An Enticing Marketing Story, Theory Without Measurement?

[Guermantes]

I find that adaptation doesn't generally cause listener fatigue, but it can in some environments. For example, I find overly live listening rooms (or headphones!) fatiguing. And adapting to the loud background noise while talking to someone in a restaurant or party can become fatiguing. Of course, I wonder whether the fatigue is caused by the loud noise or the adaptation. I think both contribute.

I would say both but may depend on how easy the adaption is. As the process is one of extracting something meaningful (e.g. linguistic or musical) from the perceptual environment, then the harder it is to extract that meaning (i.e. the concentration required) is probably proportional to the fatigue induced.

I find it more fatiguing working with poor SNR recordings when attempting to extract intelligible speech from "noise" (this is a large part of my job).

I find it more fatiguing to sit down and follow a score while listening to an orchestral piece due to the concentration involved in processing parallel symbolic (reading notes on a page) and indexical (hearing musical pitches in a harmonic relationship) semiotic inputs.

I also find it more fatiguing to extract meaning from conversations in my second language than in my native tongue, especially if the speech is highly idiomatic and relies heavily on connotation. What is subconscious in the native language becomes a very conscious process in the second one.

That said, acoustic environments can convolve with sounds to enhance their meaning. Simple choral works like plainchant can take on a whole new resonance in reflective acoustics such as a cathedral. In this case, we are not trying to parse a complex harmonic structure or understand precisely what is being sung (unless we are fluent in Latin), instead we are hearing the convolution as part of an overall aesthetic (think of the famous example of Allegri's Miserere written for the Sistine Chapel). I suggest we actually are hearing the room as an integral part of the performance and this may not be fatiguing at all. Then, when someone gets up to give a speech or sermon in the same environment, we have to hear through the room to extract the linguistic information. Suddenly the acoustics may be hindering that process. Same room but different modes of listening.

 

[DDF]

I assume you were at the National Arts Center. If so, when in Ottawa I was a regular and my chosen seats were front mezzanine to avoid the seat-dip effect (which takes about 3 rows to develop) and sucks the mid bass level down. I designed the first "hifi" audio system for the hall (the original was truly PA only) and did a lot of measurements in it. It eventually got replaced as modern pop groups insisted on deafening sound levels. I remember measuring one of the roadie installations and discovered several loudspeaker drivers connected out of phase . They were glad, but . . .

Your neighbor was undoubtedly "fatigued" by excessive reverberation. It is an awful affliction to live with.

As for doing more science, it seems that reality has done much of it for us. If we are to reproduce what the artists created (circle of confusion), a useful benchmark is the acoustical performance of mixing and mastering rooms. They are not consistent, but typically are slightly deader than typically furnished homes. The 0.3-0.4s RT range is common, but a few are even lower - like the "non-room" philosophy. None that I know of are higher. It matters most in raw stereo, and there it is not so much the RT per se, but the early reflection levels. With multichannel or upmixed stereo the room matters less. Personally, I almost always use judicious upmixing for stereo sources - as a regular concert goer here in LA (Disney Hall) I appreciate envelopment. Front terrace seats, naturally

The NAC was last week, this week was Ottawa Symphony at Dominion Chalmers. Spoiled for choice DC's typically used for chamber but its the OS's new home: "louder than God" playing Mahler in such a small space! All seats are relatively close to the stage so the sound is never overwhelmed with reverberation, and is wonderful! At the NAC, I'm partial (maybe from familiarity) to Orchestra (J). Knowing skimming loss, I tried the mezzanine years ago but had a hard time adjusting to the (to me) heavy bottom end and reverberation.

Thanks for sharing your experiences at the NAC, that was funny. You might be interested to know that the NAC made major changes to the acoustics last year, and "first arrival" off the stage is louder and brighter through the addition of diffuse wood at stage sides and back. The orchestra sound is more immediate and "clean" for lack of a better word. The performers say they don't need to work so hard at projecting. I think you'd like the change.

The RT60 range you mention was also about ideal for intelligibility (I also used to design teleconference rooms), so that sounds about right. I'm helping my poor neighbour with some room treatment ideas, and maybe indirectly his marriage

 

[svart-hvitt]

SOME SERIOUS WORK ON LISTENING ROOMS: The anechoic virtual symphony orchestra and concert hall acoustics

Professor Tapio Lokki et al. make some interesting work in the article «Throw away that standard and listen: your two ears work better».

http://www.caa-aca.ca/conferences/isra2013/proceedings/Papers/P040.pdf

From the abstract:

