The new speakers I am co-designing with diyaudio.com members will reproduce source material so accurately that I fear that many of my favorite recordings will sound badly overcompressed, given the dynamic range capability of today’s digital audio technology. Many of my favorite soundtracks and pop and jazz music were originally mastered for release on vinyl. So the first or second gen tape masters used to produce those old vinyl releases were hit with lots of compression to ensure trackability by cheap vinyl players and to extend playing time on vinyl singles and LPs.
It remains to be heard how much the masters of recordings of this vintage had escaped this sad fate. https://www.hdtracks.com/
However, even if they sound reasonably undistorted and with respectable frequency response, many of my favorites that I have on CD-and most released by the original label-definitely had their dynamics squashed out of them by high compression rates. But how soon before there would be AI software that could remap every audible aspect of a recording and via powerful extrapolating algorithms restore the original dynamics?
What I was proposing is software with AI driven capabilities-either self contained and/or enhanced by late model processors-that could do this remapping on good quality source material, like CDs. True, the process would likely sound far better if the original analog-or professionally digitized-tape masters were so worked upon. But many of my CDs sound quite good but for being dynamically squashed from the compression.
So why couldn’t a software solution-one selling for about the cost of an audio editor platform like https://www.steinberg.net/wavelab/ but marketed for audiophile use make it happen?
Wouldn’t that be lovely? And no recording, mastering or software engineers need have their jobs threatened by AI in the process.