diff --git a/src/xpra/client/ui_client_base.py b/src/xpra/client/ui_client_base.py
index 73e3684fb3..bad4425733 100644
--- a/src/xpra/client/ui_client_base.py
+++ b/src/xpra/client/ui_client_base.py
@@ -1994,7 +1994,7 @@ def stop_receiving_sound(self, tell_server=True):
         ss = self.sound_sink
         self.speaker_enabled = False
         if tell_server:
-            self.send("sound-control", "stop")
+            self.send("sound-control", "stop", self.min_sound_sequence)
         self.min_sound_sequence += 1
         self.send("sound-control", "new-sequence", self.min_sound_sequence)
         if ss is None:
diff --git a/src/xpra/server/source.py b/src/xpra/server/source.py
index dc1669682d..a7b04d2fae 100644
--- a/src/xpra/server/source.py
+++ b/src/xpra/server/source.py
@@ -988,6 +988,12 @@ def stop_receiving_sound(self):
     def sound_control(self, action, *args):
         soundlog("sound_control(%s, %s)", action, args)
         if action=="stop":
+            if len(args)>0:
+                sequence = int(args[0])
+                if sequence!=self.sound_source_sequence:
+                    log.warn("sound sequence mismatch: %i vs %i", sequence, self.sound_source_sequence)
+                    return "not stopped"
+                soundlog("stop: sequence number matches")
             self.stop_sending_sound()
             return "stopped"
         elif action in ("start", "fadein"):