target_rtp_time had better be even for stereo

diff --git a/server/speaker.c b/server/speaker.c
index 4bb1d2c6c4b6630b34fbcad53082dc0102e11c39..aebd930e490413b496d7ecf899a070c69041ba07 100644 (file)
--- a/server/speaker.c
+++ b/server/speaker.c
@@ -974,6 +974,7 @@ static size_t network_play(size_t frames) {
      * transmitting packets with overlapping timestamp ranges, which does not
      * make sense.
      */
+    target_rtp_time &= ~(uint64_t)1;    /* stereo! */
     if(target_rtp_time > rtp_time) {
       /* More time has elapsed than we've transmitted samples.  That implies
        * we've been 'sending' silence.  */