«The acoustics of a concert hall is often described with the help of ISO3382-1:2009 parameters. However, several recent studies suggest that ISO3382-1 cannot explain the details of subjective perception nor preferences of the listeners. The current parameters, averaged over listener positions, and the use of only mid frequencies are definitely inadequate. We have recently measured over 15 concert halls around Europe with a novel measurement system -- the loudspeaker orches- tra. It consists of 34 calibrated loudspeakers on stage to simulate an orchestra, the most common sound source in concert halls. For objective analysis and spatial sound reproduction, we measure spatial impulse responses from each loudspeaker at accurately defined receiver positions. This enables accurate comparison of the properties of spatial impulse responses between halls. Fur- thermore, we have developed methods to convolve anechoic symphony orchestra recordings with the measured spatial impulse responses for multichannel loudspeaker listening. Both subjective comparison of halls and objective analysis with time-frequency and spatiotemporal properties of impulse responses have helped us to link architectural features and subjective perceptions. This presentation will explain these links and their relationship to properties of binaural hearing. Several examples are given to highlight the differences between vineyard and shoe-box type concert halls. In particular, the extreme importance of early reflections for engaging sound is explained».

The article draws upon for example this PhD thesis, «A Virtual Symphony Orchestra for Studies on Concert Hall Acoustics»:

http://lib.tkk.fi/Diss/2011/isbn9789526042916/isbn9789526042916.pdf

The idea seems to be that you’d need truly anechoic recordings of instruments to evaluate the room’s influence on measured and perceived sound in room. So they use 34 speakers to make a virtual orchestra. Fascinating, isn’t it?

Even if this material is concert hall related, I believe the logic can be applied to our understanding of music recording and playback in general as well.

A couple of quotes from Lokki et al.:

«The studies made in-situ while listening to a real orchestra have revealed that important perceptual features are loudness, reverberance, and intimacy. Laboratory studies have confirmed these discriminating factors, but the possibility for instant comparison inherent in the test conditions helped subjects to also hear differences in clarity, openness, spaciousness, and timbre».

«A few years ago, the ambivalent interpretation of interviews done in-situ in concert halls and the methodological problems of traditional listening tests in laboratories made us rethink the whole subjective evaluation process. We borrowed a sensory evaluation methodology from the food and wine industry».

«It is not a surprise that the best seat examples are from shoe-box halls. Those halls also have a flat floor on the audience area enabling nice enveloping reverberation. If the audience area is strongly inclined the seats behind block the enveloping reverberation. This is quite well seen in the directional analyses shown in Figure 4. Moreover, due to the lack of the side reflections the thick 30 ms curve is oval and the final energy (red curve) is not round. In Berlin Philharmonie there are quite late side reflections that might be perceived as echoes. The section plots show nicely the ceiling reflections typical to these geometries. In Cologne Philharmonie the far back wall of the hall reflects energy that reaches the listening position very late, making the perceived sound image even more monophonic.
The final examples are from two vineyard type halls, seats on the side or behind the orchestra, see Figure 5. Such seats have obvious balance problems between instrument sections and loss of high frequencies due to the directivity of instruments. The lack of early lateral reflection is clearly seen and the envelopment is also very weak. Moreover, even though the seats are quite close to the stage the overall sound power is weak. The section plots show strong and quite late ceiling reflections that might result, together with attenuated high frequencies, in a too distant sound image».

I made a previous remark that listening to recordings with room contribution in an anechoic chamber could make you believe that you were in that specific room where the recordings took place, inastead of sitting in a an open space type room. Some people questioned that idea. However, what Lokki et al. did is the same, but in opposite order: Anechoic recordings without room contribution to be played back in concert halls. Interesting isn’t it? Does the logic work only one way? Or is the logic of Lokki et al. flawed from the outset?

Some final thoughts that may be of relevance for this thread: Lokki et al. do NOT (of course!) make changes to the speaker response to optimize the sound in concert halls. On the other hand, they document that some seats are sort of nulled out, i.e. short on certain frequencies resulting in anemic sound. Is this suggesting that we cannot listen through the room? Some rooms are obviously better than others. Can this (old) insight be used to make compensations in the sound? Orchestras routinely adapt to the room, and play differently depending on the room acoustics. That’s why Lokki et al. made a virtual orchestra that does NOT adapt to the room. So if you can change the audio output of a real orchestra to (some extent) compensate for acoustics, why not use the same logic for speakers ? I guess that question remains unanswered in absence of research including observations (experiments, blind tests) on digital room correction.

 

[Dimifoot]

Simple choral works like plainchant can take on a whole new resonance in reflective acoustics such as a cathedral. In this case, we are not trying to parse a complex harmonic structure or understand precisely what is being sung (unless we are fluent in Latin), instead we are hearing the convolution as part of an overall aesthetic (think of the famous example of Allegri's Miserere written for the Sistine Chapel). I suggest we actually are hearing the room as an integral part of the performance and this may not be fatiguing at all.

