On Sat, 26.07.08 13:01, Sean McNamara (smcnam at gmail.com) wrote: > 0.9.11 (which I'm assuming contains this change) works for me in Yes, 0.9.1 contains 5edbb57516835688b1a47a3e0813b3965b014db2 > system mode, for a very loose definition of "works". I can create and > run streams and they play... eventually. The 2-second dropout at the > start of the stream is still a mystery to me. tsched=0 reduces this > from ~2 seconds to the duration of 1 fragment as declared in > daemon.conf. But it's still a pop at the start of each stream. Hmm, what client is sthis? > With system mode disabled and tsched=0, I get zero dropouts on all but > the most incompatible PA clients. But the need for tsched=0 is > probably no surprise to you, since my card is based on Intel HDA. > > Since many of the HDA problems are bugs in ALSA, could you do a list > announce once you are reasonably confident that patches have been > applied to ALSA which allow us HDA'ers to turn on time-based > scheduling? Filtering the alsa-devel commits for relevant patches > takes hours, but I'm assuming any requisite fixes for this would > probably be initiated by you... :) I think for now that's only the three patches commited to alsa-lib shortrly after 1.0.17. Lennart -- Lennart Poettering Red Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4
