Alsa/Pulseaudio conflicts

[Gilles Casse]

Hynek Hanke wrote:
> The current implementation of Pulse Audio in Speech Dispatcher
> is slow because the implementation of it contains a problem.
> For each sound output, it opens and closes the whole connection,
> so there is a big overhead which is completely unnecessary.

Hello Hynek, all,

 From pulseaudio 0.9.8, pa_stream_update_sample_rate has been added for 
updating the rate. But if the format of the sound icon is different from 
the speech samples, I guess that the stream must be still closed and 
reopened.

I have updated the code consequently but unfortunately can't go further 
  at the moment. A patch is copied below as is (but code is not yet 
operational).

Two remaining issues:
- on a Hardy based environment (beta release), the sound icons remain 
mute (tested using sd from cvs (or from package Hardy) + espeak 1.36 (or 
1.32), the sound-icon package has been installed, but no sound using 
run_test or speechd-el).

- If I add the PA_STREAM_VARIABLE_RATE flag to 
pa_stream_connect_playback (for adding the "update rate" capability), 
spd-say does not play any sound. Perhaps a conflict with the other flag.

Copied below is the current patch.

Best regards,
Gilles



--- pulse.c.old 2008-04-20 21:51:21.000000000 +0200
+++ pulse.c     2008-04-21 18:41:55.000000000 +0200
@@ -723,7 +723,8 @@


    SHOW_TIME("pa_connect_playback");
-  if (pa_stream_connect_playback(id->pulse_stream, NULL, &a_attr, 
(pa_stream_flags_t)(PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_AUTO_TIMING_UPDATE), 
&id->pulse_volume, NULL) < 0) {
+  //  if (pa_stream_connect_playback(id->pulse_stream, NULL, &a_attr, 
(pa_stream_flags_t)(PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_AUTO_TIMING_UPDATE), 
&id->pulse_volume, NULL) < 0) {
+  if (pa_stream_connect_playback(id->pulse_stream, NULL, &a_attr, 
(pa_stream_flags_t)(PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_AUTO_TIMING_UPDATE|PA_STREAM_VARIABLE_RATE),
 
&id->pulse_volume, NULL) < 0) {
      ERR("Failed to connect stream: %s", 
pa_strerror(pa_context_errno(id->pulse_context)));
      goto unlock_and_fail;
    }
@@ -1268,6 +1052,7 @@
    int ret = PULSE_OK;
    size_t a_total_free_mem;
    pa_sample_spec ss;
+  pa_operation *o = NULL;

    if (id == NULL){
      ERR("Invalid device passed to %s()\n",__FUNCTION__);
@@ -1303,9 +1088,31 @@
      /*  if the sample spec changes, the stream will be rebuilt. */
      const pa_sample_spec *ss2 = pa_stream_get_sample_spec 
(id->pulse_stream);

-    if (!pa_sample_spec_equal (&ss, ss2)) {
+    //    if (!pa_sample_spec_equal (&ss, ss2)) {
+    if ((ss.format != ss2->format) || (ss.channels != ss2->channels)) {
        _pulse_close(id);
      }
+    else if (ss.rate != ss2->rate) {
+      // just update rate (no need to reopen the stream).
+      // require pulseaudio >= 0.9.8
+      id->pulse_success = 0;
+
+      if (!(o=pa_stream_update_sample_rate(id->pulse_stream, 
track.sample_rate, _stream_success_cb, id))){
+       ERR("pa_stream_update_sample_rate error (%s)\n",__FUNCTION__);
+       ret = PULSE_ERROR;
+       goto terminate;
+      }
+      SHOW("pulse_play: %s (call)\n", "pa_stream_update_sample_rate");
+      while (pa_operation_get_state(o) != PA_OPERATION_DONE) {
+       //      CHECK_DEAD_GOTO(id, fail, 1);
+       pa_threaded_mainloop_wait(id->pulse_mainloop);
+      }
+      SHOW("pulse_play: %s (ret)\n", "pa_threaded_main_loop");
+
+    if (id->pulse_success == 0)
+      ERR("pa_stream_trigger() failed: %s", 
pa_strerror(pa_context_errno(id->pulse_context)));
+
+    }
    }

    if (!id->pulse_mainloop) {