Floyd Toole: Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms (AES Presents) 3rd Edition - Discussion Thread

[Sonnie Parker]

With my Magneplanar 3.7's I started with the Cardas recommendations and went from there. I tried a lot of absorption initially with 3" Sonex on the front wall and up the sides, then when diffusers of various types became available I experimented with some thin 1D diffusers initially and from that point on the absorption came down and more diffusers went up. I settled on the Vicoustic Multifuser and have lots of them on the front wall and four feet up the sides. When Wendell Diller did his 30.7 North America demo tour a couple years ago he said the best sounding rooms had lots of diffusion.

My room is 19.5' wide by 23.5' deep and I have removed all of my front wall and side wall absorption. I have some diffusion on front wall (more on order), and more absorption on the back wall. My MartinLogan 15A's are about 60" from the front and side wall.

Link?
FYI, you do know he tested (at NRC), was suitably impressed and owned Bipolar speakers, prior to his forward firing Salons, yes?
They also tested Quads and at least one, older ML.

cheers

Is this testing publicly available... do you know what the bipolars were? Just curious, not that am getting rid of my ML's.

 

[Marc Lombardi]

Link?
FYI, you do know he tested (at NRC), was suitably impressed and owned Bipolar speakers, prior to his forward firing Salons, yes?
They also tested Quads and at least one, older ML.

cheers

I was referring to the work done at Harman and the follow on work that Olive and others published in AES papers.

 

[Marc Lombardi]

I would love to see the reference to the recordings used in some of the testing. I've never been able to find that information.

In the 2017 study published in AES Convention Paper 9855 (after Toole's era) they reference Steely Dan Cousin Dupree as being from previous studies. In this study they used the following in addition: DeadMau5 Ghosts n Stuff; Drake One Dance; Emmylou Harris Black Caffeine; Bruno Mars Uptown Funk; Norah Jones It's a Wonderful Time for Love.

I can't put my finger on the papers from the previous studies.

 

[natty]

I guess it’s worth pointing out that Toole sees value in not always having side wall reflections, and that it largely depends on the speakers in use. From his book, almost like he anticipated this discussion (maybe because it has happened so often?):

7.1 The Physical Variables: Early Reflections

Since the appearance of the first edition of this book, a few people took the position that “Toole” is in favor of side-wall reflections, and let their views be known in various Internet audio forums. I became the “straw man” opponent to their views that, not surprisingly, have been known to encourage the purchase of acoustical products they sell. In fact, if my critics had read to the end of the book, figure 22.3 shows some suggested room treatments, and it is stated that acoustical treatment of those contentious side-wall areas is “optional: absorb, diffuse, reflect”—the choice is left to the designer, with, one would hope, input from the customer.

It has been alleged that I have this commitment to lateral reflections in ignorance of, or dismissal of, decades of professional audio tradition in which these reflections are absorbed in recording control rooms. These people don’t know that early in my career I designed a few recording studios as a learning exercise. One of them could hold a 75-piece orchestra, had separate vocal and drum booths, and two control rooms. Many of the high-power monitor loudspeakers at the time were less than impressive, so I also designed the monitor loudspeakers that went into a couple of them—see Section 18.3.1. In those days the powerful main monitor loudspeakers were moderately directional mid- and high-frequency horns, and side walls were usually angled to direct the residual first lateral reflections into the broadband back wall absorber. Recording engineers preferred to be in a strong direct sound field, and that is what they got. Some control rooms of that period had a “dead end” that absorbed all early reflections in the front of the room. These practices in control room design evolved at a time when high-power monitor loudspeakers were simply not very good. Some misbehaved so badly off-axis (see Figures 12.8 and 18.5) that the only way to alleviate the problem was to absorb the off-axis sounds.

However, as a research scientist, I saw a number of interesting questions to be answered, and so, it turns out, did several other investigators over the years. Toole (2003) gives an updated opinion of what I thought might be useful guidelines for control room practice, and Toole (2015a) is an invited “distinguished lecture” to an audience that included recording engineers and tonmeister students at CIRMMT, the Centre for Interdisciplinary Research in Music Media and Technology, housed at the Schulich School of Music at McGill University (available for viewing on YouTube). All of this is also discussed in detail in several chapters of the original book, and a revised and expanded version follows.

