Who do we trust?

[Floyd Toole]

Toole is basically correct but as far as I know he never investigates the small corrections (about +/-1 dB) that are required at certain frequencies in terms of what listeners prefer.

Sorry for the delay in responding. Listener preferences from the earliest double-blind evaluations favored loudspeakers without resonances and flattish axial frequency responses. The next research was to determine the audible thresholds for resonances so that design engineers could make intelligent compromises in lower cost products. Section 4.6 in the 3rd edition describes audible thresholds for spectral tilts and resonances of various Qs. We can detect low Q resonances with measured amplitude deviations of much less than 1 dB when listening to revealing program or pink noise. Toole, F. E. and Olive, S.E. (1988). “The modification of timbre by resonances: perception and measurement”, J. Audio Eng. Soc., 36, pp. 122-142.

So, if listeners prefer neutral - non-resonant - sound these are the ultimate criteria to be applied. Naturally, they can only be addressed using anechoic data.

 

[Krunok]

Sorry for the delay in responding. Listener preferences from the earliest double-blind evaluations favored loudspeakers without resonances and flattish axial frequency responses. The next research was to determine the audible thresholds for resonances so that design engineers could make intelligent compromises in lower cost products. Section 4.6 in the 3rd edition describes audible thresholds for spectral tilts and resonances of various Qs. We can detect low Q resonances with measured amplitude deviations of much less than 1 dB when listening to revealing program or pink noise. Toole, F. E. and Olive, S.E. (1988). “The modification of timbre by resonances: perception and measurement”, J. Audio Eng. Soc., 36, pp. 122-142.

So, if listeners prefer neutral - non-resonant - sound these are the ultimate criteria to be applied. Naturally, they can only be addressed using anechoic data.

I'm sorry but I don't understand this. Are you saying that we can detect non-linear amplitude response of "much less than 1dB" when it is caused by speakers resonance? Does that mean that we have a higher tolerance for amplitude non-linearity when the reason for non-linearity is not resonance?

 

[Floyd Toole]

I'm sorry but I don't understand this. Are you saying that we can detect non-linear amplitude response of "much less than 1dB" when it is caused by speakers resonance? Does that mean that we have a higher tolerance for amplitude non-linearity when the reason for non-linearity is not resonance?

Loudspeaker transducers are minimum-phase devices. This means that any deviation in amplitude response is associated with a time-domain behavior. So, if there is a peak, it is a resonance of some Q, and there will be some amount of ringing. Broadband spectral variations are low-Q and narrower band variations are higher Q resonances. It turns out that we humans pay little attention to ringing, so the evidence of importance is in the amplitude response. That is why smooth, flat direct sound response wins listening tests.

As I said, we are talking about the sound radiated from the speakers, not the combined direct and reflected sounds measured in rooms.

 

[Thomas_A]

Sorry for the delay in responding. Listener preferences from the earliest double-blind evaluations favored loudspeakers without resonances and flattish axial frequency responses. The next research was to determine the audible thresholds for resonances so that design engineers could make intelligent compromises in lower cost products. Section 4.6 in the 3rd edition describes audible thresholds for spectral tilts and resonances of various Qs. We can detect low Q resonances with measured amplitude deviations of much less than 1 dB when listening to revealing program or pink noise. Toole, F. E. and Olive, S.E. (1988). “The modification of timbre by resonances: perception and measurement”, J. Audio Eng. Soc., 36, pp. 122-142.

So, if listeners prefer neutral - non-resonant - sound these are the ultimate criteria to be applied. Naturally, they can only be addressed using anechoic data.

Thanks for the response and agree with all of that. The specific question I have is whether the sound a mono source (one speaker) is preferred (tonal balance) from a stereo setup with the same source playing. While the speakers may be perfectly linear in both cases you will measure and hear a timbral change between the two situations.

 

[Krunok]

Loudspeaker transducers are minimum-phase devices. This means that any deviation in amplitude response is associated with a time-domain behavior. So, if there is a peak, it is a resonance of some Q, and there will be some amount of ringing. Broadband spectral variations are low-Q and narrower band variations are higher Q resonances. It turns out that we humans pay little attention to ringing, so the evidence of importance is in the amplitude response. That is why smooth, flat direct sound response wins listening tests.

As I said, we are talking about the sound radiated from the speakers, not the combined direct and reflected sounds measured in rooms.

Thank you for the reply! I understand that transducers are minimum-phase devices but I thought they become less so once you install them in the loudspeaker box and much less so when that box is playing in the room. What would you say is a reasonably desired linearity when measured combined direct and reflected amplitude response in the room?

 

[Floyd Toole]

Thanks for the response and agree with all of that. The specific question I have is whether the sound a mono source (one speaker) is preferred (tonal balance) from a stereo setup with the same source playing. While the speakers may be perfectly linear in both cases you will measure and hear a timbral change between the two situations.

