On Fri, 17 Dec 2004, Jan Weil wrote: Actually, as much as I'd like to recommend this combination, I'll have to instead recommend using libsamplerate [1], and the sndfile-resample tool that comes with the package, directly: sndfile-resample -to 48000 foo44100.wav foo48000.aiff sndfile-resample -to 44100 foo48000.wav foo44100.wav Works great and with good quality. I've received some bug reports about resampling with the ecasound+libsamplerate combo, and the 44.1<->48 conversions seems to trigger this bug. I hope to solve this issue for the next Ecasound release (2.3.6/2.4.0). But for pure conversion+resampling, sndfile-resample is just an excellent tool and has no problems with 24bit audio. [1] http://www.mega-nerd.com/SRC/ -- http://www.eca.cx Audio software for Linux!

On Fri, 17 Dec 2004 19:32:58 -0500 Jesse Chappell &lt;jesse(a)essej.net&gt; wrote: The simple API was designed to convert a whole file at once. Please, please, please do not use the simple API on multiple blocks of one contiguous piece of the same audio, unless you *know* there is silence around the join point. Yes, the callback and process API both store state between calls. Good catch Jesse. Cheers, Erik -- +-----------------------------------------------------------+ Erik de Castro Lopo nospam(a)mega-nerd.com (Yes it's valid) +-----------------------------------------------------------+ "Every time microshaft's stock price drops again, I rejoice. I want to see that bunch of criminals brought to their knees. Preferably at the chopping block." -- rixt in http://linuxtoday.com/stories/20659_flat.html

On Tue, Dec 21, 2004 at 04:12:48AM +0200, Kai Vehmanen wrote: Kai, Will this also affect the -ei pitch shifter effect? -Eric Rz.

On Tue, Dec 21, 2004 at 08:27:17PM +0200, Kai Vehmanen wrote: Great. Thanks Kai. I had some sort of issue with -ei being limitted to less than the full 256 times shift supported by libsamplerate. I found the limit somewhere in ecasound code (or was this inside libsamplerate itself? <scratches_head> ) and changed it in my copy. Sorry to be so vague. There was a comment somewhere in there about the shift being limitted due to glitches in the output when extreme values were used. Raising the limit accomplished my goals, but I do recall getting some lack of smoothness in the results. I haven't updated any audio stuff on my audiobox in several months and I'm at work at the moment. So, I'm not exactly sure what versions of everything i'm currently using. I've been preoccupied with building a studio space in my basement for the past month. Once I get the room in shape I'll update to the latest versions of everything and let you know if I still have any issues. On another ecasound note, but unrelated to the Subject: of this thread: I've been recording the studio construcion sounds with a python-ecasound script that randomly mixes the input of 4 microphones down to mono. I've recorded so much that I've almost maxed out my audio drive. So, the upshot of this is that I'll have to convert my pool of audio and also all of my ecasound python eci scripts to flac. According to some rough calculations doing this will let me put off buying larger drives for another 6-8 months. I know you added flac support earlier this year and I did some preliminary tests at that time. But, I ran into some issues and didn't have time to work them out so I stuck with wav files. Have you had many reports of success or problems with ecasound's flac support? Do you use it yourself? Anyway, I will necessarily need to have solid flac support in the next few months. So, that stuff will be receiving some testing on my system soon. -Eric Rz.