This applies when attending the live event.
On the contrary, when we are reproducing this live event in our listening room we are hearing a sum of:
the room which is "an integral part of the performance"+our random room=???? unpredictable
And as @Floyd Toole mentions

It matters most in raw stereo, and there it is not so much the RT per se, but the early reflection levels. With multichannel or upmixed stereo the room matters less.

Shouldn't we, as consumers, have some clear guidelines from the experts, regarding room acoustics? RT60, early reflections, absorption, bass eq-treatments etc?

These of course will not be the same for a dedicated room vs a family living room, stereo vs multichannel...and the level of quality/fidelity of reproduction will be also vary as a result. The approach of "don't worry , we listen through rooms" leads to tiresome listening and dissatisfied listeners. And as a result , the snake -oil business flourishes.
I am not sure that "we listen through rooms" has been actually proven, since the study mentioned by Dr.Toole was done in rooms with similar RT's (0.3-0.4).
So the actual result should have been (in my opinion) "we listen through rooms when the RT's are between 0.3-0.4s"

On the other hand we have the study by S.Olive, back from 2009, where he compared 5 room (loudspeaker?) Eq +no Eq for listeners preference, and the result was in favour of the Eq for most of the cases (the ones that had resulting curve similar to the well known suggested tilted downwards). The study was done on (allow me) ancient software, and did not include the most advanced approaches, todays Trinnov and Dirac. (These two are the ones used today in Harman's Group processors).
Should this study be repeated now, 10 years later? Wouldn't we have a better view on Room (Loudspeaker) Eq, updated for 2019?

 

[Guermantes]

This applies when attending the live event.
On the contrary, when we are reproducing this live event in our listening room we are hearing a sum of:
the room which is "an integral part of the performance"+our random room=???? unpredictable

Shouldn't we, as consumers, have some clear guidelines from the experts, regarding room acoustics? RT60, early reflections, absorption, bass eq-treatments etc?

These of course will not be the same for a dedicated room vs a family living room, stereo vs multichannel...and the level of quality/fidelity of reproduction will be also vary as a result. The approach of "don't worry , we listen through rooms" leads to tiresome listening and dissatisfied listeners. And as a result , the snake -oil business flourishes.
I am not sure that "we listen through rooms" has been actually proven, since the study mentioned by Dr.Toole was done in rooms with similar RT's (0.3-0.4).
So the actual result should have been (in my opinion) "we listen through rooms when the RT's are between 0.3-0.4s"

On the other hand we have the study by S.Olive, back from 2009, where he compared 5 room (loudspeaker?) Eq +no Eq for listeners preference, and the result was in favour of the Eq for most of the cases (the ones that had resulting curve similar to the well known suggested tilted downwards). The study was done on (allow me) ancient software, and did not include the most advanced approaches, todays Trinnov and Dirac. (These two are the ones used today in Harman's Group processors).
Should this study be repeated now, 10 years later? Wouldn't we have a better view on Room (Loudspeaker) Eq, updated for 2019?

Oh, I agree wholeheartedly. But the "listen through the room" argument is fascinating as it leverages aspects of our bio-cultural evolution. My point is that acoustic spaces are heard differently depending on context.

In the case of focusing on the phantom image from speakers, it is a mode of listening where we do want to hear through the room. The design of studio control rooms and mastering suites are predicated on this desire. But, it appears that totally erasing the listening room acoustic is counterproductive as we haven't bio-culturally adapted to anechoic conditions. So we are seeking the Goldilocks solution: an immediate acoustic that doesn't impede our perception of the stereo image while still satisfying a psychological need for an auditory sense of space. In short: a room we can listen through.

I think, as you do, that the majority of living spaces are far from this ideal. I can see that room EQ offers an enticing promise: the ability to "correct" the room only when it matters most which is during music replay as we switch to that particular mode of listening. But it appears to me that EQ is mostly effective with room modes, not the parts of the spectrum where we are sensitive to early reflections.

I think a compromise approach is best. Treat early reflections with diffusion where possible and let EQ clean up lower frequency response problems that cannot be smoothed otherwise.

But all this discussion may be moot, as it appears consumer lifestyle products for music streaming and IoT applications are using DSP to adapt to acoustic spaces automatically anyway. My son's generation will probably consider room EQ as a standard part of music reproduction.

 

[svart-hvitt]

Oh, I agree wholeheartedly. But the "listen through the room" argument is fascinating as it leverages aspects of our bio-cultural evolution. My point is that acoustic spaces are heard differently depending on context.

