the research Mr Jones refers to is the so-called "Archimedes" Eureka-funded research project: Project - EUREKA
it was joint project of Danmarks Tekniske Universitet, Bang & Olufsen and KEF
Bech is now Head of Research at Bang & Olufsen
on the part of Bang & Olufsen the follow-up to the project was the decision to licence the acoustic lens from Sausalito (of Moulton and LaCarruba) and to introduce their current line of horizontally omnidirectional Beolab speakers
what Jones says in this recent interview is the same what Moulton said in His 1999 interview for Recording Magazine: Moulton Laboratories :: Nick Batzdorf Interviews David Moulton
linked here many times, with no positive feedback (just no feedback or the Moulton's got His marketing agenda... blah blah stuff)
both gentlemen refer to the same "Archimedes" research project led by Bech, no wonder their conclusions are the same
yet when Jones says it then everyone should read it, but when Moulton says it it is just a childrens table
furthermore when Bech publishes His research papers He is great, but the fact that He is behind B&O-Moulton coooperation and Beolab line of omni speakers is conveniently overlooked:
http://www.bang-olufsen.com/graphics/bogo/News/BeoLab5-PR/BeoLab5_Pressrelease.pdf
From the interview with Andrew Jones:
"Which gets us back to what are, truly, the important speaker characteristics. It confirmed a lot of beliefs that ceiling reflections are one of the worst. Sidewall reflections can be good, adding to a sense of spaciousness, if the stereo speaker possesses well-controlled directivity. This same sidewall phenomenon happens in concert halls, on a different scale. The tall, narrow, long concert halls -- traditional ones -- were always the best halls. Sidewall reflections are lower-correlation than ceiling or floor, so they add spaciousness. Correlated signals, by contrast, add coloration."
To me the mention of correlation is a key part. There is a certain correlation
between the direct sound from the real sound sources and the reflections
(early to later ones).
And there is a certain interaural correlation in direct sound and reflected as well.
Correlation between direct and reflected sound will cause coloration, especially if
combing occurs somewhere say within the telephone bandwidth.
Too strong interaural correlation of the reflected sound will decrease spaciousness and
make the listening room more prominent to the perception.
My thinking is, that if you cannot manage to have the reflections (even the early
ones) diffuse enough, the only way is to avoid them by increasing the initial time
gap. The more diffuse reflections get, the more you can tolerate even the early ones without
combing artifacts causing to much coloration and having too much of image smear.
Because floor and ceiling reflections have higher interaural correlation than side reflections
and can follow the direct sound rather early, ceiling reflections should be diffused
(directed to the sides preferably) or (to some extent) absorbed in a broadband manner.
Carpets may adress floor reflections, some woofer arrangements in a speaker can reduce
LF combing due to floor reflections.
Whether to prefer absorbtion over diffusion to tame the perceptional role of early
reflections (or maybe make them "useful" in the thinking of some) is IMO answered
quite easily:
If you cannot make absorption near the speakers broadband and well balanced to leave
(get) the power response flat (or smoothly falling with frequency) you are better off
by using diffusivity. It is surely not forbidden to make use of both ...
If you have speakers with high DI you may want less absorption than when using
a speaker with lower directivity. A speaker having a directivity pattern
discontinuously variing with frequency causes the most trouble, because you cannot
find a consistent strategy in a real room to handle that, and the success
of any strategy will be highly dependent from listening position.
In the end the treatment of the room has to match the dispersion characteristics of the
speaker and its positioning. There is no room treatment without knowing what speakers
i have.
Having low DI speakers in a room with low absorbtion at least calls for high
diffusivity and listening distance quite close, if "critical listening" is the goal.
"Which gets us back to what are, truly, the important speaker characteristics. It confirmed a lot of beliefs that ceiling reflections are one of the worst. Sidewall reflections can be good, adding to a sense of spaciousness, if the stereo speaker possesses well-controlled directivity. This same sidewall phenomenon happens in concert halls, on a different scale. The tall, narrow, long concert halls -- traditional ones -- were always the best halls. Sidewall reflections are lower-correlation than ceiling or floor, so they add spaciousness. Correlated signals, by contrast, add coloration."
To me the mention of correlation is a key part. There is a certain correlation
between the direct sound from the real sound sources and the reflections
(early to later ones).
And there is a certain interaural correlation in direct sound and reflected as well.
Correlation between direct and reflected sound will cause coloration, especially if
combing occurs somewhere say within the telephone bandwidth.
Too strong interaural correlation of the reflected sound will decrease spaciousness and
make the listening room more prominent to the perception.
My thinking is, that if you cannot manage to have the reflections (even the early
ones) diffuse enough, the only way is to avoid them by increasing the initial time
gap. The more diffuse reflections get, the more you can tolerate even the early ones without
combing artifacts causing to much coloration and having too much of image smear.
Because floor and ceiling reflections have higher interaural correlation than side reflections
and can follow the direct sound rather early, ceiling reflections should be diffused
(directed to the sides preferably) or (to some extent) absorbed in a broadband manner.
