Audio Science: Does it explain everything about how something sounds?

[amirm]

What is a felt underlay, can you post a picture? Also, does this work as effectively if you put a thick rug or a Persian rug on top?

I have a thick wool carpet with some sort of rubber type underlay

Just to show an example of science being helpful there, listening tests show that floor reflections (as different from side reflections) have a negative effect on listener preference. Fortunately further analysis shows that the effect is the coloration caused in frequencies above 500 Hz. So this calls for absorption of floor reflections > 500 Hz. This is great news because the thickness of an absorber is dependent on frequency. For 500 Hz and above, a thick carpet with padding is enough to absorb those frequencies.

There are a lot of choices here between carpet type, thickness and that of the underlayment. Getting such data is very hard but there is a great AES paper with a ton of analysis. Here is a graph from it that I have saved up:

All the way on the right, we have the graph of the wool carpet with no padding. We see that it doesn't become effective until about 2 Khz. Sponge/rubber type padding only improves things a bit. But porous foam, felt and hair padding all substantially improve lower frequency absorption, giving us very good performance starting at 500 Hz which is what we desire.

Without this data, I would say just use the thickest carpet you can.

 

[Al M.]

I find it interesting that you think you can make reproduced timbres more real. If that is true, why stop at timbre? Why not really drive your point home by saying superior room treatments also make harmonics, tonality, warmth, and a small host of other coveted attributes sound more real?

Harmonics, tonality, warmth are of course part of timbre. Yet timbres can also be cold, when musicians perform in a cold acoustic. If that is recorded, the playback system should reproduce it as such. A system/room that always sounds warm, regardless of recording, exhibits coloration.

Nevertheless, whether these characteristics are already embeded in the recording (but remain inaudible at the speaker) or somehow not included in the recording at all, you are essentially saying that acoustic treatments act as a recovery mechanism to somehow restore some of the missing realism.

Exactly how do you suppose acoustic treatments recover lost music information that never made it to the speaker output in the first place? I’m sure Ethan Winer would love to know the answer to this question.

Nobody here, neither me nor anyone else, has claimed any such thing. Room treatments cannot somehow miraculously recover missing timbral information at the speaker level, they can only make sure that it is preserved as it reaches the ear of the listener. Room treatment does not make timbres more real if they are not emanating as realistic from the speakers, it preserves timbral realism that otherwise may be impaired by unfavorable speaker/room interactions. When I said "room treatment makes timbres more real" then this was sloppy wording but from the context it should have been clear what I meant.

But as a minimalist, you give me an empty room with reasonable dimensions, carpeting and pad, an absorptive material chair and 2 ottomans, and enough room to move the speakers into an optimal placement, and I’m good to go.

So you use these objects to dampen, diffuse and break reflections. Then why are you so opposed to room treatment devices that do the same? Your position makes no logical sense.

 

[jkeny]

I'm a novice in this area & I'm just wondering how these sort of adjustments to room reflections are measured - how is the direct sound teased out & measured separately from the reflections?
Does Griesinger's LOC measurement have relevance here (even though it might require some adjustments)?

There have been several requests that I put the matlab code for calculating the acoustic measure called LOC on this site. The measure is intended to predict the threshold for localizing speech in a diffuse reverberant field, based on the strength of the direct sound relative to the build-up of reflections in a 100ms window. The name for the measure is a default - If anyone can come up with a better one I would appreciate it. The measure in fact predicts whether or not there is sufficient direct sound to allow the ear/brain to perform the cocktail party effect, which is vital for all kinds of perceptions, including classroom acoustics and stage acoustics.

The formula seems to work surprisingly well for a variety of acoustic situations, both for large and small halls. But there are limitations that need discussion. First, the measure assumes that the speech (or musical notes) have sufficient space between them that reverberation from a previous syllable or note of similar pitch does not cover the onset of the new note. In practice this means the even if LOC is greater than about 2dB, a sound might not be localizable, or sound "Near" if the reverberation time is too long. A lengthy discussion with Eckhard Kahle made clear that the measure will also fail - in the opposite sense - if there is a specular reflection that is sufficiently stronger than the direct sound. This can happen if the direct sound is blocked, or absorbed by audience in front of a listener. In this case the brain will detect the reflection as the direct sound, and be able to perform the cocktail party effect - but will localize the sound source to the reflector. With eyes open a listener is unlikely to notice the image shift.

