Then standardize the room if it makes a difference. There is an IEC standard. There is also a North American (and European, and Asian) typical room in terms of dimensions and average absorption.
I'm not sure that gets you anywhere.
If we are speculating on the ideal speaker we can speculate on an ideal room for it and the room doesn't have to match any norms or present standards.
Will any speaker type (response and directivity and polar pattern) coupled with any room create a better music experience than all others?
Bring back the Bose 901!
David S.
Depth of abstraction layers to general perception in enormous. Ability to focus down to subsets and particular layers of abstraction in profound.
Take written language for example; recurring letters, words, phrases, all are ignored in comprehending underlying meaning. Mosaics are another fine example. And then look at musical composition, expression and comprehension.
Musical notation may be played mechanically, devoid of dynamics and variation in timing; as such it is an empty carrier for musical expression and meaning to be imbedded in.
The data is the universe; our senses conduits and perceptual mechanisms filter banks.
Take written language for example; recurring letters, words, phrases, all are ignored in comprehending underlying meaning. Mosaics are another fine example. And then look at musical composition, expression and comprehension.
Musical notation may be played mechanically, devoid of dynamics and variation in timing; as such it is an empty carrier for musical expression and meaning to be imbedded in.
The data is the universe; our senses conduits and perceptual mechanisms filter banks.
Those standards aren't strict enough nor adhered to in each and every studio. So, no they don't get us far.
Currently the speaker is thought of as the end of a transmission chain. This is not correct. Stereo "happens" at our ears. Our HRTF is part of the decoding, so is the room. We would need to look here for what is possible and desirable. Virtually no research in that area.
My conclusion of this discussion (so far..) is that the big picture is same for us all. The discussion is about correlation coefficients of spesific factors and weighting some factors differently. This is normal practise. Even if some day there will be a new standard, someone still disagrees about the paradigm. I remember when Kopernikus started arguing something about the sun...
Perception is both, but for the most part our perception of sound qualtiy is tied to the early sound. For the 20 speakers in Toole's first listening tests the anechoic axial response was enough to rank order the systems by listener preference. The power response absolutely would not allow rank ordering. No correlation.
I see it more as the power response (and therfore directivity index curve) must fall within an acceptable range not to muck things up. If it does, then the direct response will determine sound quality.
You are correct that the reverberent field does not dominate the direct, at least perceptually. However, even in a fairly dead domestic room we will be at or a little beyond the critical distance. In larger spaces we can be well beyond the critical distance. For those cases the reverberent energy will dominate conventional measurements.
The thing I am always trying to point out to others is that plunking down a microphone and measuring the steady state response can easily lead us astray. The steady state curve is not the perceptual curve. The bigger the room the more misleading it will be. This is why we can sit in a cinema equalized to the X curve and still enjoy the sound track. (Also why I am scratching my head now that Sean Olive is jumping on the room correction bandwagon.)
David
Read Kates and Salmi
FFT, Impulse response with frequency related time window. Long at low frequencies (essentially steady state) and short at high frequencies (essentially the direct).
David
Dave your conclusion from your first point is, of course, incorrect. The data would only imply that the correlation of the direct field is greater than the correlation of the reverberant field, but you cannot say that there is "no correlation".
And, I could not agree with your final point more. I have said this time and time again, that the speaker has to be designed such that the direct field is near flat and smooth and that EQ of the reverberant field can only mess up the direct field. However, it is possible for the reverberant field to be so bad that EQing it will improve the overall perception (even Olive notes this). But if the direct field and the reverberant field are comparable, i.e. the DI is fairly flat and smooth, then no correction of the EQ reverberant field should be required. If the reverberant field is not as flat and smooth as the power response of the loudspeaker then something is seriously wrong with the room. Will "room EQ" fix this? It might "improve it, but it is the room that needs the fix, not the speakers. As Toole points out the rooms steady state response does tend to track the loudspeakers power response. Thus a near flat and smooth DI is going to be a desirable feature of the loudspeaker.
This is in essence what I do when I use frequency dependent smoothing according to Moore's ERB data. This is the same as a frequency dependent time window, except that it is not a linear change as is usually done.
I did but I don't remember them showing the end-all approach of measuring timbre of stereo reproduction in acoustically small rooms.
Again, yes and no. Weighting some factor differently is fine as long as those weights make sense given the facts. For example, weighting phase below 700 Hz as high is not reasonable, so its not "OK".
Yes, my point was incorrect. I should have said there was a negative correlation. A number of speakers that were well ranked had quite messy power responses. At the bottom of the rankings there were a couple with quite smooth and flat power response (their horrid frequency response happened to average out nicely when spatially averaged).
