keladrin said:
Instead of "wasting brainwaves" (which I do not believe to be the case), ask.
When a DBT is performed, ask the following questions...
1. Does the source present to the speakers, the correct relative amplitude and timing to simulate an object in space? Does it have the correct ITD and IID for the spacial location that is being simulated? Or is it a mixdown which uses ONLY constant power panning without time shifts?
Big list of questions, eh??
Current source material does not present any semblence of reality to what we are hearing...Constant amplitude pan pots do not present reality...they are a slap-dash method of moving an image right to left by gross negligence, omission of the other part of what we use to localize, left to right delays.
Humans adjust to the aural presentation...if you give us long enough, we adapt to the stimulus..key word, long enough..
I don't see that in testing to date..
Cheers, John
Are you suggesting that double-blind listening tests have only used multi-tracked recordings as their source material?jneutron said:
Just out of curiousity, I'm going to ask a friend of mine to help conduct some ABX type tests the best I can thing of with speaker cables and interconnects. I'll report back later this week.
Dumbass said:
Some do. Others don't.
What is important is...the use of a multitrack recording that mixes down to two channels via a pan pot, introduces an apparent localization using only one half of the information humans use. Pan pots produce a soundfield that does not exist in nature..ever.
So, why would we use an artificially produced entity to test us?
Recording with two mikes spaced 6 inches apart seems good, but that produces the correct soundfield only when a septum is in place, preventing the ears from hearing the signal designed for the other. The crosstalk is confounding..and, the crosstalk is also frequency dependent. (check Griesinger).
Producing the correct soundfield stimulus is very difficult. Use of sources where localization parameters are not considered introduces another type of error.
Consider target practice using a pair of glasses where one lens is tilted 5 degrees, skewing the input to one eye vertically. Initially, nothing but misses are to be expected. But eventually, one can adapt to this with time. The same applies to our hearing, we adjust, conforming our interpretation to our experience.
My preface to a DBT test writeup would be: ""This test protocol uses recorded material which presents localization information that does not exist in nature..as such, the results of this test can only be used to determine if a difference can be detected in a "system" using artificial stimuli, to which human adjustment capability has not been determined..""
Cheers, John
Please recommend what source material should be used in a double-blind listening test.jneutron said:
keladrin said:
I was trying to figure out what jneutron was trying to say as well. I still don't see what this would have to do with people identifying the difference between the audible effects (if any) of different cables.
Either you can hear it with music program material (wide variety) or you can't. End of test.
Dumbass said:
For a single tone, a sinx/x signal might be ok.
The problem comes with trying to simulate a virtual source using two speakers when the stimulus is multifrequency (like everything on the planet, of course)
Work and research are required to truly simulate a virtual multifreq source using two sources, as the diffraction around the head causes the ITD to go all to heck..
It may be sufficient to simply apply the mathematically correct IID and ITD to the signals to the speakers, but I am concerned about crosstalk affecting the outcome.
Cheers, John
macgyver10 said:
Those who claim audibility seem to have adjusted to the existing system setup long before swapping cables.
Pulling them into a situation where they need to readjust prior to any testing whatsoever, may be confounding the test protocol.
I do not believe the test protocol is adequate.
Cheers, John
Jneutron,
While I understand what you are trying to say WRT localization not being properly reproduced in the current paradigm of stereo recording/reproduction. It simply isn't relative to the test.
If, in the future, a recording/playback method comes about that meets the criteria that you propose, then the speaker cable test can be repeated to your liking.
Currently, we are testing the claim that people have made regarding the audible effects between different speaker cables, using CURRENT technology and source material.
It's this specific claim that's being challenged. Not your theorectical one.
While I understand what you are trying to say WRT localization not being properly reproduced in the current paradigm of stereo recording/reproduction. It simply isn't relative to the test.
If, in the future, a recording/playback method comes about that meets the criteria that you propose, then the speaker cable test can be repeated to your liking.
Currently, we are testing the claim that people have made regarding the audible effects between different speaker cables, using CURRENT technology and source material.
It's this specific claim that's being challenged. Not your theorectical one.
jneutron said:
This can be accomodated. How much time would be required to "adjust"? In my experience, people who claim differences makes those claims almost immediately. Often this is NOT done in their home sweet home environment, but amongst peers in a suitable meeting place.