In the case of focusing on the phantom image from speakers, it is a mode of listening where we do want to hear through the room. The design of studio control rooms and mastering suites are predicated on this desire. But, it appears that totally erasing the listening room acoustic is counterproductive as we haven't bio-culturally adapted to anechoic conditions. So we are seeking the Goldilocks solution: an immediate acoustic that doesn't impede our perception of the stereo image while still satisfying a psychological need for an auditory sense of space. In short: a room we can listen through.

I think, as you do, that the majority of living spaces are far from this ideal. I can see that room EQ offers an enticing promise: the ability to "correct" the room only when it matters most which is during music replay as we switch to that particular mode of listening. But it appears to me that EQ is mostly effective with room modes, not the parts of the spectrum where we are sensitive to early reflections.

I think a compromise approach is best. Treat early reflections with diffusion where possible and let EQ clean up lower frequency response problems that cannot be smoothed otherwise.

But all this discussion may be moot, as it appears consumer lifestyle products for music streaming and IoT applications are using DSP to adapt to acoustic spaces automatically anyway. My son's generation will probably consider room EQ as a standard part of music reproduction.

@Guermantes , you wrote: «Treat early reflections with diffusion where possible...»

In post #183 I referenced Lokki et al. who wrote the following (see my bolding of the text):

«So far we have measured 19 concert halls with our loudspeaker orchestra (...) These studies have already revealed a lot of novel information on the discriminating subjective attributes that can be used to describe aural differences between halls and how different people compare the halls. In addition, the results made us study detailed properties of reflections and we found that early reflections from diffusors might lower the sound quality and lateral reflections are preferred as they are perceived louder than median plane reflections due to the shape of the human head».

So I wondered if your recommendation was based on a reference?

Remember, Lokki et al. used anechoic recordings where the room is taken out of the equation. Normal recordings have lots of room information in them. Maybe it all shows that the perfect in-room sound is impossible when the amount of room information in recordings is unknown, and the playback room is an unknown as well. The sum of room information in the recording plus the added room information in the playback situation is arbitrary. Arbitrary factors make it impossible to attain a precise target. So maybe we should all relax a bit and just play record(ings) with our flat and smooth speakers?

 

[Floyd Toole]

This applies when attending the live event.
On the contrary, when we are reproducing this live event in our listening room we are hearing a sum of:
the room which is "an integral part of the performance"+our random room=???? unpredictable
And as @Floyd Toole mentions


Shouldn't we, as consumers, have some clear guidelines from the experts, regarding room acoustics? RT60, early reflections, absorption, bass eq-treatments etc?

These of course will not be the same for a dedicated room vs a family living room, stereo vs multichannel...and the level of quality/fidelity of reproduction will be also vary as a result. The approach of "don't worry , we listen through rooms" leads to tiresome listening and dissatisfied listeners. And as a result , the snake -oil business flourishes.
I am not sure that "we listen through rooms" has been actually proven, since the study mentioned by Dr.Toole was done in rooms with similar RT's (0.3-0.4).
So the actual result should have been (in my opinion) "we listen through rooms when the RT's are between 0.3-0.4s"

On the other hand we have the study by S.Olive, back from 2009, where he compared 5 room (loudspeaker?) Eq +no Eq for listeners preference, and the result was in favour of the Eq for most of the cases (the ones that had resulting curve similar to the well known suggested tilted downwards). The study was done on (allow me) ancient software, and did not include the most advanced approaches, todays Trinnov and Dirac. (These two are the ones used today in Harman's Group processors).
Should this study be repeated now, 10 years later? Wouldn't we have a better view on Room (Loudspeaker) Eq, updated for 2019?

One needs to be careful about assuming that newer is always better. Without naming names, some of the latest offerings do things that should not be done - indeed cannot be done - with EQ. It is a problem created when clever math is combined with powerful DSP without guidance from acoustical and psychoacoustical expertise. It is not uncommon to have to manually override automatic EQ solutions to restore the inherent qualities of excellent loudspeakers. If the loudspeakers are flawed, not all of the audible flaws can be seen in room curves and not all flaws can be corrected by EQ (e.g. frequency dependent DI, a common problem).

The one area where EQ is unquestionably needed is in the bass, below about 400-500 Hz - room modes and adjacent boundary effects. It is necessary to attenuate resonant peaks, avoiding filling narrow acoustical interference dips. With multiple subwoofers it is possible to attenuate room modes and for the EQ to benefit more than a single listener. It is not difficult, but not everybody does it. Other mistakes result from trying to "fix" non-minimum-phase ripples in steady-state room curves. EQ at mid and high frequencies should be broadband "tone control" kinds of spectral balance adjustments, but too many systems think they know better.

It is a work in progress, and not assisted by the ambitious claims of algorithms claiming to perfect the sound from any loudspeaker, in any location in any room. Two ears and a brain know the difference.