It is not a black and white situation, and whatever personal opinions I may have, they are overwhelmed by experimental data from others; I am the messenger. So, read on and draw your own conclusions. It may be that “one size does not fit all.”
Before getting into experimental evidence, it is important to understand the variables involved. Our focus will be on the first reflections because these are the second most energetic sounds to reach the listener.

And by the way that YouTube video is worth watching. Sure wish more of his lectures were available.

 

[natty]

I suppose his reputation is based in part on observations like this one here, based on other people’s research that he discusses. In this example, he points out that when absorbing side wall reflections, comb filtering in stereo listening creates a null around 2khz.

However, when the side wall reflections are allowed to exist, those reflections add to the total sound hitting the listener in ways that can fill in the null. (Message people take away: Toole loves side wall reflections!)

Actually he is really using this data (not his studies but a half dozen other peoples work) to argue for using a center channel, which solves it and solves other things as well.

But it is also a good reminder how side wall absorption can hurt stereo listening.

 

[natty]

I may be wrong. But I've read that "resistance" absorbers like 703 are "velocity" based, and so have their greatest effect where velocity is the greatest. Velocity is zero at the wall (pressure is at its max at the wall). Velocity is greatest at the quarter wavelength of the frequency in question. Maybe it's more relevant in absorbing modal resonances than a first reflection of the low frequency wave from the speaker. But then below the transition frequency, and steady state, how would we distinguish first reflection from modal resonance? The answer is .... it should be measurable, right?

Well, in a sense the diagrams from Owens Corning are probably optimistic for how the panels of fiberglass act when used for absorption in a small room.

Toole shares data (again not his) that compares them (industry standard ways of measuring, focused on their use in large spaces) to real world measurements in small rooms.

But crucially he also notes that even when taking into account this discrepency:

Thicker material will move the curves downward in frequency, thereby absorbing sound over more of the frequency range. Fibrous absorbing panels that are 3 inches (75 mm) or more in thickness should usefully attenuate frequencies down to the transition frequency, below which we enter the room-mode.

 

[natty]

In the 2017 study published in AES Convention Paper 9855 (after Toole's era) they reference Steely Dan Cousin Dupree as being from previous studies. In this study they used the following in addition: DeadMau5 Ghosts n Stuff; Drake One Dance; Emmylou Harris Black Caffeine; Bruno Mars Uptown Funk; Norah Jones It's a Wonderful Time for Love.

I can't put my finger on the papers from the previous studies.

Found this. Looks like the 1985/6 study used several recordings of natural sounds/acoustic music in real space:

 

[AJ Soundfield]

I was referring to the work done at Harman and the follow on work that Olive and others published in AES papers.

I know that work. I'm referring to your quote where you read him in a recent (presumably online forum) thread.

In particular I recently read a comment like that where he was discussing the listening tests

 

[Marc Lombardi]

I know that work. I'm referring to your quote where you read him in a recent (presumably online forum) thread.

Oh okay .... this one ... https://www.audiosciencereview.com/...ut-room-curve-targets-room-eq-and-more.10950/

BTW, I was looking for evidence on the use of the "Harman Curves" as EQ targets for speakers. I have yet to find what I'm looking for, and in fact in this series of posts Toole specifically says the curves should not be used as EQ targets.

I've seen people use these curves which have been translated to text files as targets for Dirac Live. In fact every time I report an issue to Dirac and send them a project file (they are very good at responding) they say ... oh by the way we notice you use flat target curves and we really recommend you use the Harman +4 or +6. I ignore them.

 

[AJ Soundfield]

Is this testing publicly available... do you know what the bipolars were? Just curious, not that am getting rid of my ML's.

I can't recall what's in the book vs my communications with him, but here are his former Mirage M1 speakers/room.
I'd have to dig for the Quad and ML tests, but they're are out there on forums. Per my communications with him, the ML he tested was an older one with resonances, prior to ML being acquired, which he stated lead to better engineered products which would probably test better, but he was out of testing by then.

 

[natty]

There are some Quad tests in his 1985/86 AES papers, one page of which I shared, above. I think these were done at the Canada national labs rather than at Harman.

