Re: [ecasound] buffer overruns

From: Matthew Becker <matt_becker@email-addr-hidden>
Date: Thu Jul 20 2006 - 20:39:53 EEST

Hi Kai,

thanks for your response! I did try the two methods below,

the recording to /tmp continued to produce the xruns, ( as in

 ecasound -i:/dev/dsp -o /tmp/test.wav )

and the recording with non-double-buffering didn't produce the xruns but did crash the device ( ecasound -i /dev/dsp -o test.wav -z:nodb -z:intbuf )

somehow i thought the sd card would be fast enough, but upon doing some research (which i should have done on the front end :-) ) CF's generally have the edge on data transfer (especially writes?) over SD's i guess. I just ordered a 2 gb CF, hopefully the fact that i'll be recording to a CF (that is also a different device from the rootfs, which will remain on the SD card) will make the difference. Do you by any chance remember what you were recording with when you got ecasound going on your ipaq 37xx? I've started to mess with trying to get the low-latency patch going on the current familiar (0.84-rc2 ) kernel - 2.6.16. It seems to apply ok but fails a compile, once i get a bit of free time i'd like to get it working if possible. Do you remember if the latency was horrible on your ipaq as far as monitoring what you were recording and playback at the same time? (i.e. full-duplex recording -- i'd like to be able to do recording one track and listening to that one and maybe a
couple of others -- very simple track-at-a-time recording). If i can get this working maybe i'll even try to get a simple drum machine on there as well, maybe even something like a stripped-down hydrogen, so i could do simple 3 or 4 track recordings with a drum machine on there. That ( and desoldering the ipaq mic and replacing it with a line-in jack) would help me replace my zoom ps-04 :-) with a cd-quality linux handheld recording solution. (mixing and mastering on the desktop pc later).



>> Could you try:1) Make a quick test recording to a ram-based filesystem (/tmp or some tmpfs partition)...? The recording cannot be too long anyways, but you can check whether it's the SD writing, or something else.2) Record with: sh> ecasound -i /dev/dsp -o test.wav -z:nodb -z:intbuf I.e. skip the double-buffering when writing to the SD card.--

Join Excite! -
The most personalized portal on the Web!

Take Surveys. Earn Cash. Influence the Future of IT
Join's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
Ecasound-list mailing list
Received on Fri Jul 21 00:15:02 2006

This archive was generated by hypermail 2.1.8 : Fri Jul 21 2006 - 00:15:02 EEST