Mike,
I've been reading bits and pieces of this thread and finding it to be interesting and even funny sometimes. Your observations on "stretching" are, IMHO, on target. It certainly is a factor in getting better looking FR graphs for your speakers.
FWIW: I dabbeled in painting some years back (I try to emulate the "Hudson River School") and had found it necessary to do exactly what you've mentioned to get the effect that was, to me, a better protrayal of what I wanted.
Best Regards,
TerryO
One is an average of direct sound from a variety of axis. The other is an average of direct sound + reflected sound from a variety of axis.
Like so many terms in audio, "Power Response" is a very poor term because it isn't sufficiently defined.
Toole's "sound power" essentially "hints" that we don't simply perceive direct sound as coming from a discreet axis. Of course if you look back to his mantra of direct over reflected sound, it shouldn't come as any shock that "sound power" is far more important than "power response". 😉
Recent posts remind me of a cute little book from a half century ago and still in print: "How To Lie With Statistics:"
Amazon.com: How to Lie with Statistics (9780393310726): Darrell Huff, Irving Geis: Books
Despite the title, the preface is clear that the purpose of the book isn't to teach the reader to lie, but to be aware of the lies and misrepresentations in, for example "alarming" statistics printed by newspapers and illustrated with graphs.
Amazon.com: How to Lie with Statistics (9780393310726): Darrell Huff, Irving Geis: Books
Despite the title, the preface is clear that the purpose of the book isn't to teach the reader to lie, but to be aware of the lies and misrepresentations in, for example "alarming" statistics printed by newspapers and illustrated with graphs.
Yes a great book.
We have to think like painters. In perception, we talk of "realistic" or "correct" perceptions (what psychologists call "veridical" - a deck of playing cards looks a few inches tall even when held 10 feet away). But like in the interesting examples above, to produce veridical perceptions may require interventions of various sorts.
Do the columns of the Parthenon look vertical, straight, parallel, etc.? You bet they do just as some of my better records sure sound kind of like orchestral sound (never mind for the moment that this can not be defined, let alone produced in my music room).
But as everybody ought to know by now, the ancient Greek architects did not make them flat or straight. Anybody not know this? Anybody have trouble seeing the relevance to sound (re)production?
Last edited:
I agree with "SL's idea" because I've been posting about it for years on other forums, having arrived at it independently long ago.
Forgetting about room contribution for a moment and only considering the inherent error in stereo itself, attached is an example curve of the heightened perceive treble in a center image. This model includes crosstalk.
This is no way a universal definitive answer since there is significant variability in HRTF from person to person but this illustrates the average trend.
Standard pan pots don't eq this out, and equ'ing it out in the speaker results in error for all left or all right signals, but its hard to argue that stereo is perfect and flat response is always correct. This is just one of many tonal errors inherent to stereo.
Dave Dal Farra
Forgetting about room contribution for a moment and only considering the inherent error in stereo itself, attached is an example curve of the heightened perceive treble in a center image. This model includes crosstalk.
This is no way a universal definitive answer since there is significant variability in HRTF from person to person but this illustrates the average trend.
Standard pan pots don't eq this out, and equ'ing it out in the speaker results in error for all left or all right signals, but its hard to argue that stereo is perfect and flat response is always correct. This is just one of many tonal errors inherent to stereo.
Dave Dal Farra
Well I don't. Probably Terry doesn't either. Tho I sure don't have any clear guidelines to this.
Does it sound real - or is it correct? That's a pretty refined choice, actually. Most folks just want it to sound good. (More bass, please).
Hocus pocus.... applied physics will interpret human perception.
The staff consists of a professor of physics and a techie from the electric propulsion group.
Who are they kidding (besides the Shostack Foundation)?
Gimme a break!
Hocus pocus, eh? Looks to me like this physicist has done his homework.