With these reservations, here is the matlab code. It accepts a windows .wav file, which should be a stereo file of a binaural impulse response with the source on the left side of the head. The LOC code analyzes only the left channel. There is a truncation algorithm that attempts to find the onset of the direct sound. This may fail - so users should check to be sure the answer makes sense. The code also plots an onset diagram, showing the strength of the direct sound and the build-up of reflections. If this looks odd, it probably is. In this version of the code the box plot has a Y axis that starts at zero. The final level of a held note (the total energy in the impulse response) is given the value of 20dB, and both the direct level and the build-up of reflections are scaled to fit this value. The relative rate of nerve firings for both components can then be read off the vertical scale. So - if the direct sound (blue line) is at 12, you know the eventual D/R is -8dB. Rename the .txt file to .m for running in Matlab. "Matlab code for calculating LOC"

 

[amirm]

I'm a novice in this area & I'm just wondering how these sort of adjustments to room reflections are measured - how is the direct sound teased out & measured separately from the reflections?

You mean in studies analyzing effects of reflections at different levels and distances relative to direct sound? If so, an anechoic chamber is used so there are no reflections. A loudspeaker generate the direct and only direct sound in this respect. Secondary loudspeakers are used and a delayed signal with varying levels is sent to it to represent reflection at that location.

Here again is a picture I have saved on this kind of system in an AES paper (by Dr. Olive):

Is this what you were asking?

 

[jkeny]

You mean in studies analyzing effects of reflections at different levels and distances relative to direct sound? If so, an anechoic chamber is used so there are no reflections. A loudspeaker generate the direct and only direct sound in this respect. Secondary loudspeakers are used and a delayed signal with varying levels is sent to it to represent reflection at that location.

Here again is a picture I have saved on this kind of system in an AES paper (by Dr. Olive):

Is this what you were asking?

No, I meant how can you judge the effectiveness of a particular treatment in a particular room regarding direct Vs reflected power.
It just seemed to me that Griesinger's work was directly addressing this direct to reflected energy ratio (modelled on psychoacoustic principles) & wondered how applicable it might be to consumer listening rooms?

 

[Atmasphere]

^^ you'd have to set up a mic or set of mics in the room and measure the effectiveness of a the treatment.

And BTW reflected energy is not a bad thing if done right- it can improve the soundstage by taking advantage of the ear/brain perceptual rules. When the ear hears a sound the brain makes a copy and compares it to other sounds heard just a little later. If it detects a match the ambient signature is used to help locate the source of the sound.

 

[jkeny]

^^ you'd have to set up a mic or set of mics in the room and measure the effectiveness of a the treatment.

And BTW reflected energy is not a bad thing if done right- it can improve the soundstage by taking advantage of the ear/brain perceptual rules. When the ear hears a sound the brain makes a copy and compares it to other sounds heard just a little later. If it detects a match the ambient signature is used to help locate the source of the sound.

Sure, I know a mic (or set of mics) would be needed but what is the measure of the effectiveness of the treatment? I know reflected energy is not a bad thing - I wasn't suggesting it was but just referencing Griesinger about the ratio of direct to reflected power being a measure of the intelligibility of the sound i.e direct power should always be higher than reflected.

To be a bit stupid about this, let me use an example which I know will be easy to answer but it illustrates what I mean - let's say there's a dip at a certain frequency & a flat FR is the goal - how does one know that the correction to this dip hasn't been achieved by creating many extra reflections which now bring this freq dip up to the level wanted but at the expense of muddiness?

 

[amirm]

No, I meant how can you judge the effectiveness of a particular treatment in a particular room regarding direct Vs reflected power.

That emulation aims to determine exactly that. There are also other schemes like an RPG Trifusor that had three sides and you could rotate them to see reflected versus absorption:

This is still a lot more limiting than the method I described.

It just seemed to me that Griesinger's work was directly addressing this direct to reflected energy ratio (modelled on psychoacoustic principles) & wondered how applicable it might be to consumer listening rooms?

Again, the setup I mentioned is precisely what is used to measure listener preference and thresholds for reflections of all sorts.

 

[amirm]

^^ you'd have to set up a mic or set of mics in the room and measure the effectiveness of a the treatment.

There is next to no method to interpret such measurements above transition frequencies of a few hundred hertz. Perceptual effects of reflections is very complex and no microphone or measurement system can remotely capture what we hear. Listening tests rules here. See my article on this: http://www.madronadigital.com/Library/RoomReflections.html

 

[Atmasphere]

Sure, I know a mic (or set of mics) would be needed but what is the measure of the effectiveness of the treatment? I know reflected energy is not a bad thing - I wasn't suggesting it was but just referencing Griesinger about the ratio of direct to reflected power being a measure of the intelligibility of the sound i.e direct power should always be higher than reflected.