As for the 0.3-0.4s RT range. That includes the majority of domestic listening rooms, according to several studies. Below that rooms are uncomfortably "dead" even for conversation, and the cost of doing so prevents it from happening in most domestic spaces. Above that, rooms can be uncomfortably reflective/reverberant, but, as I said in an earlier post, some people seem to like a certain amount of it. In the intervening years more tests have been done in more rooms and it remains true that the highest rated loudspeakers are those radiating flat and smooth direct sound (audible in rooms of any RT), smooth off-axis performance (indicating an absence of resonances) and smooth DI confirming that reflected sounds exhibit timbral similarity to the direct sound (important in any room but most important in more reflective rooms). As loudspeakers get better, room problems are less obvious - but differences remain, which is why acoustical treatments exist.

EDIT: A final thought. For those with the technical skills, everything needed of in-situ EQ can be done with a low cost laptop measuring system and a DSP box capable of parametric filters adjustable in Q, amplitude and frequency. What is much more difficult, and possibly cannot be done, is the numerous anechoic or quasi anechoic measurements necessary to identify good loudspeakers.

Those looking for guidelines will find them in the 3rd edition of my book and the open access companion website: www.routledge.com/cw/toole

 

[svart-hvitt]

One needs to be careful about assuming that newer is always better. Without naming names, some of the latest offerings do things that should not be done - indeed cannot be done - with EQ. It is a problem created when clever math is combined with powerful DSP without guidance from acoustical and psychoacoustical expertise. It is not uncommon to have to manually override automatic EQ solutions to restore the inherent qualities of excellent loudspeakers. If the loudspeakers are flawed, not all of the audible flaws can be seen in room curves and not all flaws can be corrected by EQ (e.g. frequency dependent DI, a common problem).

The one area where EQ is unquestionably needed is in the bass, below about 400-500 Hz - room modes and adjacent boundary effects. It is necessary to attenuate resonant peaks, avoiding filling narrow acoustical interference dips. With multiple subwoofers it is possible to attenuate room modes and for the EQ to benefit more than a single listener. It is not difficult, but not everybody does it. Other mistakes result from trying to "fix" non-minimum-phase ripples in steady-state room curves. EQ at mid and high frequencies should be broadband "tone control" kinds of spectral balance adjustments, but too many systems think they know better.

It is a work in progress, and not assisted by the ambitious claims of algorithms claiming to perfect the sound from any loudspeaker, in any location in any room. Two ears and a brain know the difference.

As for the 0.3-0.4s RT range. That includes the majority of domestic listening rooms, according to several studies. Below that rooms are uncomfortably "dead" even for conversation, and the cost of doing so prevents it from happening in most domestic spaces. Above that, rooms can be uncomfortably reflective/reverberant, but, as I said in an earlier post, some people seem to like a certain amount of it. In the intervening years more tests have been done in more rooms and it remains true that the highest rated loudspeakers are those radiating flat and smooth direct sound (audible in rooms of any RT), smooth off-axis performance (indicating an absence of resonances) and smooth DI confirming that reflected sounds exhibit timbral similarity to the direct sound (important in any room but most important in more reflective rooms). As loudspeakers get better, room problems are less obvious - but differences remain, which is why acoustical treatments exist.

EDIT: A final thought. For those with the technical skills, everything needed of in-situ EQ can be done with a low cost laptop measuring system and a DSP box capable of parametric filters adjustable in Q, amplitude and frequency. What is much more difficult, and possibly cannot be done, is the numerous anechoic or quasi anechoic measurements necessary to identify good loudspeakers.

Those looking for guidelines will find them in the 3rd edition of my book and the open access companion website: www.routledge.com/cw/toole

@Floyd Toole , you wrote:

«As loudspeakers get better, room problems are less obvious - but differences remain, which is why acoustical treatments exist».

If you look at the measures taken and efforts put into concert halls, it’s obvious there are certain principles behind good and bad rooms, wouldn’t you agree? Regrettably, those principles aren’t fully understood, right? So listening through the room brings you just so far, but not all the way to the goal?

Regarding speakers that are adjusted to the room: An orchestra makes big adjustments to their sound output depending on the room they play in. What makes this adjustment right and speaker EQ wrong?

 

[March Audio]

@Floyd Toole , you wrote:

«As loudspeakers get better, room problems are less obvious - but differences remain, which is why acoustical treatments exist».

If you look at the measures taken and efforts put into concert halls, it’s obvious there are certain principles behind good and bad rooms, wouldn’t you agree? Regrettably, those principles aren’t fully understood, right? So listening through the room brings you just so far, but not all the way to the goal?