You might check the bibliography from this paper:
http://www.princeton.edu/3D3A/Publications/BACCHPaperV4d.pdf
Optimal Crosstalk Cancellation for Binaural Audio with Two Loudspeakers
Edgar Y. Choueiri
Princeton University
choueiri@princeton.edu
Frank - I am very glad that you introduced that paper by Physicist Choueiri because it will help me explain my point of view.
The Physicist is interested in some kind of perfection of "3D" sound perception using two speakers (and for some reason probably related to getting research grants) he thinks commercial surround sound is faulty. His paper has many pages of detailed math but no pages of human testing.
The paper has references to human testing - I think that's what you meant by "done his homework." Of course, I could write a paper with many references to strength of steel and even show I had a basic mastery of this subject matter. But take my word for it, if I designed a bridge with steel, you would not want to drive across it.
If you look up the Bauer effect for stereo headphone listening, you will find that all or most of the parameters he is interested in have been studied as much as they need to be. I have a gizmo I made 50 years ago that implements some head-in-the-way frequency issues for headphone listening although, at the time, no easy way to work with time shifting. Since the dawn of stereo time, people have been promoting various stereo enhancers, though without the mathematical sleight-of-hand of Dr. C. Guess what - they really work well too! These are old and quite self-evident principles.
Even with my "steel bridge" example, have I failed (once again) to explain the difference between the psychological study of human perception and the mathematical analysis of sound waves? Funny, I failed at this for many years when I taught engineering students.
It would be very disrespectful of me not to add that I am a hifi enthusiast and an amateur like most of the the crowd at this website. We are self-taught in acoustic science but, I have great admiration for what I read here. On the other side, you have people like Toole who seems to be pretty good in psycho-acoustics despite coming from engineering/science or his very talented successor at Harman who is trained in psychology. Yes, you can have a good grasp of something coming at it as an amateur. Fortunately, in this forum, there are few "steel bridge decisions" we have to make.
Last edited:
Bentoronto, I am glad you brought this up.
1. There is frequently an arrogance from folks outside the field when they come in and do what they are comfortable at (in this case the math modelling) and assume the rest is trivial (the psychoacoustical measurement, conceptualization, and background literature). Well folks, it turns out that it is not trivial.
2. Toole was not only an engineer, he also had a background in psychophysical (behavioral) measurement. In fact, something we published (in psychoacoustics) a couple of years ago cited some his (psychoacoustic) research from the mid 1960s. After all these years it is still relevant. Do you think the fellow mentioned above is going back in the older leiterature from 4-5 decades ago?
1. There is frequently an arrogance from folks outside the field when they come in and do what they are comfortable at (in this case the math modelling) and assume the rest is trivial (the psychoacoustical measurement, conceptualization, and background literature). Well folks, it turns out that it is not trivial.
2. Toole was not only an engineer, he also had a background in psychophysical (behavioral) measurement. In fact, something we published (in psychoacoustics) a couple of years ago cited some his (psychoacoustic) research from the mid 1960s. After all these years it is still relevant. Do you think the fellow mentioned above is going back in the older leiterature from 4-5 decades ago?
Bentoronto,
I'm not sure how to interpret this. You mean you and him published articles together? What is your affiliation with Toole?
Wrong guy.
I published - and he (Toole) was cited.
My point is that Toole has a history in psychoacoustic research and interestingly his work from all those years ago is still relevant.
Sorry for the ambiguous pronouns
Absolutley, as is the side wall and even later reflections. Doesn't matter for this discussion.
What I'm illustrating is a tonal error inherent to stereo independant of the off axis response to cast light on the limitations of "flat is the only right answer, all the time" orthodoxy.
I think it also explains why different people perceive the tonal response of speakers slightly differently. Very often this is expressed as high frequenciy differences, usually attributed to taste. Based on this data, its not only taste that informs these perceptions. Different people hear different levels of error due to the 30 degree spacing of the speakers and the uniqueness of the individualized HRTF.
Toole's 1985 study didn't get into this level of shades of gray, it had bigger fish to fry.
Dave
I also prefer acoustic instruments/music for evaluating loudspeakers. But what about your listening levels?
- Status
- Not open for further replies.