Observations are not anecdotes? Science is founded on observation, we knew thousands of years ago the earth travelled around the sun, we did not have to go there or measure it to know this, it was established by observation. It is impossible to measure what one person hears in comparison to another, the only possible method is by observation.
Oh, please.
The issues are well-documented:
1. Any difference in level, even as small at 0.1 dB, will affect the subjective listening impression when comparing devices.
2. Any knowledge of which unit is being played will affect the subjective listening impression. That does not mean that we don't hear what we hear, it means that our interpretation of it goes through a range of conscious and unconscious filters that are biased by what we think we know of the devices under test.
This isn't a matter of observation versus anecdote, a common refrain from those who don't understand what those words mean to a scientist. We use the term "anecdotal data" to describe careful and controlled observations that nevertheless provide an insufficient sample size to assert a conclusion that can be transferred outside the test case. A doctor notes that a disease condition in Person 1 is improved by taking Med A. This is a careful and controlled observation to be sure that Med A really did improve Person 1's condition, and not Med B that the person was also taking, or the effect of Person 1 just being in a better mood. But it is anecdotal, because it says nothing about the likelihood of Med A improving the same condition in Person 2. A study of sufficient sample size to characterize a population requires careful and controlled observations, but it also requires something more: enough of those observations to demonstrate that the
transferability of the effect to others is not mere chance.
You have not even demonstrated that your comparisons of CD players are careful and controlled--were they level-matched? Were you testing without knowing which device you were listening to (commonly known as blind testing)? If not, your own impressions should not even be reliable in reference to you (which requires care and controls), let alone transferrable to anyone else (which requires a sufficient sample size not to be anecdotal). That others heard the same effect without knowing what was playing isn't persuasive, because level-matching (using a millivolt meter and a test tone, not just "by ear") is the critical first step, and because you asked them to listen to something, which is already an indication that you made a change and desire approval. You are not being careful about this.
I own a range of CD players, including the vaunted (in its day) Magnavox CDB-650, a Tascam 401 (which has no audiophile credentials at all), a Cambridge CXC with various DACs, and a Naim. In the case of the Tascam, I also compared its internal DAC to various external DACs. These CD players range widely in price, new or used. I assert that I can't tell any difference between any of them. Now, what makes your observation more valuable than mine?
A transport that sounds different is introducing flaws in the data stream, but how would that happen without causing audible error artifacts such as dropouts or track-jumping? The transport isn't responsible for jitter--the DAC deals with that issue. Error survival either works or it doesn't. If there are data errors, these should be obvious because they will cause dropouts and tracking artifacts that are not subtle. But I can compare rips with the database accessed by Exact Audio Copy, and in nearly every case, the checksum on each track is exact--the files are essentially bit perfect even read at high speed in a $5 plastic computer drive. So, your observations not only don't demonstrate the care and controls they should for you to consider them reliable in your own case, but they violate what has been demonstrated analytically--that the same stream of bits delivered to a DAC will be the same no matter what the delivery device is.
Rick "whose perceived differences usually melt away with matched levels alone, not even requiring blind testing" Denney