Regarding speakers that are adjusted to the room: An orchestra makes big adjustments to their sound output depending on the room they play in. What makes this adjustment right and speaker EQ wrong?

I'm not sure I see that a large concert hall acoustic bears much relation to what's going on in a small living room. Modes aren't an issue and RT is massive in comparison (for a start)

 

[svart-hvitt]

I'm not sure I see that a large concert hall acoustic bears much relation to what's going on in a small living room. Modes aren't an issue and RT is massive in comparison.

The physics are the same.

I hate it when physics become irrelevant in an area of much emotion.

So I turn the question: Do we need a physics theory of the concert hall and another physics theory of the listening room? Is a concert hall a listening room?

 

[andreasmaaan]

The physics are the same.

I hate it when physics become irrelevant in an area of much emotion.

So I turn the question: Do we need a physics theory of the concert hall and another physics theory of the listening room? Is a concert hall a listening room?

I'm not sure I understand where you're heading.

I think @March Audio is suggesting not that the physics of concert halls are fundamentally different, but rather that the psychoacoustics of listening in concert halls are different.

Which principles in your view aren't fully understood?

 

[Floyd Toole]

@Floyd Toole , you wrote:

«As loudspeakers get better, room problems are less obvious - but differences remain, which is why acoustical treatments exist».

If you look at the measures taken and efforts put into concert halls, it’s obvious there are certain principles behind good and bad rooms, wouldn’t you agree? Regrettably, those principles aren’t fully understood, right? So listening through the room brings you just so far, but not all the way to the goal?

Regarding speakers that are adjusted to the room: An orchestra makes big adjustments to their sound output depending on the room they play in. What makes this adjustment right and speaker EQ wrong?

OK, let's pause and take a deep breath.

Concert halls are fundamentally different from small listening rooms, and a 100 piece orchestra comprised of sources having very different directivities is fundamentally different from two (or preferably more) loudspeakers having well defined and unchanging directional properties. Concert halls need diffusion to deliver the precious orchestral sounds to all parts of a large audience. In a home we have an audience of one, or at most a few. Diffusion confuses the sound field, diluting the critical direct sounds that provide key information about direction and timbre. A concert hall is an inseparable part of the orchestral sound we hear - for reasons known only to independent minded architects and acousticians they are all different from one another. So debates about relative excellence are endless, and, yes, conductors do their best to accommodate to the differences. As I say in my book, this is sound production and each performance is what it is and may never be repeated again.

Sound re-production attempts to reconstruct the final performance in a mixing/mastering creative process. That mixing and mastering are not performed in uniform circumstances is the origin of the circle of confusion. Live symphonic concerts, which I attend about a dozen times a year, simply cannot be replicated in homes, and certainly not through two channels. No amount of "room EQ" will make it happen. However, for the vast majority of music created in studios we can get dangerously close, and when it works it is wonderful. There we are dealing with loudspeakers in rooms in all cases. Sadly, most music is stuck in a two-channel rut.

 

[svart-hvitt]

I'm not sure I understand where you're heading.

I think @March Audio is suggesting not that the physics of concert halls are fundamentally different, but rather that the psychoacoustics of listening in concert halls are different.

Which principles in your view aren't fully understood?

I was influenced by the Lokki et al. article that I referenced previously in post #183:

http://www.caa-aca.ca/conferences/isra2013/proceedings/Papers/P040.pdf

From the article:

«A few years ago, the ambivalent interpretation of interviews done in-situ in concert halls and the methodological problems of traditional listening tests in laboratories made us rethink the whole subjective evaluation process. We borrowed a sensory evaluation methodology from the food and wine industry».

I am at a loss if scientists in concert hall acoustics rely on another set of theory than scientists in listening room acoustics. If the two types of scientist have little overlap in methods and input - theory and measurement - it would be a sign of quasi science going on.

The concert hall acousticians now use loudspeakers to replicate an orchestra to understand concert hall acoustics. I think this research may ignite ideas to think through for those who are interested in «hifi science».

Wouldn’t you agree?

 

[svart-hvitt]

OK, let's pause and take a deep breath.

Concert halls are fundamentally different from small listening rooms, and a 100 piece orchestra comprised of sources having very different directivities is fundamentally different from two (or preferably more) loudspeakers having well defined and unchanging directional properties. Concert halls need diffusion to deliver the precious orchestral sounds to all parts of a large audience. In a home we have an audience of one, or at most a few. Diffusion confuses the sound field, diluting the critical direct sounds that provide key information about direction and timbre. A concert hall is an inseparable part of the orchestral sound we hear - for reasons known only to independent minded architects and acousticians they are all different from one another. So debates about relative excellence are endless, and, yes, conductors do their best to accommodate to the differences. As I say in my book, this is sound production and each performance is what it is and may never be repeated again.