Auditory "memory" is notoriously poor. So is visual. A properly constructed test protocol would make instantaneous switches between DUTs, without alerting the subject with any cues.
For us to become so concerned with "adjustment of the subject", seems to indicate that these "differences", if there are any, are insignificant at best?
macgyver10 said:
My so called "theoretical one" takes into account, our ability to adjust to soundfield parametric changes, as well as our ability to create a virtual image using horrendously incorrect stimulus.
You are speaking about measuring an entity (human perceptual abilities) using inadequate tools, with an inadequate and incomplete understanding of our ability to adjust to aural stimulus, to produce results which you are predisposed to.
Scientific endeavors do not follow that method.
You should be asking "how do we hear", rather than saying we can't hear that..as, "that" is a signal that does not exist in nature.
You are speaking of completely uncontrolled testing.
Cheers, John
jneutron said:
I respectfully disagree. If I was testing for engineering the system that you describe, then I would be in full agreement.
However, we are testing the claim that many people have made, using the very same "inadequate tools". They maintain that they can hear it, so I'm suggesting that they be put to the test.
I'm not presupposing anything. In fact, although I'm skeptical, I would be interested to see the results of testing those claims. Maybe there IS something to it?
The test is controlled appropriately for what is being tested. We are testing a pre-existing claim, with tools that were used by the claimant. My hypothesis is that if visual and psychological cues are removed from the equation, they won't be able to detect what they earlier claimed they could. It's that simple.
I do not know. And, I have not found researchers who know the answer, as I was the first to ask them that embarrasing question.macgyver10 said:
macgyver10 said:
Yah, I know..claims like my wife heard it in the driveway. I do not pay attention to silly claims....but not all are silly.
macgyver10 said:
NO....No....no
You have lost it right there.
Do not compare a cableset to another cableset.
Compare a cableset to an actual source in an actual location.
Three speakers, one placed where the virtual image is supposed to be. The other two placed for stereo reproduction.
Then compare the fabricated virtual image to the real one.
It is a far more powerful method, finding the relative position between the virtual and the real. It can be instantaneous..
Cheers, John
You are not talking about ABX testing. You are off in aery-faery land with idealistic ideas about sound reproduction.jneutron said:
You also seem to have a gross misunderstanding of what constitutes a controlled experiment.
The ABX test is designed to test whether or not switching a specific component in a sound system is humanly audible. Nothing more, nothing less.
P.S. to jneutron before going to lunch:
I am not saying that your ideas are without merit, just that they have nothing to do with ABX testing, or with determining whether speaker cables "make any difference".
I am not saying that your ideas are without merit, just that they have nothing to do with ABX testing, or with determining whether speaker cables "make any difference".
Dumbass said:
I took no offense. Both you and Macg have been very nice to discuss with.
The point I am making to both of you, is that the stimulus being used is not one that occurs naturally. So, adaptation is required by humans.
That adaptation is not being accounted for, nor controlled.
If the entity being tested has a settling time, and we do not know it, then how do we interpret the results?
We could fix the timeframe, and assume the settling time is constant. But that is not a consideration in any audibility testing.
I prefer using a natural stimulus, one we have adapted to during our lifetime.
Humans have a demonstrated ability to discern time delays at the 1.5 uSec level, and I won't even start with differential IID capability, nor both together for depth perception..
What, in any test protocol published to date, considers those capabilities?
I only caution testers, you are using an un-natural stimuli on an uncharacterized interpretation system.
Enjoy your lunch.
til tomorrow..
Cheers, John
macgyver10 said:
The technology exists. It actually existed before multi-track, so recorded material is available.
I believe what jneutron is saying, and i agree with him, is that to do a proper test, the source material needs to be choosen so that it is not confounding the results.
dave
Dumbass said:
And as designed is statistically unable to do so. To also be scientifically valid the results also need to be duplicatable... and if you think about the methodologies used, that in itself would be a real challenge.
Lets forget about ABX. It is a flawed test with no redemtion.
dave
Dumbass, what's the test for determining whether ABX testing captures all audio distinctions people actually hear? Just wondering.
- Status
- Not open for further replies.