Carpets may adress floor reflections, some woofer arrangements in a speaker can reduce
LF combing due to floor reflections.
Whether to prefer absorbtion over diffusion to tame the perceptional role of early
reflections (or maybe make them "useful" in the thinking of some) is IMO answered
quite easily:
If you cannot make absorption near the speakers broadband and well balanced to leave
(get) the power response flat (or smoothly falling with frequency) you are better off
by using diffusivity. It is surely not forbidden to make use of both ...
If you have speakers with high DI you may want less absorption than when using
a speaker with lower directivity. A speaker having a directivity pattern
discontinuously variing with frequency causes the most trouble, because you cannot
find a consistent strategy in a real room to handle that, and the success
of any strategy will be highly dependent from listening position.
In the end the treatment of the room has to match the dispersion characteristics of the
speaker and its positioning. There is no room treatment without knowing what speakers
i have.
Having low DI speakers in a room with low absorbtion at least calls for high
diffusivity and listening distance quite close, if "critical listening" is the goal.
Last edited:
From the interview;
"You have a hierarchy: a mathematician, a physicist (which is a failed mathematician), and an engineer (which is a failed physicist)."
IMO he got it backwards (knowing enough physicists myself which is how he describes himself that comes as no surprise to me.) A mathematician lives in a world of pure abstraction apart from the real world. He lives in a system of closed logic all its own. Physicists use mathematics to explain the real world. Their mathematical models can never be exactly right of course, they are inherently approximations but they are often useful. Their theories stand until they are disproven by examples that cannot be reconciled with their current theory and so it is dismissed even if a better one hasn't come along. Naturally they don't have the time to devote to develop their knowledge to as much mathematical skill as mathematicians do. Engineers must take the knowledge gained by physicists and other disciplines such as chemists, biologists etc. and create circumstances in part of the real world that operate predictably, reliably to control that world to the degree and in a way that those they design for expect and require. Their mistakes can have catastrophic consequences such as a defective vehicle or nuclear power plant design. They must learn and use physics and mathematics to understand the problem they are trying to solve and how to apply it. Those who pay them don't accept excuses for failure and don't give them forever to come up with answers. Naturally engineers don't have the time to devote to learn as much physics as physicists know or as much mathematics as mathematicians know. I'm an engineer, my mother was a mathematician. In college my roommate was a physics major. I respect these people for what they know but I recognize that their knowledge while useful is invariably an abstraction while mine must be applied in ways that prove their worth in the world of reality, not just of the mind.
"You have a hierarchy: a mathematician, a physicist (which is a failed mathematician), and an engineer (which is a failed physicist)."
IMO he got it backwards (knowing enough physicists myself which is how he describes himself that comes as no surprise to me.) A mathematician lives in a world of pure abstraction apart from the real world. He lives in a system of closed logic all its own. Physicists use mathematics to explain the real world. Their mathematical models can never be exactly right of course, they are inherently approximations but they are often useful. Their theories stand until they are disproven by examples that cannot be reconciled with their current theory and so it is dismissed even if a better one hasn't come along. Naturally they don't have the time to devote to develop their knowledge to as much mathematical skill as mathematicians do. Engineers must take the knowledge gained by physicists and other disciplines such as chemists, biologists etc. and create circumstances in part of the real world that operate predictably, reliably to control that world to the degree and in a way that those they design for expect and require. Their mistakes can have catastrophic consequences such as a defective vehicle or nuclear power plant design. They must learn and use physics and mathematics to understand the problem they are trying to solve and how to apply it. Those who pay them don't accept excuses for failure and don't give them forever to come up with answers. Naturally engineers don't have the time to devote to learn as much physics as physicists know or as much mathematics as mathematicians know. I'm an engineer, my mother was a mathematician. In college my roommate was a physics major. I respect these people for what they know but I recognize that their knowledge while useful is invariably an abstraction while mine must be applied in ways that prove their worth in the world of reality, not just of the mind.
furthermore when Bech publishes His research papers He is great, but the fact that He is behind B&O-Moulton coooperation and Beolab line of omni speakers is conveniently overlooked:
http://www.bang-olufsen.com/graphics/bogo/News/BeoLab5-PR/BeoLab5_Pressrelease.pdf
OT but Bech's AES papers are generally phenomenal. He's one of the most under-appreciated lights in audio IMO.
Dave
The latter eaily done by havinf a sloped ceiling. My room has virtually no ceiling reflections, but (probably) strong floor reflections. Room sounds very good.
http://www.diyaudio.com/forums/blogs/planet10/493-fall-listening-space.html
dave
"Easily done" say you? Some of us live in ordinary houses. Are we unworthy of good sound? OK it can be done but I agree, the ceiling is a major issue. I've addressed it with a kind of line source approach but it requires multiple midbass drivers and enclosures with more than 6 pieces. Oh well. Most of the extra bits aren't real big.
I don't think there is any research to support that reflection absorption has to somehow correct off axis frequency response. Certainly the case for some required power response is not supported by any of Toole's research (or Bech's) and if a particular power response is not required then it is hard to argue that a particular response of each room reflection is required.
