The format is still beta'ish; one would expect the tagging-thing to be dealt with at some point in time.
This wouldn't solve the fundamental problem (if skwire understood the specs properly), though - that a lot of existing programs would possibly have to be updated in order to not crash / eat up memory like crazy when dealing with mp3hd files.
So my guess is that MP3HD will excel with Classical music, as an example, but not with Rock.
All variable-bitrate compression formats depend on the input in one way or another - either by having a constant output filesize (and trying to "spend less bits on less active passages") or achieving different output filesize depending on the input.
Other than that, the example is plain stupid; any classical concert contains tons more dynamic tunes than any rock music concert ever, and should therefore also take up much more bits, I would imagine.
Don't know about that, but I expect it to depend very much on the compression algorithm. As I understand it, MP3 works by doing frequency analysis, and discarding frequencies we don't pay as much attention to, in order to achieve better bitrate for the more interesting frequencies... lossless codecs obviously cannot do this, so they work differently
- I would expect classical music to achieve relatively small filesizes because there's silent passages and slow progressions, whereas rock, industrial, etc is often full-volume-all-the-time and has a lot of "harsh" sounds (shredding guitars, noise, whatever) that I would guess
results in a larger bitrate requirement.
But I'm pretty much a layman when it comes to audio and audio compression, so I could be totally wrong