On Tue, 2008-12-23 at 20:09 +1030, Matthew Smith wrote: > I'm not all that sure about how good it is to "normalise" 24 bit to 32 > bit - that would mean some sort of interpolation and thus introduction > of errors. To normalize doesn't refer to changing bit-depth at all. It's about detecting the loudest sample in a file (or selection), calculating a multiplier that will give it the highest allowed value and then applying this multiplier to all the samples. This is of course neutral at best regarding quality, but makes handling material from many different sources a bit easier. So I hope, at least :) > I think I'll just record using the H2, 24 bit @ 48kHz, trim and save. > Then I will expand that to 32 bit so you have got the pure and > interpolated versions. I asked for 32 bit float mainly to not have people recording with a JACK application submitting exports to an integer bit-rate. I want to avoid conversions and 32 bit float is a given because I will edit in Ardour :) -- Thorsten Wilms thorwil's design for free software: http://thorwil.wordpress.com/ _______________________________________________ Linux-audio-user mailing list Linux-audio-user@xxxxxxxxxxxxxxxxxxxx http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user