You really should pull out the "part 1 and part 2" papers and look at the ranked ordered response page again as it is very enlightening.
To make it easy I'll add scans of the good bits. 0-15 is the on axis listening window average. Clearly units with wide and smooth response are near the top of the list. "Power Response" is the same speakers in the same rank order but with power response shown. 1, 2, and 4 not so good (3 is nice). 10 is the nicest power response possible, but a so-so ranking. 12 is nice. 16 is very nice and 18, two from the bottom of the rankings, is also nice.
What is the correlation between power response and listener preference?
David
Attachments
Last edited:
Their goal was to model human hearing and mimic it with a more correct measuring system. Once that is done we should be able to measure perceptual frequency response in any room.
David
Aalto University in Helsinki, Finland is one of the leading universities in acoustics, here are some of the latest dissertation papers about sound localization.
https://aaltodoc.aalto.fi/handle/123456789/52/discover
Departmnt of Signal Processing and Acoustics Communication acoustics: Spatial sound and Psychoacoustics - Department of Signal Processing and Acoustics - Aalto University
Aalto is a wave in English! Aalto University was named after one of it's students, architect Alvar Aalto. His wife and colleague was Aino Aalto, who designed these
https://aaltodoc.aalto.fi/handle/123456789/52/discover
Departmnt of Signal Processing and Acoustics Communication acoustics: Spatial sound and Psychoacoustics - Department of Signal Processing and Acoustics - Aalto University
Aalto is a wave in English! Aalto University was named after one of it's students, architect Alvar Aalto. His wife and colleague was Aino Aalto, who designed these
Last edited:
Juhazi - something wrong there.
Dave - you are doing an eyeball "correlation" and I don't accept that. We already agree that the direct field is dominant, but we seem to disagree on the weight that should be put on the power response. You sometimes sound like you think that it should be zero, but that is just not supported by the data. I don't think that it is high, certainly not as high as the direct component, but it is certainly not zero.
I think that #3 contradicts your position because its direct response is not that great but its power response is and it ranks rather high. Clearly it is NOT all direct field!!
Dave - you are doing an eyeball "correlation" and I don't accept that. We already agree that the direct field is dominant, but we seem to disagree on the weight that should be put on the power response. You sometimes sound like you think that it should be zero, but that is just not supported by the data. I don't think that it is high, certainly not as high as the direct component, but it is certainly not zero.
I think that #3 contradicts your position because its direct response is not that great but its power response is and it ranks rather high. Clearly it is NOT all direct field!!
Last edited:
In Toole's own words: "In fact, if the fine structure were removed from these curves (power response averages) by one-third-octave or wider bandwidth filtering, it would be difficult to discern any trend whatsover in the data." That sounds to me like zero correlation.
Now you are doing eyeball correlation.
Lets agree that the link between axial response is stronger than in the power response and that there are strong counter-examples in the power response curves (well ranked speakers with poor power response and poorly ranked speakers with good power response).
Which parameter are you going to measure and optimize when you want a speaker to be a success in the marketplace?
Now, I admit it would be interesting if he had a group of 20 speakers with identical axial response and a wide range of power response curves....
David
If anyone would like to read the paper that Dave is referencing, Harman has it linked on their site here:
Harman - Scientific Publications
It's titled "LoudspeakerMeasurementsand Their Relationship
to Listener Preferences
art 2*"
That's easy: Do it all.
But then appearance ranks above both the direct response and the power response, combined! so this is just an academic discussion as far as the marketplace is concerned.
About a week ago I was visiting a relative who happens to have a hu-mon-gous house, and who happens to be very very wealthy.
One thing I've noticed abut wealthy people is that they buy expensive cars, expensive homes, electronic gadgets and TVs. But they often have cheap speakers!
At one point in the evening we were watching a movie, and my relative complained that 'the movie wasn't loud enough.'
And then they fiddled with a bunch of remotes, and noticed that their Samsung soundbar wasn't turned on. Basically they were watching the movie and listening to the built in speakers that were in the TV.
The part that was astonishing to me was that there was virtually no difference between listening to the TV's speakers, and listening to the Soundbar. The Soundbar sounded no wider, no louder, no bigger, or less distorted. The Soundbar basically sounded exactly the same as the TV did.
And this is your typical consumer. They have a car in the driveway that costs as much as a house (seriously), but they're listening to movies over a $300 soundbar that is basically a step down from a $50 boombox.
I have another relative that's actually IN the music business, and it's the same story there. He's very well off, big expensive house, big expensive cars, but he listens to music and movies over a stereo that he bought at a department store. Bose would be a big step up.
Last edited: