Before we fall into a bottomless pit of self-pity over the relativism of the senses and start sounding like engineers puzzling over human testing they aren't trained to perform....
We can distinguish two realm of understanding that can help keep our minds clear.
1. Good methods of testing with people. Everybody (except Golden Ears) knows the importance of blind or double-blind test*. Also there are strategies such as preferring comparisons to absolute judgments and lots of other testing smarts out there.
2. Important to be clear about the nature of various perturbations of hearing. Some are momentary influences of context, some are cognitive (where you improve over time in decoding a room's sound), some are learning where you learn what to listen for, some are long-term losses, and so on. Each perturbation gets addressed in different ways. For example, in certain extreme cases you can do what we call a "World's Fair" design where each person being tested gets just a single exposure to whatever-the-issue-is because their judgment is "spoiled" afterward (at Bell Labs we used to invent World's Fair tests for fun during our coffee-klatches).
*on the weekend, I was part of a police thief identification line-up (nope, not at that side of it) which was more than triple-blind, bless 'em. Don't ask.
We can distinguish two realm of understanding that can help keep our minds clear.
1. Good methods of testing with people. Everybody (except Golden Ears) knows the importance of blind or double-blind test*. Also there are strategies such as preferring comparisons to absolute judgments and lots of other testing smarts out there.
2. Important to be clear about the nature of various perturbations of hearing. Some are momentary influences of context, some are cognitive (where you improve over time in decoding a room's sound), some are learning where you learn what to listen for, some are long-term losses, and so on. Each perturbation gets addressed in different ways. For example, in certain extreme cases you can do what we call a "World's Fair" design where each person being tested gets just a single exposure to whatever-the-issue-is because their judgment is "spoiled" afterward (at Bell Labs we used to invent World's Fair tests for fun during our coffee-klatches).
*on the weekend, I was part of a police thief identification line-up (nope, not at that side of it) which was more than triple-blind, bless 'em. Don't ask.
Which is why you should have a well known reference system at hand. It will at least help you through those moments when a head cold of hay-fever among other things temporarily dulls your senses.
Rob🙂
Toole: "...on the topic of the role of reflections in the corruption or enhancement of timbre - sound quality - it is now evident that in normal listening rooms there is little risk of corruption (by comb filtering)..."
Thank goodness.
Of all the things that Floyd says, this is the most controversial. I tend to agree with Floyd in most everything he says, but on this point I do not agree. Very early reflections have an effect on sound quality. Maybe its not timbre and maybe some people like it, but there is an effect.
very early reflections are linear distortion from a listener's pov, aren't they?
so their subjective effect would be level dependent?
I, also disagree. I think Floyd was talking in very general terms (that many reflections exist and most are innocuous). Certainly the Bech studies disagree. Early reflections of sufficient magnitude will modify timbre, especially if received from the front rather than the sides.
Play pink noise through a speaker and walk away from it. You will always hear a comb filter pitch that rises in frequency as you recede.
David S
I think we've all at times convinced ourselves that we knew what we were talking about, formed an opinion, and later realized that there were more variables than we had realized. Floyd is human too.
There was a paper written pre 1983, called Backward Inhibition, and I thought it was written by Lipshitz or Toole, but I apparently lost my copy and can't remember for sure. It was a great paper (possibly from an AES journal) that described how the ear-brain mechanism reacts to different amounts of a signal being delayed, and then added back into itself. There were 4 or 5 different delay amounts that had distinctly different psycho-acoustic effects, completely independent of the comb filtering aspect. I would sure love to get my hands on that paper again. I vaguely remember the author, or one of the authors, being from Canada (?) Can anyone help me find this paper? Every audio engineer should have a copy. One thing it said was that an echo that is more than 40mS delayed from the original, will effectively replace, in the minds perception, the original. It was a very fascinating paper. I must have loaned it out and never gotten it back.
There was a paper written pre 1983, called Backward Inhibition, and I thought it was written by Lipshitz or Toole, but I apparently lost my copy and can't remember for sure. It was a great paper (possibly from an AES journal) that described how the ear-brain mechanism reacts to different amounts of a signal being delayed, and then added back into itself. There were 4 or 5 different delay amounts that had distinctly different psycho-acoustic effects, completely independent of the comb filtering aspect. I would sure love to get my hands on that paper again. I vaguely remember the author, or one of the authors, being from Canada (?) Can anyone help me find this paper? Every audio engineer should have a copy. One thing it said was that an echo that is more than 40mS delayed from the original, will effectively replace, in the minds perception, the original. It was a very fascinating paper. I must have loaned it out and never gotten it back.
Towards the end of Sound Reproduction Dr. Toole does talk about absorbing early reflections altering "spectral Balance"--pg 417. To me this would also suggest that wide or narrow, controlled or uncontrolled directivity speakers will will also effect spectral balance. Dr. Toole agrees on pg 323. That would then extend into how well controlled the directivity is and how low that control goes to get you spectral balance. IOW, off axis performance modifies the timbre/spectral balance d/t reflections of the on axis performance. Most professional monitor speakers are flat/smooth on axis, start wide, rather omni in the bass and get progressively narrower toward the treble with a fairly smooth off axis declination. It would seem a reasonable starting point for the home enthusiast.
There are other parts in the book where Dr. Toole discusses "timbre" effects of off axis behavior and well as perceived detail and giving the ear a "second look" with reflections. I'm too lazy to find them now.
Dan
There are other parts in the book where Dr. Toole discusses "timbre" effects of off axis behavior and well as perceived detail and giving the ear a "second look" with reflections. I'm too lazy to find them now.