Ah, the dilemma of stereo. It is a fundamentally flawed delivery system and no amplitude panned phantom image created by sounds radiated by both loudspeakers can sound the same as, or as "accurate" as, sound radiated by a single loudspeaker. This is discussed at length in my book, starting with the obvious 2 kHz octave wide cancellation dip caused by acoustical crosstalk. When stereo is compared to mono listening - see section 7.4.2 in the 3rd edition - people were much more sensitive to loudspeaker faults in mono, but interestingly enough simple pop music was similarly revealing in that there were two "mono" sound sources - the hard panned left and right images. Classical music and pop with high production values generate enough spatial information to distract from many common small timbral faults. Spatial effects rank with timbral accuracy in our evaluations of sound quality. Obviously, one should start with neutral loudspeakers and let the chips fall where they may.

Even at my age, with accumulated hearing degradations, I can hear these effects and although I still derive pleasure from the music it grieves me that in 2019 we still are listening to the first improvement on mono - from the 50s! Will it never end?

 

[Floyd Toole]

Thank you for the reply! I understand that transducers are minimum-phase devices but I thought they become less so once you install them in the loudspeaker box and much less so when that box is playing in the room. What would you say is a reasonably desired linearity when measured combined direct and reflected amplitude response in the room?

You gotta read my book. Steady state room curves are not reliable indicators of sound quality. If not the book, at least have a look at:
Toole, F. E. (2015). “The Measurement and Calibration of Sound Reproducing Systems”, J. Audio Eng. Soc., vol. 63, pp.512-541. This is an open-access paper available to non-members at www.aes.org http://www.aes.org/e-lib/browse.cfm?elib=17839

It is an open access, no cost download, and addresses many of the issues. Humans have a remarkable ability to "listen through" rooms, to hear the essential qualities of sound sources, whether they are loudspeakers, voices or musical instruments.

 

[Krunok]

You gotta read my book. Steady state room curves are not reliable indicators of sound quality. If not the book, at least have a look at:
Toole, F. E. (2015). “The Measurement and Calibration of Sound Reproducing Systems”, J. Audio Eng. Soc., vol. 63, pp.512-541. This is an open-access paper available to non-members at www.aes.org http://www.aes.org/e-lib/browse.cfm?elib=17839

It is an open access, no cost download, and addresses many of the issues. Humans have a remarkable ability to "listen through" rooms, to hear the essential qualities of sound sources, whether they are loudspeakers, voices or musical instruments.

Well, I bought your book, hopefully summer vacations will provide enough free time for me to read it.

 

[Rja4000]

Well, I bought your book, hopefully summer vacations will provide enough free time for me to read it.

Same for me

 

[Rja4000]

Ah, the dilemma of stereo. It is a fundamentally flawed delivery system and no amplitude panned phantom image created by sounds radiated by both loudspeakers can sound the same as, or as "accurate" as, sound radiated by a single loudspeaker.

So, you think we should select loudspeakers by listening to only one of them, in mono ?

(I can hardly think how I could get an idea of the loudspeaker 'sound' by just looking at measurements. Unless there is an obvious flaw, of course.)

 

[Floyd Toole]

So, you think we should select loudspeakers by listening to only one of them, in mono ?

(I can hardly think how I could get an idea of the loudspeaker 'sound' by just looking at measurements. Unless there is an obvious flaw, of course.)

When you have read my book, your questions will have been answered - with proofs - along with a few others, I suspect. Enjoy!

Yes, we should evaluate loudspeakers in mono. After many decades of doing this, winners in mono tests always have prevailed in stereo tests, but we are fussier when listening in mono. Measurements can reveal flaws that humans can and cannot hear. The greatest problem with enjoying recorded music is that it is not at all standardized. There is a substantial "circle of confusion" between what is heard by the creators of recordings and by those who hope to enjoy them. The most "perfect" loudspeaker/room combination cannot always sound "good".

 

[Thomas_A]

Ah, the dilemma of stereo. It is a fundamentally flawed delivery system and no amplitude panned phantom image created by sounds radiated by both loudspeakers can sound the same as, or as "accurate" as, sound radiated by a single loudspeaker. This is discussed at length in my book, starting with the obvious 2 kHz octave wide cancellation dip caused by acoustical crosstalk. When stereo is compared to mono listening - see section 7.4.2 in the 3rd edition - people were much more sensitive to loudspeaker faults in mono, but interestingly enough simple pop music was similarly revealing in that there were two "mono" sound sources - the hard panned left and right images. Classical music and pop with high production values generate enough spatial information to distract from many common small timbral faults. Spatial effects rank with timbral accuracy in our evaluations of sound quality. Obviously, one should start with neutral loudspeakers and let the chips fall where they may.

Even at my age, with accumulated hearing degradations, I can hear these effects and although I still derive pleasure from the music it grieves me that in 2019 we still are listening to the first improvement on mono - from the 50s! Will it never end?