Sound re-production attempts to reconstruct the final performance in a mixing/mastering creative process. That mixing and mastering are not performed in uniform circumstances is the origin of the circle of confusion. Live symphonic concerts, which I attend about a dozen times a year, simply cannot be replicated in homes, and certainly not through two channels. No amount of "room EQ" will make it happen. However, for the vast majority of music created in studios we can get dangerously close, and when it works it is wonderful. There we are dealing with loudspeakers in rooms in all cases. Sadly, most music is stuck in a two-channel rut.

@Floyd Toole , I see - needless to say - your point on differences between concert halls and listening rooms. But are there common grounds?

You wrote:

«Concert halls are fundamentally different from small listening rooms, and a 100 piece orchestra comprised of sources having very different directivities is fundamentally different from two (or preferably more) loudspeakers having well defined and unchanging directional properties».

So I wonder, did you read the Lokki et al. article (and others) I referenced in post # 183? They set up 34 loudspeakers in 19 concert halls to study room acoustics.

It seems like concert hall acousticians are not satisfied with the advice to just listen through the room, wouldn’t you agree? And I believe the loudspeakers’ role is to take the orchestra out of the equation in room acoustics research in order to control for room depending orchestra performance.

I like the research of Lokki et al. because it is systematic in its approach to control for factors that influence sound quality in rooms, even if these rooms are extremely big.

EDIT: Let me add that I dislike theories that work only small-size, not big-size.

 

[andreasmaaan]

I am at a loss if scientists in concert hall acoustics rely on another set of theory than scientists in listening room acoustics. If the two types of scientist have little overlap in methods and input - theory and measurement - it would be a sign of quasi science going on.

I don't quite understand this point. What is the set of theory that the former rely upon, what is the set of theory that the latter rely upon, and how do the two differ? (I read the link btw, just couldn't quite see exactly what you were referring to in it.)

The concert hall acousticians now use loudspeakers to replicate an orchestra to understand concert hall acoustics. I think this research may ignite ideas to think through for those who are interested in «hifi science».

Again I'm not 100% sure I understand what you're suggesting. Could you be more specific?

 

[svart-hvitt]

I don't quite understand this point. What is the set of theory that the former rely upon, what is the set of theory that the latter rely upon, and how do the two differ? (I read the link btw, just couldn't quite see exactly what you were referring to in it.)



Again I'm not 100% sure I understand what you're suggesting. Could you be more specific?

My point is: If research and science on listening rooms acoustics and sound quality in the usual hifi/pro speaker setting is not relevant for rooms acoustics and sound quality in concert halls, I question the value of one or both types of research.

I believe the same robust theories work in big and small settings. Lessons from mono are relevant for stereo are relevant for 5.1, for 7.2 for x.y.z setups etc. Theories for rooms measuring 1x1 are relevant for rooms measuring 10x10 and 100x100. I believe good science is scalable.

When I read concert hall acoustics, it seems like the researchers are obsessed with the room.

When I read @Floyd Toole - and more so @Cosmik - it seems like we will listen through the room so no need to worry too much about the room.

Conductors and performers in symphony orechestras are very room conscious too, and «EQ» their sound to compensate for the room. But speaker EQ is a sin according to some.

Call me confused...

In my area, which is not audio or sound related obviously, I am allergic to theories that work only in certain areas but not in others. To me it seems like the obsessions of audio and sound people change depending on which area they work in. Am I wrong to raise the question why this is so?

 

[Guermantes]

In a home we have an audience of one, or at most a few. Diffusion confuses the sound field, diluting the critical direct sounds that provide key information about direction and timbre.

Thank you for the input on this, Dr Toole. So would you favour treating specular reflections with absorption over diffusion? How about diffusion on the rear wall which is popular in RFZ designs?

There are certainly other perspectives on this:
http://www.myroom-acoustics.com/hosted/myroom-acoustics/MyRoom_Design-white_paper.pdf
Of course, this is probably not a room I would want to live in!

Sound re-production attempts to reconstruct the final performance in a mixing/mastering creative process. That mixing and mastering are not performed in uniform circumstances is the origin of the circle of confusion. Live symphonic concerts, which I attend about a dozen times a year, simply cannot be replicated in homes, and certainly not through two channels. No amount of "room EQ" will make it happen. However, for the vast majority of music created in studios we can get dangerously close, and when it works it is wonderful. There we are dealing with loudspeakers in rooms in all cases. Sadly, most music is stuck in a two-channel rut.