Dan
When read carefully you find that Floyd was talking about "preference" and "spatiousness", mostly in the context of large venue recordings. "Preference" is a tricky thing to measure when there is no comparitive basis. In private discussion with Jens Blauert, he commented that he believed that very early reflections would add to spatiousness and ASW, but by deffinition this would reduce "imaging" in the sense of being able to localize an individual source. Hence, in terms of "preference" it would depend almost entirely on what one was looking for and listening to. Consistant with previous comments, venues with high "spatiousness", like orchestras in large auditoriums, have almost no precise imaging without the visual sense to locate them and substantial "spatiousness". On quite the contrary a small group of players, say Eric Clapton in "Unplugged", have very precise images and source locations.
This is clearly one area where there may never be agreement and certainly not without the use of a wide selection of source material and comparative assesments of two different setups. This, of course, is difficult to impossible to do. So we argue here ad-infinitum with no conclusions - there aren't any.
A lot of loose talk here. We have to be careful not to confuse things when words or names overlap.
First, there are the acoustic properties of rooms - hall-sized and small-sized - that influence how things sound in them.
Second, there are the acoustic properties of speakers, rooms, the placement of speakers, the location of drivers, and all their interaction (not to mention how the source was created), which influences how speakers work in the room. In turn, some aspects of quality relate to a kind of general enjoyment (and in turn, that is sometimes related to concert hall ambience) and some relate to specific close reproduction of you-are-there instruments playing.
Nothing gets enmeshed in this confusion more than "comb fitltering".
First, there are the acoustic properties of rooms - hall-sized and small-sized - that influence how things sound in them.
Second, there are the acoustic properties of speakers, rooms, the placement of speakers, the location of drivers, and all their interaction (not to mention how the source was created), which influences how speakers work in the room. In turn, some aspects of quality relate to a kind of general enjoyment (and in turn, that is sometimes related to concert hall ambience) and some relate to specific close reproduction of you-are-there instruments playing.
Nothing gets enmeshed in this confusion more than "comb fitltering".
Dr. Toole discusses image problems created by front wall (the wall behind the speakers) reflections in ch 8. With listeners reported front wall absorption "improved localization and reduced coloration".
Dan
Dan
Last edited:
Not familiar with it but a search showed 2 AES papers, one by JR Ashley and one by Jens Blauert, using that term in the title.
I've thought these paper might be useful to read, but never have.
AES E-Library: A New, Psychoacoustically More Correct Way of Measuring Loudspeaker Frequency Responses
AES E-Library: The Influence of the Room and of Loudspeaker Position on the Timbre of Reproduced Sound in Domestic Rooms
AES E-Library: Perception of Reproduced Sound: Audibility of Individual Reflections in a Complete Sound Field, II
Anyone else? They also seem to contradict the 'reflections don't matter' type of thought.
Dan
AES E-Library: A New, Psychoacoustically More Correct Way of Measuring Loudspeaker Frequency Responses
AES E-Library: The Influence of the Room and of Loudspeaker Position on the Timbre of Reproduced Sound in Domestic Rooms
AES E-Library: Perception of Reproduced Sound: Audibility of Individual Reflections in a Complete Sound Field, II
Anyone else? They also seem to contradict the 'reflections don't matter' type of thought.
Dan
Floyd told me that his opinion in this regard changed over the years. He used to believe that early reflections mattered, but then came to believe that they "added spaciousness" to the sound. Its clearly an area where he is not as firm as other areas.
Hello,
Interesting is also the usage of the term "time dependent frequency response".
- Elias
Interesting is also the usage of the term "time dependent frequency response".
- Elias
Looking at the number of microphones in sight in the promotional materials for "Unplugged" I have to conclude that any "precise images" or localization is entirley synthetic . . . and I have a number of such recordings where the easy localization persists with dipoles, despite the substantial front wall reflections. "Spacious" and "localization" are not opposites . . . both can exist simultaneously, depending on the recording and the (properly set up) loudspeakers, of course.
Read the description.
And read what I wrote again. You have a lot of trouble with that.Dan
I tend to find side walls being too near being much more of a problem with affecting localization of images than the wall behind the speaker - although that depends on the baffle size, with a narrow baffle design more influenced by the wall behind it than a wider baffle.
I wonder how much of the change in "localization" reported could be attributed simply to the change (increase) in frequency response below the baffle step frequency that having the speaker too close to the wall behind it has.
Particularly on narrower speakers this will boost the lower midrange by a couple of dB or more which is more than enough to throw out the delicate balance between the lower midrange and the upper midrange presence region that is critical in establishing good localization and imaging.
I've had some success in EQ'ing the lower midrange / upper bass down a dB or two (subjectively, not trying to do it based on room measurements) when I'm forced to place speakers too close to the wall behind them, and a lot of the "lost" localization and imaging returns.
I haven't read the reference, but was any attempt made to correct for the change in frequency balance of moving the speaker closer to the wall behind it, or was it just moved with no other changes ?
Colouration is still a problem though, which as well as coming from comb filtering I think is related to resonances of the wall panels themselves. Typical wall materials will have their own resonant behaviours not unlike speaker panels, so any sound reflected by the wall will be coloured by those resonances.
The further away the speaker is from the wall behind it the less the relative amplitude of the reflection vs direct single and the greater the time delay, until a point is reached where the colouration from the wall drops below the perception threshold both in amplitude and time delay.
A little bit of judicious EQ of a speaker designed for free standing use when placed too close to the wall behind it can still give reasonable imaging though, whereas I find if the side walls are too close there's not really anything you can do with EQ to improve imaging or localization.
Probably the article by von Bekesy,"Auditory Backward Inhibition in Concert Halls". That was '71. This writer mentions it:
Sound structure in music - Google Books
Ashley did a riff
IEEE Xplore - Auditory backward inhibition can ruin a concert hall
- Status
- Not open for further replies.