I just dont understand why this is any different than audioquest cables though. Both companies make products that dont do anything appreciable at best, and are poor performers at worst.
While I don't care too much about the specific opinions of individual members on the matter, no matter what perceived status they might have. I am still curious to know what
@amirm's reaction is to this?
What I would like even more is to have
@amirm actual knowledge of losses encodings chip in here for a more in-depth analysis on what things might actually be possible in a compression scheme like MQA provides. I'd love to know for instance what bitrate would be needed to encode let's say a 20 to 40 kHz music signal with half-decent quality. What kinds of tradeoffs can be made here? What about effects like time smearing in lossy codecs (remember those first 128k MP3's).
And at
@GoldenOne: Would be interesting to do another analysis on your tracks from Spotify. What does the lossy compression do with those test signals? What is the difference? Does the encoder also go haywire?
Also, note that you could actually encode a 320 kbit Vorbis file, noise shape the output (basically just whitening the data), and then shove them into the lower 5 to 6 bits of a 24-bit track. It would fit, and would still contain 18 bits of "lossless" audio, 5 to 6 bits of additional noise, and a fully Vorbis encoded version of the same track. Obviously, these two data sources are then totally uncorrelated, so realistically there is no good use-case for this, but it does indicate how much room there actually is to encode additional lossy information in those lower bits.