Quote:
Ok - briefly... doubting non-existent evidence is not a flaw on the part of the doubters. Subjective impressions, without objective measurements do not count as evidence. They can support actual evidence, but our brains are too good at adapting and there are so many psychological (not to mention physical - pad break in, for instance) other factors, not to mention our really, measurably BAD audio memory, that impressions, alone - mean very little.
This is not because we are perpetual doubters. It's because our standards of evidence mean something. Show me actual repeatable, testable evidence and I'll spin on a dime and support this. My position is the null hypothesis. Until is it show to be, it cannot be claimed as fact to be.
No. The burden of proof is on someone making a positive claim of something. Not on me to show why your claim is bunk.
Show me these frequency graphs showing audible changes - then we'll talk. Even Tyl's graphs (the only ones I've seen in support) showed changes on the very small fractions of decibles, and did not account for sample variation and any number of other testing errors. A fact he acknowledges.
Outside of Tyl's graphs (I'll explain more on this in a bit), we have shown you 2-3 other graphs that I can remember on Head-Fi. Will I find them again, they are long gone, you can look if you feel. They were just thrown out the window and
doubted for accuracy. Everything to you guys seems innacurate. I know other factors come into play, sure they do. However, the apparent change cannot be accounted for human error after these observations have been done, and confirmed time and time again. Everyone (that hears the burn in) hears these changes, and not only do they hear these changes, the hear them over the
same time frame. This second part in itself debunks that it's psychological... If it was, we would actually hear this change in a
different time frame. As for bad memory, you have to remember, one persons memory may not be trustable, but when 50+ people report the same results for a given IEM (reduces your padding claim to null here) those results must be accepted, other wise it looks foolish to doubt those results (what you just did with the numerous observations).
A lack of proof against the argument that doesn't rely on psychological (since this was debunked before you above) or doubting accuracy shows invalidity in that stance. You keep giving the same reasons over and over. We give you data, and observations, you doubt it as being accurate. Now, who does this look bad? Let me give you a parallel example:
One doubter stands, not trusting his senses (doesn't believe in them, or just doesn't have them). One person stands there and tells him there is a cat on the mat. These are two physical objects, a cat that is sitting on a mat. The doubter doesn't believe it, he doubts himself, and the other person (let alone, he himself might not see the cat). So that one person brings another to come and confirm that that cat is indeed on the mat. Once more, the doubter doubts the accuracy and validity of both people. So the man goes out once more and brings back 20 people to tell the skeptic there is a cat on the mat. Once more, the skeptic does what he does best, doubt with the same reason (you're senses are invalid and inaccurate; not to mention, it could be psychological, you can be hallucinating). He goes out to bring 200 people... You get the point, the doubter keeps calling the following into question (accuracy, validity, psychological).
Now, that example has nothing to do with audio until you change cat on the mat to burn in. Then it makes sense. A skeptic is just that, doubting all accuracy and observation regardless of how many tell them about it. They deny all claims due to the reason of inaccuracy. Is this not what you're doing? When presented with observations, and graphs, you deny that they are true, or find some excuse to
make them invalid? Aren't you being a full on 100% skeptic here? Believe it or not, that story above parallels perfectly what is happening here. You state this, "...not to mention our really, measurably BAD audio memory, that impressions, alone - mean very little." When you do this, do you not doubt validity and accuracy of what countless people have found (please note that IEMs do not have pads that break in and the psychological factors are broken above).
Now on to Tyl's graphs... You're right, for one frequency range, there is very little change, at most, I saw a 2dB change. Although audible, was very small change. However, music in itself (why test 1 frequency alone when any given instrument cannot create a pure tone of one frequency) uses multiple frequency ranges for a given
note. So now we don't have one frequency (what you are comparing), but a number of frequencies. How do we measure change in that? An integral, more importantly, the integral of the absolute value of the freq graphs subtracted from each other so the integral would look like this:
S(|f(l)| - |g(l)|)dl
S = integral symbol
l = dB output
f = end function of dB vs frequency
g = original function of dB vs frequency
Range for this integral will be actually any range for a given sound (as this is what we hear a difference in). For some reason we don't listen to pure tones in our everyday music, why base your tests on it?
So 2 dB for a given frequency, that goes up and down will create a
huge overall change in sound. When you take all those little changes and add them up, they turn into a big change which will end up with a substantial enough change in a given note of an instrument. So although Tyl's graph shows small changes, the small changes (looking at the .5 dB spikes and other .2-.4 dB spikes; heck there are even some 2 dB changes here and there) added together create a change in sound that is bigger in itself as we are now looking at change over a frequency range (not one given frequency).
Now secondly, you also have to remember that a spike in some frequency can change the entire change in the perception in the entire frequency. A spike here = a small subtle loss everywhere else. At it's core, this loss is small, however, adding up all the the additions and subtractions, this loss grows. If you add bass and treble, do you not recess the mids? If you add mids do you not recess the bass and treble? Looking at the grand scheme of things,
small change is
big over the entire spectrum.
Just something to think about. We shouldn't be basing graphs on one frequency range, instead, the entire frequency range, and getting area between (requires an integral) them to get accurate numbers of overall change for a given frequency range.