sounds will be a hard to master test.

X. I would also add the importance of the right amount of direct sound to later reflections which is important. Maybe this is of interest.

Best
Thomas
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Hi, I made some comments that might be helpful, but I have a very low level understanding of the subject.

1) Speaker positioning from the front wall and soundstage depth:

I asked Toole for his opinion about this, and I think his response addresses your question: https://www.avsforum.com/threads/ho...-what-the-science-shows.3038828/post-57878964

For a different perspective, check here for Linkwitz's view: https://linkwitzlab.com/frontiers.htm. I don't know why he felt that 6 ms would be a sufficient delay for summation not to occur.

2) Speaker positioning from the side wall and side reflections:

Despite popular belief, it seems that the side wall reflections are important and desired. Specifically those that can come from wide angles (60°) as Toole writes “Start to think in terms of “preference,” “spaciousness,” “low interaural cross-correlation (IACC),” and “lateral reflections” as positively correlated with each other…. For maximum “preference” from the Ando (1977) data, it seems that reflections from about 30° to 90° are most effective. When IACC is measured, a broad minimum is seen around 60°, corresponding to a maximum in the preference ratings. Preference, therefore, is associated with low interaural cross-correlation….IACC exhibits a broad minimum around 60°.”

So should this goal of 60° determine the side wall distance goals?
...
Should we be setting up our speakers to ensure ~60° angle of the near wall reflections? If we do not have enough lateral room for this should this call into question the orientation of our rooms (having the width of the room be larger than the length if it allows for 60 angle reflection)?

No, look beyond the text at the relevant figure again to see that the IACC minimum depended somewhat on the musical motif (~50 degrees for music motif A, centered a bit less than 60 degrees for music motif B), also that preference seemed to continue to increase modestly beyond that. The range of variance in IACC and preference is pretty modest between ~35 and 70 degrees.

Most importantly, some listeners are particularly sensitive to lateral reflections, so individual preference should be considered. I tried to outline some thoughts along these lines here: https://www.audiosciencereview.com/...-room-reverberation-and-a-few-thoughts.16806/

3) Precedence Effect and side reflections

I am sure I just missed something somewhere when it comes to the precedence effect. But my understanding is that the primary sound will be heard first and that all reflected sounds will be combined into the first arriving sound. My understanding is that this is commonly misunderstood to mean that the delayed signals are ignored which is certainly not the case, as they are combined with the original signal, but that the original signal has all these other delayed signals added into the interpretation of it.

Toole mentions that delays greater than 30-40ms can then be interpreted as separate sources and thus everything under falls into the precedence effect does it not? So if the signals are being combined to that first signal, then does the distance of reflections matter? Does it matter if the side walls are mere inches vs feet vs meters away? Should it not all be interpreted under the primary signal? If that is the case, then would not all reflections in small rooms (under 30ms), be of no consequence when it comes to imaging? Once again, I am sure I am missing something here, but I like to think I understood that the delayed signal is combined with the original (and not ignored), but that the original signal trumps the delayed signal in terms of localization etc. But now it has me questioning the effects of all reflections and if they truly are damaging to soundstage.

I wonder if you're conflating timbral and spatial effects of reflections. Check Toole's quoting of Benade's summary of "Generalized Precedence Effect," specifically "provided (a) that these sequences are reasonably similar in their spectral and temporal patterns and (b) that most of them arrive within a time interval of about 40 ms following the arrival of the first member of the set." I don't know what you mean about the distance of reflections, but a proximal boundary inches away would significantly affect the timbre of the early sound arriving at the ear. Also, keep in mind that conventional loudspeakers are progressively forward-radiating at higher frequencies, so the timbre of reflections is likely to vary greatly, depending on the the orientation of the loudspeaker and relevant boundary in question.

4) Tooles low interaural cross-correlation (IACC) for imaging vs Geddes time intensity trading for image stability?

It seems like there are two opposing forces at play when it comes to toeing in of the speakers.

If you want truly low IACC and to isolate the L&R signals to the L&R ears (as used to the extreme with Ambiophonics or placing a mattress in between the speakers and up to your face) then one would not want to toe in the speakers to an extreme (at maximum aiming directly at the listening position, but probably further outward as ones off axis response allows). This should help with the soundstage and imaging. So I wonder if absorbing the far wall reflections (even though they are at a wider angle and thus less IACC, would help with soundstage would it not (although maybe at the cost of envelopment / spaciousness)

If you want a more stable “sweet spot” then using time intensity trading and aggressively toeing in the speakers as recommended by Geddes sure makes a lot of sense, and is easily experienced when I have personally done it. But is this improved sweet spot and image stability coming at the cost of a larger soundstage? I have not been able to definitively test this as I am not able to do blind tests myself, but I wonder what people impression is.

I wonder if it's a little more complicated, since most conventional loudspeakers are designed to be listened to on- or nearly on-axis (Geddes' speakers and the Gradient Helsinki 1.5 being notable exceptions, being designed to be listened to 22 and 15 degrees off-axis, respectively), and Geddes' recommendations depended very much on specific room setups, including the use of constant (not just controlled) directivity loudspeakers for time-intensity trading. Less toe-in is likely to result in image broadening or increased ASW, as well as perception of a wider soundstage. Aggressive toeing-in is likely to contribute to later reflections, which may contribute to the perception of envelopment in a directionally treated room.

