Eric, thanks! I'll try to compare sndfile-xxx- and ecasound- reformatted files. Probably, results will be the same :-) Andrew ======= On Monday 10 October 2005 18:31, Eric Dantan Rzewnicki wrote: ======= On Sun, Oct 09, 2005 at 11:03:32PM +0400, Andrew Gaydenko wrote: > Eric, > > Thanks! > > Are there any quality differences in comparison with sndfile-xyz tools? If you use the resample-hq option ecasound uses libsamplerate, which I think is used by sndfile-resample since Erik de Castro Lopo is the author of both libsndfile and libsamplerate. I'm pretty sure ecasound uses libsndfile for most sound file handling, though I'm not 100% sure whether that includes sample width changes. Anyway, the syntax from man ecasound: Resample - access object of different sample rates Object type 'resample' can be used to resample audio object's audio data to match the sampling rate used in the active chainsetup. For example, ecasound -f:16,2,44100 -i resample,22050,foo.wav -o /dev/dsp, will resample file from 22.05kHz to 44.1kHz and write the result to the soundcard device. Child sampling rate can be replaced with keyword 'auto'. In this case ecasound will try to query the child object for its sampling rate. This works with files formats such as .wav which store meta information about the audio file format. To use 'auto' in the previous example, ecasound -f:16,2,44100 -i resample,auto,foo.wav -o /dev/dsp. If ecasound was compiled with support for libsamplerate, you can use 'resample-hq' to use the highest quality resampling algorithm available. To force ecasound to use the internal resampler, 'resampler-lq' (low-quality) can be used.
