Wow, you seem pretty worked up about this. Let me explain.
Quote:
Originally Posted by JBroll
Wrongo. An ideal compressor ONLY ALTERS VOLUME, not timbre.
|
That's exactly what I said, it even says so in the very quote you posted. It is possible to alter timbre with multiband compression, but normally, you'd keep all the bands pretty even (hence: when done correctly).
Of course a normal singer "sounds" better when compressed, if by sound better you mean "signal is more even, doesn't change between inaudible and drowning everything else out". But I meant the actual singing performance itself: Being on pitch, conveying emotion, all that crap isn't affected by the use of a compressor. When you are screaming though, it is. I should have been more clear on that, sorry.
Quote:
Originally Posted by JBroll
But just think about the implications of 'compression brings up quiet little parts of the signal all the time' - the signal-to-noise ratio of EVERYTHING would be 0dB.
|
Compressors have a threshold value. Everything below it does not get altered. Above the threshold however, the signal gets compressed by whatever ratio you have set. So above the threshold, the signal-to-noise ratio DOES suffer.
I am not exactly sure how you came up with 0dB, but it implies that you think a compressor is an "all or nothing" effect. I could interpret that as some kind of "fundamental failure to understand how exactly a compressor works", but I'm sure you just meant something else and I wouldn't want to risk coming across as an arrogant know-it-all.
One example is: Set the threshold to -20dB, and the ratio to 1:infinite, you have a brickwall limiter. If your attack and release times are fast enough, most of the signal that is above that threshold now has an SNR of 20dB less than it did before. Everything below the threshold has the same distance to the noise floor as before, so its SNR doesn't change.
Now after such stupidly heavy limiting you'd compensate with gain, raising the noise floor by whatever amount of gain you apply. So overall, the SNR tends to suffer, that's why most vocalists who compress also use a noise gate or an expander.
Another example:
Imagine screaming...I don't know, RAAAAAAKE into the microphone, because you're really upset with gardening tools.
You'd set your threshold so that the R is below it and the A is above and gets compressed. The K get's compressed too, obviously, but the barely audible "-h" at the end does not. What happens? The R gets a lot louder relative to the A or the K.
Try screaming the R at volume, you'll see what I mean. Those are the kinds of noises that the compressor can amplify, and as a result the whole scream sounds alot more powerful and effortless.