adjust the volume just before sending it out to the sound card and no…

…t before entering the queue where things may get dropped (leaky queue), and delayed (by definition)

git-svn-id: https://xpra.org/svn/Xpra/trunk@16829 3bb7dfac-3a0b-4e04-842a-767bc560f471

src/xpra/sound/sink.py

@@ -111,7 +111,6 @@ def __init__(self, sink_type=None, sink_options={}, codecs=get_decoders(), codec
             pipeline_els.append(decoder_str)
         pipeline_els.append("audioconvert")
         pipeline_els.append("audioresample")
-        pipeline_els.append("volume name=volume volume=0")
         if QUEUE_TIME>0:
             pipeline_els.append(" ".join(["queue",
                                           "name=queue",
@@ -120,6 +119,7 @@ def __init__(self, sink_type=None, sink_options={}, codecs=get_decoders(), codec
                                           "max-size-bytes=0",
                                           "max-size-time=%s" % QUEUE_TIME,
                                           "leaky=%s" % QUEUE_LEAK]))
+        pipeline_els.append("volume name=volume volume=0")
         sink_attributes = SINK_SHARED_DEFAULT_ATTRIBUTES.copy()
         #anything older than this may cause problems (ie: centos 6.x)
         #because the attributes may not exist