 

[AJ Soundfield]

Oh okay .... this one ... https://www.audiosciencereview.com/...ut-room-curve-targets-room-eq-and-more.10950/

Ahh, ok. The first post is good, will read thread when possible. I presume Floyd chimes in at some point. I have been trying in vain to explain these things to audiophiles/studiophiles on audio forums forever, but that usually ends poorly.

BTW, I was looking for evidence on the use of the "Harman Curves" as EQ targets for speakers. I have yet to find what I'm looking for, and in fact in this series of posts Toole specifically says the curves should not be used as EQ targets.

I've seen people use these curves which have been translated to text files as targets for Dirac Live. In fact every time I report an issue to Dirac and send them a project file (they are very good at responding) they say ... oh by the way we notice you use flat target curves and we really recommend you use the Harman +4 or +6. I ignore them.

Yes, I'm usually ignored or branded a heretic, but here goes: https://www.avnirvana.com/threads/harman-curve-and-rew.8131/#post-61365
https://www.avnirvana.com/threads/avr-for-2-ch-and-ht.8204/page-2#post-62161
https://www.avnirvana.com/threads/t...ch-barnett-of-accurate-sound.8270/#post-62554
etc, etc, etc.

cheers

 

[AJ Soundfield]

But it is also a good reminder how side wall absorption can hurt stereo listening.

Toole did his own studies on what "normal"/everyday listeners prefer regarding absorb/diffuse/untreated domestic living space sidewalls (in book) and of course the overwhelming majority preferred the latter...blind of course.
This created quite a backlash by (perceptual evidence free) studiophile believers as I noted here, insisting that their self trained hearing would produce different results (as if that mattered, since the "normal"/average end consumer is not a studiophile).
Well, McGill U said, ok, lets test that
https://www.aes.org/e-lib/browse.cfm?elib=16640

In contrast to earlier studies that used normal listeners, this study uses trained audio engineers to perform selected tasks in a variety of acoustic settings.

The results were very funny.

cheers

 

[Marc Lombardi]

Ahh, ok. The first post is good, will read thread when possible. I presume Floyd chimes in at some point. I have been trying in vain to explain these things to audiophiles/studiophiles on audio forums forever, but that usually ends poorly.



Yes, I'm usually ignored or branded a heretic, but here goes: https://www.avnirvana.com/threads/harman-curve-and-rew.8131/#post-61365
https://www.avnirvana.com/threads/avr-for-2-ch-and-ht.8204/page-2#post-62161
https://www.avnirvana.com/threads/t...ch-barnett-of-accurate-sound.8270/#post-62554
etc, etc, etc.

cheers

I think the posts in that link are reposts of Floyd's posts, aggregated into one place.

I'm not saying the studies and data were not useful, at least to Harman. They're in the business of trying to correlate consumer preferences to speaker design. But #1, these curves represent the personal preferences of some bunch of people listening to music that I may or may not (mostly not) listen to myself, in a room that does not resemble mine; #2, the general public has been coaxed by many industry "professionals" to use the curves for a purpose that they were never intended.

What gets frustrating for me when I try - as I have recently - to coach people through Dirac calibrations, is that people use Dirac for its primary intended purpose which is to correct modal resonances, then they complain that Dirac "killed the bass", and are encouraged by others (including Dirac Support) to put the boom back in the room ... justifying the Harman Curves as the science behind the recommendation.

But, looks like it's just you and me AJ

 

[natty]

To be fair other folks like a tipped up bass response too eg ARC seems to do this and perceptual loudness studies support the idea that one isn’t hearing the intended bass relative to the rest of the audio range when listening below reference leave to movie soundtracks.

 

[natty]

Toole did his own studies on what "normal"/everyday listeners prefer regarding absorb/diffuse/untreated domestic living space sidewalls (in book) and of course the overwhelming majority preferred the latter...blind of course.
This created quite a backlash by (perceptual evidence free) studiophile believers as I noted here, insisting that their self trained hearing would produce different results (as if that mattered, since the "normal"/average end consumer is not a studiophile).
Well, McGill U said, ok, lets test that
https://www.aes.org/e-lib/browse.cfm?elib=16640

The results were very funny.

cheers

If I could qualify for the paywall barrier I could share in the joke I guess?

