[linux-audio-user] anti-compressing?

[Mark Knecht]

Andrew Gaydenko wrote:
> Hi all,
> 
> Few days ago there was discussion in the list about mastering-CD/compressing/
> limiting/gain-upping and so on. I have strange feeling from the discussion.
> Does it mean, that, indead, we (all CD-listeners) listen DAMAGED with such
> compressing music? I don't mean any music/sound engineering work! I'm saying
> about market demands (compressed music is more sutable for cheap audio quipment,
> isn't it? And compressed sound seems to be better at low volume level, don't it?). 

Well, nearly all music is compressed a bit, but I often do my own with 
no compression. It's really a question of how loud I want to listen and 
how much headroom my stereo has. I don't listen much about 80-85db 
average volume, and I have LOTS of headroom in my system electrically, 
so for me it works well.

Unfortunately, for analog radio and TV, where they have much lower 
signal to noise ratios, they have to compress. It's an unfortunate, but 
common, problem that Ron was running into. So many producer types think 
you have to get a CD 'radio ready'. In truth most radio stations have 
big, fat compressors anyway, so it wouldn't matter much what CD they 
played. They will make sure they squash it down to whatever they need.

'Dark Side of the Moon' wasn't compressed much. It managed to get 
played. But of course, they weren't late-teen, belly-bareing, 
dance-music titans... (Hey, I like th ebare bellys though!) ;-)


> 
> If it is true (i.e., a market demands to damage music) - are there any (software)
> tools to anti-compress a dynamic range? And - is it worth even to try to expand
> dynamic range?

Basically, no, there's no way to get back to the original. It's not even 
about the compressor equations. Once data is compressed the dynamics 
that tell the compressor what to do are gone from the compressed data. 
To undo the compression you'd have to know something about the original 
dynamics. The data's not there, so you just cannot do it. Compression is 
not 'loss-less'.

I.e. - what a compressor does to a 0.8v signal with slow dynamics is 
different than what it does to a signal that peaks at 0.8V with fast 
dynamics. since you don't know which one it was just by looking at the 
final bit stream you cannot get back to the original.

> 
> P.S. I'd like to repeat, I don't say about engineers work - they try to do
>      their work as good as possible, I think. I'm saying about a market
>      influence to music quality.

If you're interested in about how engineers work and want more ideas, I 
cannot even begin to express how good Bob Katz's book on mastering audio 
is. Everyone interested in this subject should at least check it out 
form a library if not buy a copy. Probably one of the best $40 you'll 
ever spend on an audio book.

> 
> Andrew
> 
>