From: Conrad P. <co...@me...> - 2006-12-23 00:16:26
|
On 23/12/06, pete <zen...@ze...> wrote: > On Fri, 2006-12-22 at 12:33 +1100, Erik de Castro Lopo wrote: > > peter wrote: > > > > > svn co http://svn.metadecks.org/sweep/branches/sweep-jack-testing > > > will get you the closest we have to a jackified sweep. there's no jack > > > input (recording) yet and if you playback files that have a different > > > samplerate to your running jackd then it won't sound too good given > > > sweeps rather noisy resampling. > > > > When running with Jack output and the user attempts to open > > a file at a different sample rate, I think it would be a really > > good idea to either fail with an error dialog and/or ask > > the user to sample rate convert it to the jack sample rate > > using libsamplerate. > > well, correct me if i'm wrong but, couldn't we use libsamplerate on the > fly to convert where needed at all output stages? > we could have options to select the converter, optional obstructive > warning dialogs (for the paranoid and forgetful) and some sort of > visual indicator that highlights when source and output rates > don't match? > > it's not just useful for jack either, the alsa plug layer resampler is > limited by (current) design to linear interpolation iirc, so application > side resampling (albeit via libsamplerate) would be a improvement there > too. > > i was also wondering if we could use libsamplerate to handle scrubbing? yeah, libsamplerate would be good for both these uses :-) The current architecture (doing very cheap interpolation) was done around 2001 or so. I reckon we can afford the CPU to do better quality rate conversion now ;-) cheers, Conrad. |