 

[Marc Lombardi]

To be fair other folks like a tipped up bass response too eg ARC seems to do this and perceptual loudness studies support the idea that one isn’t hearing the intended bass relative to the rest of the audio range when listening below reference leave to movie soundtracks.

Absolutely! If you are listening to anything below reference levels then a Fletcher-Munson-esque loudness compensation would be appropriate. I see people mixing this up, and suggesting you need loudness compensation all the time ... and okay if someone wants more bass fine, but there's no psychoacoustic justification. Now the question: what is reference level? 85db average? A or C weighted?

 

[natty]

“Average“ can be abused. Supposed to refer to average dialogue apparently.

Really it’s that pink noise noise at -20db below full volume, 0dbfs in the digital scale, should be 85 db at the listening position, iirc.

C weighted is the definition.

Nice discussion here including the comments: http://www.acousticfrontiers.com/2013314thx-reference-level/

Reference level for all channels except low frequency effects is calibrated by adjusting the audio chain such that a pink noise signal recorded at -20dB relative to full scale (0dB) creates 85dB sound pressure level as measured with a C weighted SPL meter at the seating locations. Volume levels are adjusted for each channel individually until they read 85dB. The master volume control setting associated with this playback level is then set to a nominal 0dB, or reference level. The history behind this is that sound engineers and producers generally work so that the average recording level for dialog in movie soundtrack is -20dB.

Unfortunately the same perceptual curve that works for much content mastered that way does not work for lots of music recordings (especially with the loudness wars).

 

[Sonnie Parker]

I'm also the same as 95% of the other folks in the world with a subwoofer in their system. The flat response that most of these DSP systems create does make you come away believing you have no bass... not sure how anyone can listen to it myself (but perhaps I've trained my brain to believe that's how it should be). When working with SVSound, we would have had every subwoofer we sold returned if we did not advise the customer to turn up their bass on the sub after their DSP system flattened out the response. I can't even begin to count the number of support calls I handled EVERY SINGLE DAY of people complaining they had no bass in their system.

So Toole was really keen on the Mirage M1 speakers, which are apparently no longer made... don't even think Mirage is manufacturing speakers now. What are his favorite music listening speakers now? I've seen his living room theater, but I'd be surprised if he is using that room for music with those speakers being wall mounted. I'd be very interested in comparing and measuring (with my mic that doesn't really show me what I'm hearing) some super neutral speakers that I have a anechoic measurement to start with, then compare those to my ML 15A's. That might give me an idea how far off I am. However, because I really like the sound of my ML's, my brain has trained me to like that sound, if the neutral speakers don't sound like this, I won't like them.

I'm also curious as to what mic he is using to record his measurements (he may mention this in his book, but I haven't see it yet). After all... apparently measuring with a mic does not really represent what we are actually hearing, so what mic does he use and how are his measurements any more accurate than what ours are today? How can he or anyone else conduct tests accurately if the mics they are using don't really represent what we are actually hearing?

This is really beginning to become comical... our mics are not really showing us what we are really hearing, and our brains have faked us into believing bad sound is good simply because we smoothed out the response... although we really didn't smooth it out because what we are measuring with our bad microphones is not what we are actually hearing. What are people to do... there is not an easy answer for most people. Train your brain to like whatever you hear and don't do anything.

 

[AJ Soundfield]

If I could qualify for the paywall barrier I could share in the joke I guess?

Of the 26 mixing engineers (males and females 24-63 yrs old, average 10yrs experience), 7 preferred (lateral) absorption, 8 diffusion, 11 no treatment.
From the paper

Historical Context: The impetus for this research was provided by research conducted by F. E. Toole. In his recent book Sound Reproduction, he concludes that the listener can adapt to reflections in a room and can also clearly distinguish between acoustic comb filtering in the listening room (caused by differences of arrival between the direct and reflected sound) from the direct sound itself [1]. This is not the case for any comb filtering that is part of the direct signal, i.e., a strong reflection that was electronically combined with the direct signal during the recording process and is, therefore, part of the reproduced sound from the loudspeaker.

