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

An Enticing Marketing Story, Theory Without Measurement?

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
So it is not correct that the first sound defines the timbre.

So if you were to hear a familiar voice inside an anechoic chamber you wouldn't recognize it by that logic? The later reflections are related to the timbre of the room, not of the source.

No, but you don't decide timbre within a millisecond.

Any adjustment within the threshold of detectability will affect the timbre, whether you like that or not.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
I will repeat this once again for the sake of argument

The noteworthy studies that have been performed have concluded that independent of the room, loudspeakers with a neutral direct and indirect sound, are rated higher compared to non-neutral speakers. This doesn't mean that the room can't affect the score, but it won't affect the preference order.

Once you start altering the steady-state room curve of a neutral loudspeaker, you are altering the direct (anechoic) sound, which, in case of a neutral loudspeaker, is direct contradiction with the results of the studies.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
So if you were to hear a familiar voice inside an anechoic chamber you wouldn't recognize it by that logic? The later reflections are related to the timbre of the room, not of the source..

Depends what you mean with "defines the timbre". Direct sound is modified by the reflections within the precedence window and affects loudness, timbre and spaciousness.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
Depends what you mean with "defines the timbre". Direct sound is modified by the reflections within the precedence window and affects loudness, timbre and spaciousness.

Timbre means the spectral signature of distinct sounds, which are made of a fundamental and harmonic resonances. Timbre is not dependent on reflections, or the room. We will recognize a familiar voice out in the open, a bathroom or the living room. This is the living example of listening through rooms.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
Timbre means the spectral signature of distinct sounds. It is not dependent on reflections or the room. We will recognize a familiar voice out in the open, a bathroom or the living room. This is the living example of listening through rooms.

You will have a comb filtering effect that affects the timbre that is dependent on the reflections. You will hear that it sounds different in the bathroom vs. the open, but still be able to recognise the voice.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
But if a familiar voice sounds a certain way in the living room, you cannot EQ it to sound the same in the bathroom - but you'd try it on loudspeakers nonetheless, it cannot work.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,422
Likes
2,407
Location
Sweden
But if a familiar voice sounds a certain way in the living room, you cannot EQ it to sound the same in the bathroom - but you'd try it on loudspeakers nonetheless, it cannot work.

That is another topic, you should not alter a good speaker response with EQ other than perhaps the bass region < 200 Hz or so. And you should not alter room reflections other than when they become problematic, e.g. when the arrive too early or too high in amplitude. Also, as I mentioned elsewhere, there is no point in having reflections in the zero degree axis from the direct sound, aka from behind the speakers. These reflections only makes the sound worse.
 

Floyd Toole

Senior Member
Audio Luminary
Technical Expert
Industry Insider
Forum Donor
Joined
Mar 12, 2018
Messages
367
Likes
3,893
No, but you don't decide timbre within a millisecond.

Evidence from hundreds of double-blind listening tests over 50+ years indicate that the highest rated loudspeakers are those with the flattest, smoothest anechoic on-axis (direct sound) frequency response coupled with smooth and constant or gently changing off-axis response (the early reflections curve in the spinorama). All of this is in JAES papers and both of my books, especially the 3rd edition which offers numerous anechoic spinoramas and data on loudspeakers good and bad.

So, is timbre "decided" in a millisecond? No, but what happens in the direct sound preconditions the perceptual process in important ways - the precedence effect for localization is an important example, and also as a reference against which to compare later arriving sounds from different directions in normally reflective rooms which contribute to perceptions of both timbre and space. Spatial impressions are comparable with timbre in evaluations of sound quality.

As for the "decay across the frequencies", reverberation times in domestic rooms and home theaters are typically in the range 0.3-0.4 s - which cannot be considered classic reverberation as the sound field is not even approximately diffuse. By any room acoustical standard, these are "deadish" rooms. The reverberation we hear in recordings is natural reflections in the recording environment (which is 1.5 s and up for live music venues) or synthetic reverb added in the recording control room (which can be anything). What happens in the recording control room is the "original" performance for recorded music - regardless of musical genre.
 

