Not my favorite study thats for sure. It was just around here and it was one that others on the forum kept throwing in my face based on falsely drawn conclusions - not the papers their own. But really again do I have to even have a controlled study to determine that if you alter one channel of a stereo signal's phase that it will mess up the imaging? Come on man you say you have mixed this stuff should be easy and I really don't understand why someone with an ounce of common sense hasn't backed me up on this point. Yet ironically Dr Geddes repeatedly is using this as some sort of example of my stubbornness and lack of a want to look at contradicting evidence. He fails to notice that despite my better judgment I did independently try to test for it again - doubting myself and double checking - yet he has the brass to say he wasn't paying much attention. Well what is it? I really don't see what I have done to him to make him view me as a threat or whatever. I just wanted to talk about some silly audio stuff.
Hi Key
I am not criticizing you so much as the audio world in general, that these things are simply not understood. I have to admit to not having the slightest idea about what you are claiming, I am only saying that your approach is just NOT the way to do science and therefor I will have to decline to participate.
I am not trying to be overly critical, but we - audiophiles - have to stop being so casual in our endeavors. That the audio world is in a serious state of decline because of its tendancies at rash conclusions is quite demonstable (just take wire discussions for example).
I would love to discuss the "phase audibility" effects that you are talking about, if you care to describe them in an objective way, as well as all the prior work that has been done in this area, but I have no interest in "listening" to examples of something that I have no idea has any merit.
And the study that you references is truely disturbing. I have glanced at the study and I have read his web site. The claims on the web site do not follow from the study, even if the studies results are correct. But a glance at the study was itself disturbing since it appeared to have a lot of loose ends. At any rate, I did not pursue this any further since I simply do not see the relavence of the work to anything that I am doing, so I ceased to pursue it. I simply don't have the time to follow all of these kinds of trails.
And please don't take any of my comments personally - you simply hit a nerve.
I am not criticizing you so much as the audio world in general, that these things are simply not understood. I have to admit to not having the slightest idea about what you are claiming, I am only saying that your approach is just NOT the way to do science and therefor I will have to decline to participate.
I am not trying to be overly critical, but we - audiophiles - have to stop being so casual in our endeavors. That the audio world is in a serious state of decline because of its tendancies at rash conclusions is quite demonstable (just take wire discussions for example).
I would love to discuss the "phase audibility" effects that you are talking about, if you care to describe them in an objective way, as well as all the prior work that has been done in this area, but I have no interest in "listening" to examples of something that I have no idea has any merit.
And the study that you references is truely disturbing. I have glanced at the study and I have read his web site. The claims on the web site do not follow from the study, even if the studies results are correct. But a glance at the study was itself disturbing since it appeared to have a lot of loose ends. At any rate, I did not pursue this any further since I simply do not see the relavence of the work to anything that I am doing, so I ceased to pursue it. I simply don't have the time to follow all of these kinds of trails.
And please don't take any of my comments personally - you simply hit a nerve.
Last edited:
Yep the conclusions they jump to are way off imo. And the ones that audiophiles are drawing from it are even further off. But anyway it was partly my motivation for posting examples of high time distorted music or just trying it myself. Usually I keep to myself but I thought this is DIYaudio.com and what better place to try it yourself. It wasn't meant as a scientific study so if you don't feel like participating that's cool with me. But don't come into the thread and dismiss it as being inaccurate without even taking a minute to examine the specifics.
And to me this is even more ridiculous than a cable debate that I would even have to argue that the phase characteristics of two loudspeakers in a stereo set should match. There is nothing subjective about it - it's like you were trying to debate with me on with the stance that the sky isn't blue because I haven't done any controlled testing and I have no degree. You offered no insight or specifics as to why what I was trying to clarify about one of your prior sweeping statements for the good community was wrong except that you do authentic controlled tests so therefor your opinion is more informed than mine and you don't even need to bother checking any real world examples I can present no matter how easily verifiable and repeatable because they are automatically tainted.
Honestly you have struck o my nerves and I let it slide a couple times already and I frankly think it's just as much audiophoolery to continue such a conversation about something so elementary as altering the arrival times at the ear will lead to a change in the stereo imaging.
Earl why so overcautious? That's the test configuration:
Looks more like a horizontal summing localization experiment set-up to me.
Best, Markus
An externally hosted image should be here but it was not working when we last tested it.
Looks more like a horizontal summing localization experiment set-up to me.
Best, Markus
Is that what you were presumably testing? That two loudspeakers in a stereo pair need to be closely matched? I think that is well known. But I recal some satatements about the audibility of symetrical distortions versus non-symetrical ones. No test that I know of has found a significant effect in that regard, although there is reason to believe that such a thing could happen because the ear is not symetrical. And those prior tests that DID find some statistically significant effects failed to control for a confounding variable, namely asymetrical nonlinearity. Unless the intervening system in the test is perfectly symmetrical then changing the symmetry of the source - the control variable - changes the total transmission, thus confounding the results with a highly correlated alternate variable.
