That does seem to be an . . . issue . . . with the CBT design (for home use, anyway).
There are some weak links between what we measure and what we hear, but this are small compared to the weakness of "expert opinions". We argue here about things that IMO are not the main criteria of sound quality, but they are not irrelavent either. Its like my beef with Floyd Toole's book. I agree with about 90% and thats the real criteria, but we disagree on the desired sirectional response. We don;t disagree that whatever it is it should be constant. We need to consider things in context. If I look at data from a speaker I would look at just a few things: flatness and smoothness of the frequency response on the listening axis; flatness and smoothness of the DI along the listening axis (although a slight downward trend to the DI is apparantly desirable). All this above 200 Hz since below that the situation will be dicated by how the modal region in the room is handled. These four simple factors will tell you worlds about the design. Far more than what anyone says based on a simple listening test at some audio show.
but I don't understand the data without an "expert opinion"
but I am advised not to trust any expert opinions...
"catch 22"
That seems very odd to me as its not brain science to look at a frequency response and Directivity Index curve (as above). It does not require the reference to an "expert" to evaluate that data.
I've come to "trust" other people's (expert) opinion in the same way and for the same reasons that I "trust" their data . . . I do spot-check confirmations. If I hear what they report hearing (or not hearing) where I can confirm it then I give greater weight to their reports overall. I also pay particular attention to "confirmation bias" and my own tendency to hear what I want to (or think I should) hear.
It is certainly not easy with things audio, as our hearing is both adaptable and suggestable, and our "auditory memory" is limited and comprimised by our imagination. But we can learn . . . and learn to hear more "accurately" than comes naturally to us. Along with "looking at" the data . . .
Being aware of our human failings is a start, but I myself have been fooled more times than I care to admit. Toole and others have shown that reliable subjective data is possible, but its not easy. But how many of us take the time to 'do it right" - take the time to listen to many familiar examples over a long period of time before that we arrive at a conclusion. Certainly necessary care is the exception rather than the norm. In those cases that lack the "due diligence", I believe that simply looking at the data is going to be a better judge of quality than subjective opinions whether they are yours or someone else's.
if you believe in the superiodity of science, yes
but honestly, a scientist who only believe in numbers I would trust more as an accountant
but maybe that is the problem with science these days
too much focus on the money
and we cannot ignore that this whole thread from the very beginning is based on opinions from people who have been payed to have an opinion
so you would also know how to fool other experts...if thats what you wanted
please, only in theory, ofcourse
but honestly, a scientist who only believe in numbers I would trust more as an accountant
but maybe that is the problem with science these days
too much focus on the money
and we cannot ignore that this whole thread from the very beginning is based on opinions from people who have been payed to have an opinion
so you would also know how to fool other experts...if thats what you wanted
please, only in theory, ofcourse
Let us not forget that the above is also just an opinion.
I tend to believe what I've actually heard. I also tend to believe expert opinions given by those I trust. If I have consistently seen or heard top quality work from someone, I am willing to trust his advice and opinions. On the internet, when we don't know the person or the work, that can be difficult. We are right to beware.
Actually its not. It is well quantified result spelled out in the book that I referenced.Let us not forget that the above is also just an opinion.
I was always skeptical, but after reading the book that I referenced above, I am now convinced that no opinions are worth listening to. If its not hard data then its just blowing in the wind. Read that book and you will be embarrassed to make the statements that you made above.
I think he has nothing to be embarrassed about. The technical approach will get you to the target if done correctly- but setting the target is just as much art as science. This isn't an amplifier we're talking about, with a single valued output for a single valued input, a completely cut and dried bit of engineering- the "accuracy" term is a lot dicier for transducers in rooms.
What should be accurate to what? Are you trying to reproduce the exact soundfield around the listener's head that existed at the microphones (it can't just be at the earholes for a head that's not clamped)? I don't believe that you can. The best that a set of transducers can provide in a room is an illusion, and there's a LOT of conflicting literature about what it takes to do that. Toole has one approach, Linkwitz has another, you have another, Ken Kantor has yet another... What you all share is enough grounding in rigorous engineering to accurately hit the targets you set. What you don't share is the same target, and I submit that NONE of your targets are "accurate," and all provide an excellent illusion in one way or another. The illusions differ, but illusions they all are, and absent a major change in the way soundfields are sampled, encoded, and reproduced, I don't see how anyone can reasonably speak of "accuracy" in the context of speakers in a room.
What should be accurate to what? Are you trying to reproduce the exact soundfield around the listener's head that existed at the microphones (it can't just be at the earholes for a head that's not clamped)? I don't believe that you can. The best that a set of transducers can provide in a room is an illusion, and there's a LOT of conflicting literature about what it takes to do that. Toole has one approach, Linkwitz has another, you have another, Ken Kantor has yet another... What you all share is enough grounding in rigorous engineering to accurately hit the targets you set. What you don't share is the same target, and I submit that NONE of your targets are "accurate," and all provide an excellent illusion in one way or another. The illusions differ, but illusions they all are, and absent a major change in the way soundfields are sampled, encoded, and reproduced, I don't see how anyone can reasonably speak of "accuracy" in the context of speakers in a room.