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,721
Likes
1,934
Location
Oslo, Norway
I was looking for the latest study done on this, and finally found it: http://epubs.surrey.ac.uk/805557/9/2013 135th AES Convention Pike.pdf
Seems very methodologically sound to me.

From page 7:
The results of Olive et al., [2] have been confirmed: timbral differences between loudspeakers and between rooms are larger when comparisons are made using a direct comparison method (involving side-by-side stimulus presentations and short listening periods) than when using an indirect method (involving time gaps between listening, longer listening periods and less attention directed to timbral differences). This effect of comparison method indicates adaptation, a decrease in sensitivity, to timbre when listening is conducted over longer listening periods. The statistical reliability of adaptation effects were not reported by Olive et al. but here results have been shown to be statistically reliable. This work has also extended the reach of Olive et al’s results to speech sounds and to other loudspeakers and rooms.

Olive et al. found that no significant differences remained between loudspeakers and rooms when the indirect comparison method was used. This shows that that adaptation was complete in their study. Significant differences remained in this study.

What this means, in simpler terms, is that we adapt to and listen through rooms, but that this adaptation is not 100 percent. The room does matter for what we hear, even though the brain compensates for it. I also think it's been shown in quite a lot of studies that reflections affect the perceived timbre. Does it make sense, then, to apply (light) eq to compensate somewhat for the timbral effects of the room, if they are pronounced, even above Schroeder? I think it's a YMMV situation.
 
Last edited:

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
So, is timbre "decided" in a millisecond? No, but what happens in the direct sound preconditions the perceptual process in important ways

We did a test once with fellow musicians, playing back identical pitches played through a variety of instruments, but without the initial -start- of the note. Can you believe it was more difficult this way, to recognize the instruments, even for musicians? Just to say..
 

Floyd Toole

Senior Member
Audio Luminary
Technical Expert
Industry Insider
Forum Donor
Joined
Mar 12, 2018
Messages
367
Likes
3,893
We did a test once with fellow musicians, playing back identical pitches played through a variety of instruments, but without the initial -start- of the note. Can you believe it was more difficult this way, to recognize the instruments, even for musicians? Just to say..
I cannot - without a lot of effort - give the reference, but there was a serious study of instrument identification that involved deleting the onset of notes. In some real instruments this "onset" interval can be 100 ms or more, obviously as the characteristic resonances in the instrument built up energy. The result, as I recall, was that gross errors in instrument identification occurred when the onset transient was removed.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
I believe it is possibly the same test that inspired us to repeat the process in the first place - it's been some years and I can't seem to find any references. It was a humbling experience in any case.
 

Absolute

Major Contributor
Forum Donor
Joined
Feb 5, 2017
Messages
1,084
Likes
2,125
Evidence from hundreds of double-blind listening tests over 50+ years indicate that the highest rated loudspeakers are those with the flattest, smoothest anechoic on-axis (direct sound) frequency response coupled with smooth and constant or gently changing off-axis response (the early reflections curve in the spinorama). All of this is in JAES papers and both of my books, especially the 3rd edition which offers numerous anechoic spinoramas and data on loudspeakers good and bad.
Hopefully all of the manufacturers are aiming for such speakers, albeit with different levels of other priorities.
Which brings me to some questions I've always pondered regarding such tests;

- Have any tests been done where the highest rated speakers were tested against itself only room-compensated?

- Do you know if there's been noted differences in relative placements of highest rated speakers depending on where it was tested?
For example if nr 2 in one test became number 1 in a smaller/bigger/different environment, perhaps because of wider/narrower directivity.

- Would it matter how far away the listening positions were? Far off you have a significant drop in higher frequencies that you won't have up close.
Or are most things above 10k rather meaningless in the big picture?