To be a bit stupid about this, let me use an example which I know will be easy to answer but it illustrates what I mean - let's say there's a dip at a certain frequency & a flat FR is the goal - how does one know that the correction to this dip hasn't been achieved by creating many extra reflections which now bring this freq dip up to the level wanted but at the expense of muddiness?

Ha! I find this is a problem with many room correction schemes... Personally I am not aware of a measurement method that could actually sort this out. Amirm seems to be addressing listener preferences rather than the room itself.

Myself I don't go for flat frequency response. The main reason being- no matter what you do you are not going to get it (for example- good luck trying to fix a standing wave...). Its important to understand that the human ear/brain system has a means of knowing that you are in a room and automatically compensates for it. Try listening your phone with your other ear rather than the one you normally use and you will find it sounds different. Its not because your ear is different (unless it is of course); the brain does a lot to equalize in its quest for intelligibility. This might be part of why there are more nerves going to the ear rather than away from it but more research is needed.

 

[jkeny]

That emulation aims to determine exactly that. There are also other schemes like an RPG Trifusor that had three sides and you could rotate them to see reflected versus absorption:

This is still a lot more limiting than the method I described.


Again, the setup I mentioned is precisely what is used to measure listener preference and thresholds for reflections of all sorts.

I'm not sure we are understanding one another.
I'm asking, how, I judge my room has been improved by such treatments (apart from using my ears)?

 

[jkeny]

Ha! I find this is a problem with many room correction schemes... Personally I am not aware of a measurement method that could actually sort this out. Amirm seems to be addressing listener preferences rather than the room itself.

Myself I don't go for flat frequency response. The main reason being- no matter what you do you are not going to get it (for example- good luck trying to fix a standing wave...). Its important to understand that the human ear/brain system has a means of knowing that you are in a room and automatically compensates for it. Try listening your phone with your other ear rather than the one you normally use and you will find it sounds different. Its not because your ear is different (unless it is of course); the brain does a lot to equalize in its quest for intelligibility. This might be part of why there are more nerves going to the ear rather than away from it but more research is needed.

Right, this is what I was getting at - as a novice to this, there seems to be no measurement system that tells us we have improved the room acoustics.
Now, I'm all for sighted listening but I thought that this was frowned upon? I can see a much greater level of expectation bias being involved in this scenario where some thousands of currency has been spent on room treatment & to remove it would involve some thousands more.
What do those who generally try to enforce blind testing on others have to say about this?

Sure, I was just using flat freq response as a simple example target

 

[amirm]

I'm not sure we are understanding one another.
I'm asking, how, I judge my room has been improved by such treatments (apart from using my ears)?

Ah. That is ultimately the answer. The iterative approach however is quite expensive, time consuming and infinite in scope. For this reason, the right approach is:

1. Buy the right speakers so that their off-axis response is good (i.e. similar to direct)
2. Put enough absorption in the room as to make the room not too live, or too dead. For this, there is a measurement called RT60 although you can design the system in advance to hit this target.
3. Use acoustic research to determine what acoustic products you need to put where.
4. Optimize low frequencies with seating location, position and number of low frequency drivers, acoustic products, and electronic EQ.

Random approach to this topic is going to generate random results.

 

[PeterA]

There is next to no method to interpret such measurements above transition frequencies of a few hundred hertz. Perceptual effects of reflections is very complex and no microphone or measurement system can remotely capture what we hear. Listening tests rules here. See my article on this: http://www.madronadigital.com/Library/RoomReflections.html

Thank you Amir. This is the kind of response I was hoping to read when I started this thread. It sort of confirms what I always suspected, namely that the combination of both measurements and listening, and how to interpret the collected data, leads to a better understanding of audio systems and what we hear. This in turn can lead to better designs and get us closer to the music.

 

[Tony Lauck]

There is next to no method to interpret such measurements above transition frequencies of a few hundred hertz. Perceptual effects of reflections is very complex and no microphone or measurement system can remotely capture what we hear. Listening tests rules here. See my article on this: http://www.madronadigital.com/Library/RoomReflections.html

This is confirmed by my experience.

I used measurement to adjust speaker placement and cross-over adjustment plus digital parametric equalization to get the response flat below 200 Hz. Above that, I used room treatment to eliminate first reflections and adjusted the tweeter controls by ear to provide a good balance between dullness and harshness across a wide range of recordings. This seemed to work very well. Getting the bass response flat caused the walls behind the speakers (which are in a corner) to disappear. The only thing that I don't like about the bass response is that it was necessary to use polarity reversal and phase adjustment in the subwoofer and this impacts the time response slightly, making it difficult to detect absolute polarity.