"Skeptical" is good (in my opinion
), but there's more to evaluating "hard data" than just saying "hard data". It is, after all, an illusion that we pursue with our silly little toys, and not all that easy to map "hard data" to our illusions of success at creating it."Hard data" can tell me how flat the frequency response, or how high (or low) the distortion and (perhaps) what kinds of distortion "sound worse" . . . but that "sound worse" is itself subjective, and how flat the frequency response needs to be to "sound good" is also subjective and varies from listener to listener.
I "like" what I can convince myself I like . . . what more do I need to know?
*edit* looks like SY was posting what I was writing as I was writing it.
Last edited:
The way I look at it is this. We are trying to create the illusion of a reconstructed soundfield out of two points in space. Now I am very much on the side of measurements and the exactness of scientific proof.....I know this to be true because whenever I read a review It takes everything I have in me to hold back the roll-eyes and sometimes rage!
Having said that, I do understand that psychoacoustics is involved, (how we perceive information) its because of this fact, it becomes more than science, it is now experience....hence opinions. I mean the IMP won folks. Are we afraid that all we have learned is for not? I am as much flabbergasted as the next guy but we have to talk about it. I'm so confused over the results that I think it's a conspiracy
I wish Gary would come back and discuss this.....but maybe he's afraid of too much questioning and negativity. It does get exhausting.
Having said that, I do understand that psychoacoustics is involved, (how we perceive information) its because of this fact, it becomes more than science, it is now experience....hence opinions. I mean the IMP won folks. Are we afraid that all we have learned is for not? I am as much flabbergasted as the next guy but we have to talk about it. I'm so confused over the results that I think it's a conspiracy
I wish Gary would come back and discuss this.....but maybe he's afraid of too much questioning and negativity. It does get exhausting.
The way I look at it is this. We are trying to create the illusion of a reconstructed soundfield out of two points in space. Now I am very much on the side of measurements and the exactness of scientific proof.....I know this to be true because whenever I read a review It takes everything I have in me to hold back the roll-eyes and sometimes rage!
Having said that, I do understand that psychoacoustics is involved, (how we perceive information) its because of this fact, it becomes more than science, it is now experience....hence opinions. I mean the IMP won folks. Are we afraid that all we have learned is for not? I am as much flabbergasted as the next guy but we have to talk about it. I'm so confused over the results that I think it's a conspiracy
I wish Gary would come back and discuss this.....but maybe he's afraid of too much questioning and negativity. It does get exhausting.
I for one don't place much on the fact that the IMP won is a setup that was basically design for it to win. Maybe Earl can chime in here. I don't know how it works with the AES but, having presented numerous scientific papers at conferences and published even more in journals, my experience is that conference papers are not peer reviewed. Journal articles are. Conference papers which are deemed significant contribution are often selected for peer review and then appear in journals. Lesser significant papers fall by the way side.
So does the AES work the same way? If so, did the paper ever make it to journal status? And, does it really matter? And, after all the years, just what is it that we have learned, speakers that sound good sound good? After all, audio is like watching TV is a room lined with mirrors. Hey, I like that. I may have to add it to my signature.
Let's not throw the baby out with the bathwater. Yes there is psychoacoustics involved, we understand a great of that, so its science not "illusion". Are there things left that avoid quantification - yes. Are they dominate factors? Hardly. The bottom line to me is that loose listening tests and expert opinions have more variance in them than what we don't know about the relationship between objective measures and subjective judgments. Add to that the book that I was talking about and its claim of an almost completely disconnected relationship between reality and personal judgment (i.e. statistically it has been shown that the more convinced someone is that they are right more likely they are to be wrong) and I simply have to throw out the subjective.
But you are all entitled to you opinions (as useless as they are).
But you are all entitled to you opinions (as useless as they are).
Adding a rear tweeter to a dipole does not make the power response more uniform. It actually make is worse. There are other reasons for adding the rear tweeter.
When I say power response I mean energizing the room evenly through the entire frequency spectrum like a true omni would do.
Let's just compare an omni to a box speaker in that regard. The box speaker will require BSC, that means more energy is present in lower frequencies as far as illuminating the room.
Let's just compare an omni to a box speaker in that regard. The box speaker will require BSC, that means more energy is present in lower frequencies as far as illuminating the room.
Something about the test being discussed has bugged me and I think I've finally realized why. Ignoring for the moment that the test was judged only on the ability to produce a believable "audio scene" (as has been mentioned already) -- what about the fact that it was based on playing only three tracks of music? If more tracks, of different types and recorded in different environments, were used, would the results have been different should the listeners have begun to (hypothetically, now!) recognize all the tracks always sounding similar on some of the speakers? Kind of like the criticism of "direct/reflecting" speakers --that they apply a fixed reverb/ambience effect, one that can't be turned off?
Last edited:
You do, yes. You've stated that for years and it's what works for you and for most of your customers. That's great. It's your approach and your goals. It works for you.
But I do not have to throw out the subjective, nor does anyone else "have to." In the end, I have to like the way it sounds. That's the point of a home system, the final determinant. If it's an IMP or an Orion or an Abbey or a Bose, I have to like they way it sounds, it has to meet or exceed my expectations. Good engineering and design are greats way to get there, but they are not the reason to be of the speaker.
I really think that most of us have the same, or very similar, goals. We just get there by different paths.
A very good point. What do you think? You sat not far from the three judges at the Dayton speaker contest a few years back. I was obvious to us very quickly how a speaker sounded. Few changed character much with the second or third track (some did, tho).
If we had taken the highest scoring three and spent more time with them, would we have changed our minds? I really don't know.