So, is timbre "decided" in a millisecond? No, but what happens in the direct sound preconditions the perceptual process in important ways - the precedence effect for localization is an important example, and also as a reference against which to compare later arriving sounds from different directions in normally reflective rooms which contribute to perceptions of both timbre and space. Spatial impressions are comparable with timbre in evaluations of sound quality.
Alot of people dealing in acoustics are talking about the first 5-7 ms as critical because that's perceived as direct sound. When you tested for preference for lateral reflections, did you find a particular point in time where the preference factor changed significantly?
 

JoachimStrobel

Addicted to Fun and Learning
Forum Donor
Joined
Jul 27, 2019
Messages
518
Likes
303
Location
Germany
Excellent post. I also play the piano, and have tried the piano in different rooms, and in different positions in the living room. Same experience here: It sounds really different! I actually try to do some "room correction" with the piano though. At the current position, it was somewhat too bright, because it is standing in a corner with reflective walls which probably reflect more of the higher than the lower frequencies. So the first thing I did was to experiment with adding more carpeting etc around it. That helped quite a lot. But I also asked my piano tech to voice it down - to make it sound a bit darker. That also helped.

When it comes to hifi and speakers, the same thing applies, I think. Light broadband eq, based on the response in the room, is never a bad thing. As is taking down excessive energy in the bass. Beyond that... hm, I'm somewhat skeptical.

And the surprising thing is, that these different sounds of a piano are not taken as a real problem, asides possibly voicing it a bit differently, but that is just a nuance.
But if a loudspeaker would sound that differently in different rooms we make a fuzz about it.
This is very similar to color temperature. We have no problem with a face in 8000 or 2000K light and enjoy the view. But a picture being displayed with the same yellow that the 2000k view had, would not be accepted as nice.
As somebody else posted, we hear through the room, preferentially with real instruments, like we see through a color cast if it is fully immersive. That is the power of our brain ant 2 eyes/ears if we sense enough cues from the “real” instrument.
I was always amused how little real musicians cared about their sound reproduction system ( that is at least my impression), and it is easy to see why: They can read the cues from the instruments and reconstruct the real thing even from a 50$ cassette recorder.
 

UliBru

Active Member
Technical Expert
Joined
Jul 10, 2019
Messages
123
Likes
337
Once you start altering the steady-state room curve of a neutral loudspeaker, you are altering the direct (anechoic) sound, which, in case of a neutral loudspeaker, is direct contradiction with the results of the studies.
Neutrality has a clear mathematical definition. A neutral speaker thus would transfer the input signal 1:1.
An according pulse response would be a Dirac pulse. Such a speaker would behave correctly under steady-state aspects of course.
As music is not steady-state such a speaker also would behave correctly with transient-state aspects. But not only the frequency response would be perfectly flat, also the time behaviour would be perfect.
As the air pressure does not follow precisely the speaker driver movement at least for lowest frequencies down to the DC range I would even concede that a speaker described by a minimumphase highpass (corner frequency e.g. around 20 Hz) but otherwise flat frequency response could be described as neutral.
But I have never ever come across a speaker with such a neutrality. All measurements I have seen up to now show deviations from neutrality.
So how much deviation is allowed? When is a speaker neutral, when is it no longer neutral?
Is it neutral with a perfect frequency response but bad time behaviour? Can a speaker sound good when the frequency response is not flat?
Is there a common sense about a certain limit a speaker has to pass to call it neutral? When "neutral" speakers sound different which one is more neutral from a scientific point of view?

Is the resulting sound correct when the so-called neutral speaker is driven by an amplifier, preamplifier, DAC (which type? which brickwall filter?) or other sound source which is deviating from neutrality by themselves?

So what's wrong with trying to get the resulting soundwave closer to the original sound signal by correction? Why do people always stare at frequency responses and thus at the steady-state behaviour when clearly music is not steady-state? What's wrong with excessphase correction to improve a speaker under time-aspects?