Opening the blinds to see the wild animals cavort on my lawn (at times including turkeys, coyotes, fischers, deer, black bear and moose, plus the usual ground hogs, robins and squirrels) makes a noticeable deterioration of sound quality due to reflections. Even more so, if I forget to push the computer monitor way back from the plane of the speakers. None of these reflections are particularly hard to hear once one starts serious experimentation. When I got my Focal near field monitors I was hoping to be able to set them up entirely by ear, but this proved to be impossible in the case of my small room (formerly a bedroom).

 

[stehno]

Harmonics, tonality, warmth are of course part of timbre. Yet timbres can also be cold, when musicians perform in a cold acoustic. If that is recorded, the playback system should reproduce it as such. A system/room that always sounds warm, regardless of recording, exhibits coloration.

Nobody here, neither me nor anyone else, has claimed any such thing. Room treatments cannot somehow miraculously recover missing timbral information at the speaker level, they can only make sure that it is preserved as it reaches the ear of the listener. Room treatment does not make timbres more real if they are not emanating as realistic from the speakers, it preserves timbral realism that otherwise may be impaired by unfavorable speaker/room interactions. When I said "room treatment makes timbres more real" then this was sloppy wording but from the context it should have been clear what I meant.

Sorry, I took your original statement on its face and since it’s not that uncommon for others to make outlandish claims I wasn't all that surprised. Shoot, I was almost ready to reach out to Ethan Winer (anonymously of course) to share your news with him.

Your rewording is certainly more palatable.

So you use these objects to dampen, diffuse and break reflections. Then why are you so opposed to room treatment devices that do the same? Your position makes no logical sense.

It’s my turn to stand corrected. Hopefully, I never quite said zero acoustic treatments as I certainly pointed out more than once my need for carpet, pad, chair, and ottomans realizing what acoustically purposes they potentially serve.

Without carpet and pad, most likely my room would be too alive. Without the chair, I’d get tired standing and stooping down trying to be slightly more in line with the tweeters and without one of the ottomans, I’d have nowhere to rest my feet and the second ottoman is a leftover from when I had a sofa in the room instead of a chair. Using these objects do damp, diffuse, and break reflections but do I need them? Not really. Just a good quality carpet and pad. The chair and 2 ottomans will do some of that, but the benefits are pretty minimal in the grander scheme of things.

Since a room needs to be somewhat neutralized, long ago I've removed most every reflective furnishing out of the room. So my idea of acoustic treatments is add the carpet and pad and remove the negatives. Those are what I consider my minimum and my maximum requirements or acoustic treatments 101. That’s really it. Hence, my corrected statement is zero room acoustics above and beyond acoustic treatments 101 (for perhaps any reasonable room).

But I’m also pretty confident that with the levels of music info remaining audible at my speakers, the recordings are pretty much able to speak for themselves without further alteration or compensation. Might I extract a bit more musicality with some level of acoustic treatments beyond the 101 intro level? Probably but unsure.

But the real question is, in the grander scheme of things, with even SOTA-level playback system performance levels where they are today in contrast to the absolute sound, which categories deserve the most attention for greatest return on investment? I’d have to put acoustic treatments (above 101) right up against Meridian’s MQA high-rez format. Both of which are at the very bottom of my priority list. From a ROI perspective.

 

[jkeny]

Going back to the practical issues about audio science - a very good post from Jason Stoddard (he of Schitt fame) about measurements used, what they mean (or not), what the equipment costs, etc.

A nice bit at the end he calls "measuring the unexpected" about multi-tone tests "one of the measurements we do that is off the beaten path. This measurement appears to correlate at least loosely to subjective impressions, and it unearths some surprising problems in gear that otherwise measures very well."

I like his story about "The Perfect DAC"

The Perfect DAC was not one of ours. It was sent to us by a friend who wanted to get some measurements for it. This was a delta-sigma DAC, manufacturer and chipset redacted, with a very fancy power supply and all the buzzword-compliant stuff people like to hear about these days. We said, “Sure, why not.” And ran it through its paces.

And…in terms of standard measurements, this DAC blew everything we’ve ever measured away. I mean, vanishingly low noise floor, virtually undetectable power supply harmonics, insanely low THD, flat frequency response…

…until you looked at the IMD, which gave numbers a bit higher than you’d expect, given the THD results. And the numbers weren’t related to the 1K spike…they appeared down low, below 100Hz.

