diff options
author | Martin Schrodt <martin@schrodt.org> | 2019-03-15 09:46:52 +0100 |
---|---|---|
committer | Gerd Hoffmann <kraxel@redhat.com> | 2019-03-18 12:21:15 +0100 |
commit | f6142777659f2e7ad143f2850f1f036f899f475f (patch) | |
tree | cded3595db62bb7213492afb1e7e50fd82e8f184 /qapi | |
parent | baea032ec751306805214190b1ac23f409e9739a (diff) | |
download | qemu-f6142777659f2e7ad143f2850f1f036f899f475f.zip qemu-f6142777659f2e7ad143f2850f1f036f899f475f.tar.gz qemu-f6142777659f2e7ad143f2850f1f036f899f475f.tar.bz2 |
audio/paaudio: prolong and make latency configurable
The latency of a connection to the PulseAudio server is determined by
the tlength parameter. This was hardcoded to 10ms, which is a bit too
tight on my machine, causing audio on host and guest to malfunction.
A setting of 15ms works fine here. To allow tweaking, I also made the
setting configurable via the new -audiodev config. This allows to squeeze out better timings in scenarios where the emulation allows it.
I also removed setting of the minreq parameter to (seemingly arbitrary) half the latency, since it showed worse audio quality during my tests. Allowing PulseAudio to request smaller chunks helped.
Signed-off-by: Martin Schrodt <martin@schrodt.org>
Message-id: 20190315084653.120020-3-martin@schrodt.org
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Diffstat (limited to 'qapi')
-rw-r--r-- | qapi/audio.json | 6 |
1 files changed, 5 insertions, 1 deletions
diff --git a/qapi/audio.json b/qapi/audio.json index 97aee37..9fefdf5 100644 --- a/qapi/audio.json +++ b/qapi/audio.json @@ -206,12 +206,16 @@ # # @name: name of the sink/source to use # +# @latency: latency you want PulseAudio to achieve in microseconds +# (default 15000) +# # Since: 4.0 ## { 'struct': 'AudiodevPaPerDirectionOptions', 'base': 'AudiodevPerDirectionOptions', 'data': { - '*name': 'str' } } + '*name': 'str', + '*latency': 'uint32' } } ## # @AudiodevPaOptions: |