Thanks for this and this is the most likely reason I prefer somewhat altered frequency response for stereo speakers. One can argue that should be in the mix already, but I prefer a slight adjustment in the frequency curve to get a more natural phantom center. Others may disagree but the chips fall where they do.

 

[Rja4000]

When you have read my book, your questions will have been answered - with proofs - along with a few others, I suspect. Enjoy!
(...)
Measurements can reveal flaws that humans can and cannot hear. The greatest problem with enjoying recorded music is that it is not at all standardized. There is a substantial "circle of confusion" between what is heard by the creators of recordings and by those who hope to enjoy them. The most "perfect" loudspeaker/room combination cannot always sound "good".

Thanks for taking the time to answer me
I will read it ASAP.

I'm an engineer (and a new ASR reader), so, of course I'm convinced about the benefits of measurements. And psychoacoustics.
My point is that there is no 'perfect' speaker, that would match all requirements perfectly. (At least none that I could afford).
So they all 'sound' quite different today.
At the end of the day, I have to make sure the overall balance, dynamic,... match my 'tastes'.

I've spent quite some time selecting speakers in the past 40 years, and I know it's not easy, requires using multiple different recordings I'm used to, good conditions,... And even then, it's difficult.
Measurements, when/if available from serious source, so far, helped saving a lot of time pre-selecting only good loudspeakers.
But ultimately it's a matter of listening.
At least, that was my method so far for loudspeakers.
(For other components, like DAC, amplifier, source, that's different. Measurements are enough. Nowadays, 'listening' is just there to check nothing's badly broken in a setup.)

I own a few good (by my standards) speakers, and I like them all, for different purposes. They are all sounding somehow differently, and remain so even after Dirac eq.
When the overall balance is set the same (or similar enough), I guess the main differences are in some details in medium transcription.
And in dynamic behaviour.

 

[Krunok]

You gotta read my book. Steady state room curves are not reliable indicators of sound quality. If not the book, at least have a look at:
Toole, F. E. (2015). “The Measurement and Calibration of Sound Reproducing Systems”, J. Audio Eng. Soc., vol. 63, pp.512-541. This is an open-access paper available to non-members at www.aes.org http://www.aes.org/e-lib/browse.cfm?elib=17839

It is an open access, no cost download, and addresses many of the issues. Humans have a remarkable ability to "listen through" rooms, to hear the essential qualities of sound sources, whether they are loudspeakers, voices or musical instruments.

I took time today to read this paper and let me here quote a part of it as I have often encounter a question what is Dr. Tool's attitude toward room EQ (original wording, text emphasis by me):

"Below the transition/Schroeder frequency equalization
has a role to play.

In the upper-bass frequency range
adjacent-boundary interactions affect the sound energy radiated
into the room. Brought to the attention of the audio
community by Allison [15, 16] and discussed in [1] chapter
12, these fluctuations can be corrected for by equalization
using a spatially-averaged measurement to reveal the underlying
curve.

In the low-bass frequency range equalization
can be very useful as a means of attenuating prominent room
resonances at a single listening location. This is more successful
in rooms with significant low-frequency absorption.
Alternatively, with passive and active combinations of multiple
subwoofers standing waves can be manipulated, seat-to-
seat variations can be reduced, and global equalization
can then be beneficial to several listeners (see [2,17–20],
and [1] chapter 13).

In-room measurements are therefore
useful at low frequencies, differently in different sizes of
rooms, and bearing in mind that the target performance for
steady-state room curves may not be a horizontal straight
line.

In conclusion, there are reasons to exercise great caution
in the application of equalization based on conventional inroom
measurements. However, it is definitely advantageous
at lower frequencies, and later discussions will show that
equalization based on anechoic data is also useful in the
creation of superior loudspeakers."

 

[Blumlein 88]

Interesting. I had seen these here and there but it had never twigged that the intention was to mimic head shadowing. If you ever do your experiment, please share

@andreasmaaan

http://www.soundkeeperrecordings.com/format.htm

Barry Diament does all his recording with Jecklin disks now. You can download samples of his work above. Uses wideband Earthworks microphones. Minimalist production. Big Maggies for monitors in his studio. Maybe a little too much audiophile belief. Still Very nice recordings.

Thought you might enjoy some high quality examples of Jecklin disk recordings.

If you aren't familiar with him, he mastered some Led Zeppelin albums and Bob Marley. But does his own recordings to suit himself now.

 

[andreasmaaan]

@andreasmaaan

http://www.soundkeeperrecordings.com/format.htm

Barry Diament does all his recording with Jecklin disks now. You can download samples of his work above. Uses wideband Earthworks microphones. Minimalist production. Big Maggies for monitors in his studio. Maybe a little too much audiophile belief. Still Very nice recordings.

Thought you might enjoy some high quality examples of Jecklin disk recordings.

If you aren't familiar with him, he mastered some Led Zeppelin albums and Bob Marley. But does his own recordings to suit himself now.

This is very interesting - thx