As there are competing philosophies on the design of mixing and mastering spaces, it makes the claims of room correction EQ even more enticing: one curve to rule them all! All of these complexities are swept under the carpet with the simple selection of a target curve and automated application of DSP

 

[Floyd Toole]

@Floyd Toole , I see - needless to say - your point on differences between concert halls and listening rooms. But are there common grounds?

You wrote:

«Concert halls are fundamentally different from small listening rooms, and a 100 piece orchestra comprised of sources having very different directivities is fundamentally different from two (or preferably more) loudspeakers having well defined and unchanging directional properties».

So I wonder, did you read the Lokki et al. article (and others) I referenced in post # 183? They set up 34 loudspeakers in 19 concert halls to study room acoustics.

It seems like concert hall acousticians are not satisfied with the advice to just listen through the room, wouldn’t you agree? And I believe the loudspeakers’ role is to take the orchestra out of the equation in room acoustics research in order to control for room depending orchestra performance.

I like the research of Lokki et al. because it is systematic in its approach to control for factors that influence sound quality in rooms, even if these rooms are extremely big.

EDIT: Let me add that I dislike theories that work only small-size, not big-size.

Lokki's work is fundamentally important in that it is the first attempt to systematically work through the variables. Loudspeakers are not the equivalent of musical instruments - and he doesn't claim they are - but they are a beginning. Moving his "orchestra" around to different venues is a smart idea. Historical evaluations of concert hall acoustics were highly subjective, biased by visual and other factors and so on, resulting in a very loose collection of guidelines. As I said, even those guidelines were thrown into confusion by the never ending panorama of designs. There has been agreement among at least some "experts" that shoebox halls are best. I have experienced the Vienna Musikverein and was underwhelmed. Sorry. Truth is that "perfection" has been a moving target, modified of course by the need to have audiences of profitable size.

From my perspective, making the subjective evaluations blind by using binaural technology is key. Historical subjective evaluations of halls have been done with all possible biases in place. Lokki can do better.

Surely the real answer is adjustable acoustics - physically and/or electronically manipulated. It is not a new idea. It would allow for more satisfactory concerts of different musical genres, including those employing loudspeakers. I would like to think that Lokki's research will provide this kind of guidance. Otherwise we will be stuck with building multi-million dollar edifices that may or may not please audiences - which are getting older; have you noticed? I see many attendees with hearing aids. Somehow I think that the musical ancients would be excited to be able to exercise some acoustical control over the performance venues. Time to move on?

It could be called "room manipulation"

 

[Guermantes]

I believe the same robust theories work in big and small settings. Lessons from mono are relevant for stereo are relevant for 5.1, for 7.2 for x.y.z setups etc. Theories for rooms measuring 1x1 are relevant for rooms measuring 10x10 and 100x100. I believe good science is scalable.
. . .
In my area, which is not audio or sound related obviously, I am allergic to theories that work only in certain areas but not in others. To me it seems like the obsessions of audio and sound people change depending on which area they work in. Am I wrong to raise the question why this is so?

It's not that the laws of acoustics are different for small rooms and large rooms, it's that different things become dominant factors in the respective spaces. For smaller rooms, room gain and early reflections close to the direct source (i.e. comb filtering) are dominant factors.

 

[svart-hvitt]

Lokki's work is fundamentally important in that it is the first attempt to systematically work through the variables. Loudspeakers are not the equivalent of musical instruments - and he doesn't claim they are - but they are a beginning. Moving his "orchestra" around to different venues is a smart idea. Historical evaluations of concert hall acoustics were highly subjective, biased by visual and other factors and so on, resulting in a very loose collection of guidelines. As I said, even those guidelines were thrown into confusion by the never ending panorama of designs. There has been agreement among at least some "experts" that shoebox halls are best. I have experienced the Vienna Musikverein and was underwhelmed. Sorry. Truth is that "perfection" has been a moving target, modified of course by the need to have audiences of profitable size.

From my perspective, making the subjective evaluations blind by using binaural technology is key. Historical subjective evaluations of halls have been done with all possible biases in place. Lokki can do better.

Surely the real answer is adjustable acoustics - physically and/or electronically manipulated. It is not a new idea. It would allow for more satisfactory concerts of different musical genres, including those employing loudspeakers. I would like to think that Lokki's research will provide this kind of guidance. Otherwise we will be stuck with building multi-million dollar edifices that may or may not please audiences - which are getting older; have you noticed? I see many attendees with hearing aids. Somehow I think that the musical ancients would be excited to be able to exercise some acoustical control over the performance venues. Time to move on?

It could be called "room manipulation"

Thanks for your answer @Floyd Toole !

I am just trying to attack sounds and reproduction from different angles and ask stupid questions in the process. My background is not in sound or audio, but I have some experience questioning models and data.

Your patience and answers are much appreciated