The 2nd Bech paper actually went into varied frequency response of different reflections. Both the first and second studies simulated a speaker in a typical room by simulating the strength, delay and arrival angle of the first 17 reflections. In the first study he used a generic polar curve for the speaker that changed level vs. angle. In the second study he used the exact frequency response the simulated speaker would have at each radiation angle. The general effect was to give some reflections considerably less level in the 500 to 2000Hz range. Rather than noticing some particular timbral change from this it mearly shifted the threshold of detection by an amount similar to the reduced midrange level.
This also follows the work of Lipshitz and Vanderkooy that found that the ear was fairly insensitive to holes in the power response.
It is an appealing audiophile notion that wall reflections must have a spectral balance compatible with the axial response but there isn't any proof of it that I have found.
David S.
My own observations say that the thing that works best in a small room is to limit the amplitude of direct reflections from surfaces that are large like a ceiling, floor or wall... it is also useful to keep direct reflections limited via diffusion where that is an option.
The general effect of high amplitude short delay reflections causes a variation in spatial information and "blenderizing" (tm) of the soundfield.
In this respect lower frequency reflections are less noticeable... but not meaningless in sonic terms.
Almost all of us are listening in a small room.
Ymmv.
_-_-bear
The general effect of high amplitude short delay reflections causes a variation in spatial information and "blenderizing" (tm) of the soundfield.
In this respect lower frequency reflections are less noticeable... but not meaningless in sonic terms.
Almost all of us are listening in a small room.
Ymmv.
_-_-bear
David,
i can remember you correcting that "audiophile notion" from the viewpoint
of known research more than once and i have no problem in admitting, that
postulating "flat" or some special (falling) curve in power response seems
an exageration and thus my wording was chosen too certain in this point.
On the other hand:
- were there always speakers used for comparison in cited research,
which were capable of continouus power response (flat or falling) ?
- is it possible that an energy whole between say 500Hz and 2000Hz has
become a kind of convention in the consumer's small room
listening situation, due to common (non CD) multiway speaker's dispersion
characteristics ?
- would you agree that e.g. a relatively extended decay in (lower) midrange,
which is present in some (e.g. dry walls and HF-absorbers only)
rooms, is not contributing to "good sound" and that a
balanced decay (usually longer towards LF) is preferable ?
- would you agree that if extended decay time in a certain frequency range
coincides with a hump in the speakers power response, these two factors could
make up a detrimental effect ?
If there should be some agreement, i would preferably concentrate room
treatment (e.g. in upper bass/lower midrange) near/behind the speakers.
And why not adressing reflection issues from a side wall (or a low and tilted
ceiling under the roof) too close to a (non CD and for some reason not sufficiently
toed in) 2-way speaker with broadband absorbers fully functional somewhere below
the crossover frequency of the tweeter ? Which does not necessarily mean killing all
ipsilateral reflections.
Those are situations i found in real homes and i am not sure whether all those
(real and often real bad) situations are always covered sufficiently by common
studies.
Last edited:
OT posts removed.
The Lipshitz and Vanderkooy paper directly listened to a variety of speaker power responses Including flat power from rising on axis response (scorchingly bright), flat power response with flat axial response (still overly bright) and more normal power responses with flat axial. They found the later were fine and that holes in the power response were fairly benign. Peaks in the power response were undesirable.
If you look at the Toole studies where he tests a lot of conventional speakers you can see a common thread to power responses but there are 2 norms: the 2 way norm and the 3 way norm. Have we gotten used to them and even adapted our recording techniques to them? Not likely but not impossible.
Seems on the face of it preferable, but that is actually the point of contention: is there evidence that it is necessary or does it just seem like a good idea?
Here is something to consider. We talk about reflections as if they are heard in issolation and have a particular frequency spectrum that is important. The reality is they combine in a particular way with the direct response based on their delay (and perceived based on direction and the ears inherent time windowing). So a flat spectrum reflection will combine with the direct sound for a very comb filtered response. The major deviation is around the first or second comb filter nulls. I have listening simulations and there are always major pitch effect determined by the particular delay. Flat plus flat does not equal flat.
At higher frequencies the comb filtering is dense, within a critical bandwidth, and the reflections add brightness.
I would follow the Bech papers and see that the floor bounce was most likely to be audible and treat it first. If speakers were near a side wall I would absorb or diffuse that bounce, probably based on whether the room generally needed absorption or not. I am intriqued by the possibility of redirecting early bounces to force multiple bounces and later arrival.
David S.
Were these tests done in mono or stereo? If they were done in stereo, they would be rather useless, because there are other effects which significantly affect the perceived brightness. It would be incorrect to draw a conclusion about "correct" power response from a stereo test.
Rather useless? I don't know why an experiment where you directly manipulate the variable of interest would end up be "rather useless" in either mono or stereo.
AES E-Library Experiments in Direct/Reverberant Ratio Modification
Here, read the paper and contact the authors. Let me know their response.
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.