I truly wonder how people talk about neutral speakers.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
You are looking at this way too complex. In any case the original signal is electric in nature, and is being converted into 3-dimensional acoustic energy - it can never be 1:1. Years of testing have determined that loudspeakers that feature smooth, flat direct sound fields, and smooth dispersion characteristics are always preferred. There have also been studies about threshold detection of anomalies such as resonances. Floyd Toole wrote a great book about it - and I've no interest in quoting entire chapters here.

In any case, if you take a loudspeaker that let's say, an anechoic listening window response between 100Hz to 10kHz of +/1dB from the reference axis, and when placed in a room, continue to adjust the response above the transition frequency - you will now have adjusted the anechoic response, making the loudspeaker less neutral - which we've learned is always preferred and simultaneously technically the most accurate.

Two "neutral", but different, loudspeakers will still sound different in an A/B comparison. They will however, sound similar - not identical. Which is right for you? Look at the measurements, select two or three models, and try and organize a little blind test for yourself.
 

UliBru

Active Member
Technical Expert
Joined
Jul 10, 2019
Messages
123
Likes
337
You are looking at this way too complex.
No, I'm just facing the difference between theory and practice. This allows me to keep away from steady-state postulations.

...and when placed in a room, continue to adjust the response above the transition frequency - you will now have adjusted the anechoic response, making the loudspeaker less neutral - which we've learned is always preferred and simultaneously technically the most accurate.
Yes, a change of the frequency by a filter will of course change the anechoic "neutrality" of the speaker. But what if the room "neutrality" of the speaker is different than the anechoic "neutrality"? Is it wrong to use a filter then?

Two "neutral", but different, loudspeakers will still sound different in an A/B comparison. They will however, sound similar - not identical. Which is right for you? Look at the measurements, select two or three models, and try and organize a little blind test for yourself.
If I'm allowed to select between so-called "neutral" speakers to meet my personal taste it should also be allowed to select another "less neutral" speaker to meet my taste. And it should also be allowed to use a filter to change any speaker to my taste.
As there is no absolute criteria (too complex ;), 3-dimensional acoustic energy can never be 1:1 the original signal) we can also follow subjective criterias which is our taste. So each way to improve the result is legal. Live and let live.
 

TimVG

Major Contributor
Joined
Sep 16, 2019
Messages
1,181
Likes
2,573
Yes, a change of the frequency by a filter will of course change the anechoic "neutrality" of the speaker. But what if the room "neutrality" of the speaker is different than the anechoic "neutrality"? Is it wrong to use a filter then?

Changing the speaker to suit the room is a fool's errand in my opinion - independent of the room, the neutral loudspeakers score higher than their non-neutral counterparts, this has been proven.


If I'm allowed to select between so-called "neutral" speakers to meet my personal taste it should also be allowed to select another "less neutral" speaker to meet my taste. And it should also be allowed to use a filter to change any speaker to my taste.

Sure, it's not about being allowed or not being allowed. It's about facts and opinions - make your own choice, but do it blind! Remove all bias, and let your ears be the judge. If you are like most, you will opt for neutrality.
 

UliBru

Active Member
Technical Expert
Joined
Jul 10, 2019
Messages
123
Likes
337
Changing the speaker to suit the room is a fool's errand in my opinion - independent of the room, the neutral loudspeakers score higher than their non-neutral counterparts, this has been proven.
Simple example:
if a room boosts a frequency by a room mode you can change the room. To change the living room to an anechoic chamber will certainly help.
But what if the spouse does not like to live in an anechoic chamber?
Is it better to leave the "neutral" speaker untouched or is it better to reduce the critical frequency by a filter? Of course the latter solution must be a fool's errand as it changes the neutrality of the speaker. I guess that the fool will not take care about this but enjoy the better sound.
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Is there any accepted, reproducible demonstration of just what on Earth "timbre" is, in any sort of technical definition that can be applied across contexts?

This word seems to have a definition ascribed upon it variously and as numerous as the people that employ it's use. Or is it just a description of "what a sound should sound like" ..as if all of audio reproduction isn't chasing this - generally speaking.
 
Top Bottom