If minreq is not explicitly specified, it was always initialized to 20 ms (DEFAULT_PROCESS_MSEC). However when the total latency is not much higher than 20 ms, this is way too high. Instead use tlength/4 as a measure: this will give a decent sink_usec in all modes (both traditional, adjust latency and early request modes). This greatly improves PulseAudio's ability to ask for data in time in low-latency scenarios. Signed-off-by: David Henningsson <david.henningsson at canonical.com> --- src/pulsecore/protocol-native.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/src/pulsecore/protocol-native.c b/src/pulsecore/protocol-native.c index 38c66c5..da705eb 100644 --- a/src/pulsecore/protocol-native.c +++ b/src/pulsecore/protocol-native.c @@ -938,8 +938,13 @@ static void fix_playback_buffer_attr(playback_stream *s) { if (s->buffer_attr.tlength <= 0) s->buffer_attr.tlength = (uint32_t) frame_size; - if (s->buffer_attr.minreq == (uint32_t) -1) - s->buffer_attr.minreq = (uint32_t) pa_usec_to_bytes_round_up(DEFAULT_PROCESS_MSEC*PA_USEC_PER_MSEC, &s->sink_input->sample_spec); + if (s->buffer_attr.minreq == (uint32_t) -1) { + uint32_t process = (uint32_t) pa_usec_to_bytes_round_up(DEFAULT_PROCESS_MSEC*PA_USEC_PER_MSEC, &s->sink_input->sample_spec); + uint32_t period = s->buffer_attr.tlength / 4; /* In low-latency, ask for data in four periods by default */ + if (frame_size) + period -= period % frame_size; + s->buffer_attr.minreq = PA_MIN(period, process); + } if (s->buffer_attr.minreq <= 0) s->buffer_attr.minreq = (uint32_t) frame_size; -- 1.7.9.5