 

[AJ Soundfield]

This is really beginning to become comical... our mics are not really showing us what we are really hearing, and our brains have faked us into believing bad sound is good simply because we smoothed out the response

I'm also the same as 95% of the other folks in the world with a subwoofer in their system. The flat response that most of these DSP systems create does make you come away believing you have no bass... not sure how anyone can listen to it myself (but perhaps I've trained my brain to believe that's how it should be). When working with SVSound, we would have had every subwoofer we sold returned if we did not advise the customer to turn up their bass on the sub after their DSP system flattened out the response. I can't even begin to count the number of support calls I handled EVERY SINGLE DAY of people complaining they had no bass in their system.

Yeah that is pretty comical.
Btw, as I have explained innumerable times, the mic "hears" pretty good at 1m (>500hz), but by 3-4m in reflective domestic room, not so much like 2 ears/brain. Incredibly simple concept to understand.

So Toole was really keen on the Mirage M1 speakers, which are apparently no longer made... don't even think Mirage is manufacturing speakers now. What are his favorite music listening speakers now? I've seen his living room theater, but I'd be surprised if he is using that room for music with those speakers being wall mounted. I'd be very interested in comparing and measuring (with my mic that doesn't really show me what I'm hearing) some super neutral speakers that I have a anechoic measurement to start with, then compare those to my ML 15A's. That might give me an idea how far off I am. However, because I really like the sound of my ML's, my brain has trained me to like that sound, if the neutral speakers don't sound like this, I won't like them.

I'm also curious as to what mic he is using to record his measurements (he may mention this in his book, but I haven't see it yet). After all... apparently measuring with a mic does not really represent what we are actually hearing, so what mic does he use and how are his measurements any more accurate than what ours are today? How can he or anyone else conduct tests accurately if the mics they are using don't really represent what we are actually hearing?

Yes, M1s. The rest might require a deconvolution filter like what Mitch spoke of.

cheers

 

[Sonnie Parker]

Yeah that is pretty comical.
Btw, as I have explained innumerable times, the mic "hears" pretty good at 1m (>500hz), but by 3-4m in reflective domestic room, not so much like 2 ears/brain. Incredibly simple concept to understand.

What happens between 1m and 3m... what if we are 2.5m away? lol

So below 500Hz... mic does okay at 3m?

No way to measure accurately with a mic above 500Hz if 3m away.... we have to trust our ears?

 

[natty]

Of the 26 mixing engineers (males and females 24-63 yrs old, average 10yrs experience), 7 preferred (lateral) absorption, 8 diffusion, 11 no treatment.
From the paper

Thanks. Doesn’t seem like they replicated the study then (based on who the test subjects were since most of the studies that Toole mentions have both professionals and lay people) but now I’m even more intrigued about the results and details. What kind of speakers, treatments, content, etc.

But what you describe matches Toole’s conclusions. That anything from zero treatment to quality* treatment can work well and may be preferred to some people, when using “good“ (relatively flat, consistent off axis, etc) speakers.




* quality =

”5.6.1 A Message about Sound Absorption and Scattering
The dominant factor in the shape of the room curves, and therefore a factor in what we hear, is the off-axis sound, the first reflections paramount among them. This observation sends a clear message that in order to preserve whatever excellence exists in the loudspeakers, the room boundaries must not change the spectrum of the reflected sound. This means that absorbing and scattering/diffusing surfaces must have constant performance over the frequency range above the transition frequency around 300 Hz, below which the room resonances become the dominant factor.”

 

[natty]

What happens between 1m and 3m... what if we are 2.5m away? lol

So below 500Hz... mic does okay at 3m?

No way to measure accurately with a mic above 500Hz if 3m away.... we have to trust our ears?

lol! I hope that’s not true.

Dummy heads, multiple mic positions, good math all help capture the power sound curve that hits the ears, so while at the end of the day my ears have to be happy, good measurement can help me get into the ballpark pretty well. But it helps in different ways depending on frequency I guess?

Toole‘s conclusions are along the lines of trusting measurement to direct action under 500hz/transition frequency, actions like the number and placement and etc of subs, and of course some eq........

......and then above that point he is reluctant to use things like eq to solve physical problems assuming one has the golden apocryphal “good” speakers.

 

[Sonnie Parker]

Yeah... like maybe the JBL M2 speakers he helped design.