What? We ran through our multitone test (it’s easy to do digital multitones on a Stanford as well, not sure about other analyzers) and the low-frequency numbers went bonkers. As in, there was a broad range of non-harmonically related distortion components from 10-90 Hz, at a fairly high level (-50dB or so). -50dB is potentially audible. And it was up nearly 90dB from the baseline measurement.
...............
So…while we putter around confidently with all of the accepted measurements, maybe there are still realms out there where “here be there monsters.”

That’s why we still listen. And measure. And come up with new measurements. And listen again.

I know people will say that this DAC was broken but here's the dilemma - when is something broken? - when we hear a problem? - when a measurement reveals this? - what if the revealing measurement isn't performed, & we hear a problem, is it still broken ? - What if measurements show no problem, a dubious blind test shows no problem but sighted listening does? I know many that would argue that they have evidence on their side that there is no problem!! As my sig says "The greatest obstacle to discovery is not ignorance – it is the illusion of knowledge."
It's a bit of the one hand clapping in the forest sorta thing.

How many broken audio devices are in circulation simply because the measurement to reveal their brokenness hasn't been done?

This is the problem with measurements - where do you stop - there's thousands of possible measurements that can be done.

And, btw, multi-tone testing is not a new test - it has a long history in telephony, apparently but has been dropped from the stock suite of audio measurements for whatever reason. Some argue that IMD results can be extrapolated to multi-tone results - in a non-linear system, I would like to see them defend that. But the hint that something was wrong above did show up in the IMD test so that somewhat proves the efficacy of the stock testing? But what about the engineers who released this DAC - did they not notice the IMD hints given in their testing?

The problem, to my mind, with a lot of talk about audio science & measurements is that it is spoken of in absolute, infallible terms - not taking account of the real-world issues that surround any application of science

 

[amirm]

Going back to the practical issues about audio science - a very good post from Jason Stoddard (he of Schitt fame) about measurements used, what they mean (or not), what the equipment costs, etc.

A nice bit at the end he calls "measuring the unexpected" about multi-tone tests "one of the measurements we do that is off the beaten path. This measurement appears to correlate at least loosely to subjective impressions, and it unearths some surprising problems in gear that otherwise measures very well."

I like his story about "The Perfect DAC"


I know people will say that this DAC was broken but here's the dilemma - when is something declared broken? - when a measurement reveals this? - what if the measurement isn't performed, is it still broken? It's a bit of the one hand clapping in the forest sorta thing.

How many broken audio devices are in circulation simply because the measurement to reveal their brokenness hasn't been done?

This is the problem with measurements - where do you stop - there's thousands of possible measurements that can be done.

And, btw, multi-tone testing is not a new test - it has a long history in telephony, apparently but has been dropped from the stock suite of audio measurements for whatever reason. Some argue that IMD results can be extrapolated to multi-tone results - in a non-linear system, I would like to see them defend that. But the hint that something was wrong above did show up in the IMD test so that somewhat proves the efficacy of the stock testing? But what about the engineers who released this DAC - did they not notice the IMD hints given in their testing?

The problem, to my mind, with a lot of talk about audio science & measurements is that it is spoken of in absolute, infallible terms - not taking account of the real-world issues that surround any application of science

Before he gives any advice on measurements, he needs to clean up his house so that he doesn't ship $400 DACs that spit out this garbage:

And this type of system software player dependency:

And that is not all. He also shipped a pair of "premium" RCA cables that upon a second insertion/extraction has its center pin broken because it did not have sufficient strain relief. If we are going to quote people, let's quote people with much cleaner record than this please.

 

[jkeny]

Before he gives any advice on measurements, he needs to clean up his house so that he doesn't ship $400 DACs that spit out this garbage:

And this type of system software player dependency:

And that is not all. He also shipped a pair of "premium" RCA cables that upon a second insertion/extraction has its center pin broken because it did not have sufficient strain relief. If we are going to quote people, let's quote people with much cleaner record than this please.

Oh, so that was the manufacturer of that DAC you had measured these problems with - I didn't know that until now.

And yes, I have had some issues with one of their products apparently having DC on it's input but that all goes to show the fallibility of applied audio science & people. If he really does perform all the measurements that he specifies then these sorts of problems shouldn't get through or should they?

Irrespective of the above, his post is a good outline of stock suite of measurements that are generally used by audio manufacturers & his comments & observations are worthwhile. I'm not holding him up as a bastion of anything - just as a good summary of the stock application of audio science!

 

[Tony Lauck]

Before he gives any advice on measurements, he needs to clean up his house so that he doesn't ship $400 DACs that spit out this garbage:

Out of curiosity, what was the data path from the computer to the DAC, expecially digital interconnect type, protocol and clocking that produced those ugly plots?