PlanetZ forums - scopeusers.com

If you use 24bit recording, I see no good enough (to me anyway) reason to software compress before recording (with most singers anyway).

Software compressing will not give you a better s/n ratio. It will however give you more bits (bits like in 16bit, 24bit etc.) on your harddisk on the quiet parts. This advantage however is only usefull, if your a/d converters work on a higher bit rate, than your recording.

Immanuel

Here´s a scenario that can bring out some relative dramatic difference... Lets say you have to record flamengo guitar. It is a very transient source to deal whith and you record that instrument at -20dB to not allow the peaks to clip. Thats 3-1/3 bits lesser accuracy than 24-bit (6db = 1bit). The quiet passages are at -32dBs. That´s a 5-1/3 bits lesser accuracy than 24-bit (144db). Then you'll try to software compress with 12dB gain reduction to reduce those peaks. By that you´ll also add makeup gain of another 12dB. The S/N now is 7-1/3 bits lesser than your maximum sample depth of 24-bit. And afterwards you decide to apply some radical eq by boosting certain frequencies, which again raises the S/N. And the more tracks you recorded that way, the worse... And all those converters just give you a real bit-depth of 20-bit.... So you finally ended up whith a 12-bit file, no matter if you recorded everything whith 32-bit float or 24-bit fixed.

But if you decide to compress a little before AD conversion, the whole conversion process might give you a few extra bits for further processing. But the compressor should really be a good one! And you´ve got to know what you are doing...

Regards,
Bernhard