I will try to answer:
*****************
How did you come to these conclusions?
e.g., did you do pre & post sound stream analysis? Review code or technical documentation? Read analysis from 3rd party sound technologists?
*****************
First I will point out again I am not a programmer. I have done nothing like what you discuss above. My understanding of this comes from my engineering background and what I know about the digital world. Analog signals are smooth like a sinewave. The AC out of your wall socket for instance. If you listened to that sinewave you would hear a 60 cycle hum. If this 60 cycle signal were digitized then it would sound nearly the same but not as clear because small pieces of the wave form (harmonics) would now be missing. The missing pieces would look like small stair steps in the digitized wave form instead of a smooth steady line.
The size of the missing pieces is mitigated somewhat by the sample rate. The higher the sample rate is, the less impact the missing pieces of the wave form have on the sound you hear, because the stair steps are smaller. This is why people usually try to use a higher sample rate to get the best quality digital recording they can. Something to note in all of this is not all digital sources have been digitized at the same sample rate. MAXD's tech (algorithm) will restore in a transparent way all of the stair steps in the digitized signal no matter what sample rate or source was used.
So, when an analog signal is sampled, one rectangular piece at a time is sampled. Then a second rectangular sample is made and so on. The stair steps I am referring to are because the samples are square. The sign wave has a rising or falling curve. Thus, the sampling leaves out a small chunk (the harmonics of the analog wave form) and you see small stair steps in the digitized signal.
How can the software determine whether the extrapolated data is accurate to the analog source? I am not certain that it actually does that. Replacing the harmonics in my view would require some mathematical understanding of what harmonics are and how they may at least be approximated and restored so that the ear may not tell the difference from the original analog signal even though the restoration may not be exactly what was in the original signal. In fact, the restoration may sound even better than the original and may be controllable (tweaked) to some extent to improve things even more. You would have to ask the programmer how this is being implemented in the algorithm.
Your comments about the subjectivity of hearing tests is spot on. The demo on the MAXD website however shows how this tech may actually allow the difference to be heard by the human ear. It is a small difference but in head phones or telephones it may be very noticeable. As I previously mentioned in my earlier post, this may not be significant enough to be sellable. They have been trying for almost 10 years to get a buyer. Either this is the calm before the end, or the calm before the boom!
I don't know if I covered everything in your response but I tried to clarify some of my thinking. Some of your questions would require someone with appropriate expertise to answer. My education is in Material Science and my work background is in System Planning for electric utilities.
Hope this helps.