5) Finding the sweet spot:

I have never found an objective way to find the sweet spot for speakers until I read this from Toole:
...
Does this seem like the best way to find and judge the sweet spot in ones room?

This is the so-called Shirley curve, so my understanding is that it should be basically be maximized in any highly symmetric setups where early reflections are suppressed and the listener is positioned the midline between the speakers, but this wouldn't tell you where along that midline would be deemed to be optimal.

Young-Ho
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Interesting topic which I have investigated during a couple of years.
short story - bringing the loudspeakers away from the wall behind them creates a dip and a peak in the frequency response thats gonna change your perceived sound quality. With the right distance, an illusion of depth can be had, though its just an illusion, it can sometimes help compensating for the very crude and primitive coding of the 2- channel system.
So theres nothing that says that damping the front wall always will be considered as a better sound.

The best thing to do, is to do your own recording ( purist 2-channel ) and compare with the real acoustic event. 2-channel recordings are just an illusion and nothing more , of the real acoustic event.
Knowing this, after making more than 20 recordings with good microphones, one has to understand that there is no easy fix with perfect sound with two chanels. You have to work with the illusion and the perceived sound quality.

That said, different loudspeaker positions regarding distance from the front wall can really make big differences in perceived quality.
But, there is no quick fix and no solutions thats gonna be the same everywhere.

Regarding the precedence-effect : after 5 ms soundtravel and more, the microphone and ear/brain are starting to function very differently. The mic takes up all the sound, the brain selects sounds.
 
Last edited:

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
1. Front wall distance:
[snip]
A good distance is about 1.5m.

So stick the speakers in the middle of the room! I'm exaggerating, but for some of the rooms described here, it's not that far off.
 

test1223

Addicted to Fun and Learning
Joined
Jan 10, 2020
Messages
503
Likes
511
So stick the speakers in the middle of the room! I'm exaggerating, but for some of the rooms described here, it's not that far off.
Yes with small rooms this is the only solution for getting good sound. In the middle everything symmetrical and almost omnidirectional speaker with a short listening distance.


Edit: the center image is weak and moving your head is bad though.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
Edit: the center image is weak and moving your head is bad though.

Are you referring to the comments on toe-in? I always prefer a strong center image to wider soundstage.

I did try putting my Buchardt S400s so the baffle is 1.5 m from my "front wall" (actually 2 walls at an obtuse angle), and the imaging is indeed more precise, but I have to move the speakers a bit closer (from 8 feet to 7 feet) so I don't have to move the listening chair too far back, and there's some loss of bass reinforcement. I haven't measured yet to see how much bass is lost (it's not too bad subjectively, but I don't listen to music with much in the way of sub-bass frequencies.) Of course, I can only get away with this because I live alone.
 
Last edited:

test1223

Addicted to Fun and Learning
Joined
Jan 10, 2020
Messages
503
Likes
511
Are you referring to the comments on toe-in? I always prefer a strong center image to wider soundstage.

I did try putting my Buchardt S400s so the baffle is 1.5 m from my "front wall" (actually 2 walls at an obtuse angle), and the imaging is indeed more precise, but I have to move the speakers a bit closer (from 8 feet to 7 feet) so I don't have to move the listening chair too far back, and there's some loss of bass reinforcement. I haven't measured yet to see how much bass is lost (it's not too bad subjectively, but I don't listen to music with much in the way of sub-bass frequencies.) Of course, I can only get away with this because I live alone.
Yes that is the usual experience with trying to find the best placement. A good placement for the bass and for the higher frequencies are not the same.
Try a closer listening distance and experimenting with the angle of the stereo triangle about 40° to 60° should provide a sweet spot where you also get envelopment which is very nice (a more omnidirectional speaker would provide better results regarding envelopment). About 1.2m listening distance should be very good for the S400.

Usually the center is more stable if you move the speaker closer to the front wall but you loose over all exactness. A smaller angle of the stereo triangle will also give a better center, but you loose envelopment and the size of the image.

A heavy toe in can help to get a better center with the cost of a more tilted tonality. There is also the issue that the listening window is not as good with a heavy toe in, which will decrease the performance and head movements are worse.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
Yes that is the usual experience with trying to find the best placement. A good placement for the bass and for the higher frequencies are not the same.

It also occurs to me that 1.5 m is smack dab in the middle of the "bad" range of 1.0 to 2.2 m for SBIR (Speaker Boundary Interference AKA Allison effect):

http://arqen.com/acoustics-101/speaker-placement-boundary-interference/

I'll try to do some measurements to see if I get any bad dips.

About 1.2m listening distance should be very good for the S400.

I've never considered a listening distance that close (this setup for relaxing, not monitoring). Maybe I'll try it in the interest of science, but I'd be afraid that it would sound very "heady".
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Thanks for the information! Certainly very appreciated!

I asked the same question at diyAudio as Dr. Geddes spends time on there, and he was kind enough to give his thoughts!
https://www.diyaudio.com/forums/roo...ng-toole-geddes-stereo-setup.html#post6541971

Thanks for linking to that thread. You're working with a very narrow room, so you're going to want to heavily damp the sidewall closest to each speaker. Usual recommendation for sweet spot setup for stereo listening is at the apex of an equilateral triangle where the base is between the two speakers, though accommodations may need to be made (also, Blumlein recordings "should" be reproduced with wider spacing to 90 degrees, not 60 degrees). The section from Toole that you quote above only helps with putting you equidistant from each speaker but would remain similarly relevant at different angles.
 
Top Bottom