Last edited:
Thanks Markus
I didn't want to come down too hard on Key since he isn't making any mistakes that most other people would'nt make as well. Much of what he has written has been open minded and objective - much more so than many. But as with most people, he seems to have held onto his own work a little too closely. Scientists learn that attacks and questions of thier results is inevitable and after awhile they come to accept such things as necessary. The novice often finds the attacks offensive and becomes defensive - both correct to an extent. But the defense must be purely objective without any subjectiveness or bias, and thats the difficult part.
No. You came into the thread which was only about symmetrically distorted transient response and made a blanket statement.
"The only phase that is in the least bit audible is non-minimum phase or group delay - more meaningfull because it is in the time domain."
I tried to clarify that statement and show you that you are overlooking some scenarios. Namely that if I was to do the same test only apply an allpass filter to one channel of the signal that it would be obviously audible because it would change the stereo imaging. You without trying it yourself shot back with another unfounded statement.
"My points are based on actual controlled subjective tests not opinions. The data says that there is no difference between symmetrical distortions and nonsymmetrical ones and in general in a loudspeaker neither is significant."
And I am sorry but this is elementary speaker wiring. The speakers must be in proper phase relationship or the stereo imaging will be off. I don't need a controlled study to show you. All I need is for you to actually think with me on the same logical topic. If you perform a phase shift on one signal of a stereo signal and not the other, do you not alter the arrival time at the ears of the listener? And would this not shift the stereo imaging? Easy stuff here.
I certainly never understood that you were talking about doing different things on the two channels of a stereo pair. When you mix discussions of doing things to both channels or one channel only, those are completely different things. What's audible as a difference in two stereo channels is completely different than what is audable when done to BOTH channels. This was never clear to me. Everyone is going to assume that what we talk about as "audible" occurs equally on the stereo pair or it is done in mono, as all my work is. If you are doing different things to the two channels then this needs to be made explicitly clear as no one is going to automaticly realize that this is what you mean. If you did make this distinction clear and I missed it then shame on me. But if you did not make this critical distinction clear and/or changed in the middle of the discussion then shame on you.
I guess I tried to make it clear. I gave two examples that I thought were simple enough for you to see what I was getting at - all pass applied to one channel of a stereo signal, and a mismatched polarity between a stereo set of speakers.
All I was trying to do was point out that exact detail with the thread. It's HOW you distort the time which can make it audible. I kept seeing the figure 6 microseconds thrown around without any specific qualification of that time margin. I tried pointing it out to them that errors which happen on only one side of your head and not the other are much much easier to spot than errors which happen equally to both sides of the head.
Having someone with your credibility entering the thread and just say something which is at complete odds with what I was proving to myself and trying to give an opportunity for others to prove to themselves. It forced me to do a little side track. I thought "oh someone will listen to this and say "well that's obviously audible compared to the symmetrical distortions"" and we could move on. But you simply didn't want to get what I was saying or just thought it was tainted and waste of your time. That's fine and all but I personally think that someone could walk away from that thread siding with you - much more credible than me - and they would be very wrong in that conclusion.
All I was trying to do was point out that exact detail with the thread. It's HOW you distort the time which can make it audible. I kept seeing the figure 6 microseconds thrown around without any specific qualification of that time margin. I tried pointing it out to them that errors which happen on only one side of your head and not the other are much much easier to spot than errors which happen equally to both sides of the head.
Having someone with your credibility entering the thread and just say something which is at complete odds with what I was proving to myself and trying to give an opportunity for others to prove to themselves. It forced me to do a little side track. I thought "oh someone will listen to this and say "well that's obviously audible compared to the symmetrical distortions"" and we could move on. But you simply didn't want to get what I was saying or just thought it was tainted and waste of your time. That's fine and all but I personally think that someone could walk away from that thread siding with you - much more credible than me - and they would be very wrong in that conclusion.
If thats the case then I appologize because that was not what I understood nor what I intended to do. When I saw the words "symmetrical distortions" this means something very precise to me, but its not what you meant. So this was all a misunderstanding. You meant symetrical from channel to channel and to me it means from plus to minus on the waveform - completely different things. One is highly audible and the other is not.
It's always difficult to jump into a thread and know exactly what is going on. Specifically when the terminology can differ from thread to thread. You (I) think that you understand but you don't.
Earl,
Will you be posting your white paper on this thread or your website (or both)?
Thanks!
Both
Only this - I had it done or about 98% finished, then the program crashed and I lost a few hours of work so I am back to only about 90%. A few more days.
That stinks.
This might be of some interest to those who followed this thread and my other thread on driver break up. I replaced my cheapy eminence tweeter with the Selenium D220t. Here's the results:
on axis
An externally hosted image should be here but it was not working when we last tested it.
11.25 off axis
An externally hosted image should be here but it was not working when we last tested it.
22.5 off axis
An externally hosted image should be here but it was not working when we last tested it.
33.75 off axis
An externally hosted image should be here but it was not working when we last tested it.
45 off axis
An externally hosted image should be here but it was not working when we last tested it.
The tweeter definitely has a wider dispersion in the upper octaves, but little less extension on the low end. I have a couple more questions:
1)The wider dispersion thing, does that mean the wavefront is more spherical at the mouth of the cd/throat of the horn?
2)When I actually mount the WG on a baffle will it get better low end extension?
Thanks again Gentlemen!
Dan
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.