backport the sync implementation from the csdr++ branch

csdr/__init__.py

@@ -258,7 +258,7 @@ class Dsp(DirewolfConfigSubscriber):
                 chain += ["csdr convert_f_s16"]
 
         if self.audio_compression == "adpcm":
-            chain += ["csdr encode_ima_adpcm_i16_u8"]
+            chain += ["csdr++ adpcm -e --sync"]
         return chain
 
     def secondary_chain(self, which):

csdr/chain/clientaudio.py

@@ -10,5 +10,5 @@ class ClientAudioChain(Chain):
             workers += [AudioResampler(inputRate, clientRate)]
         workers += [Convert(Format.FLOAT, Format.SHORT)]
         if compression == "adpcm":
-            workers += [AdpcmEncoder()]
+            workers += [AdpcmEncoder(sync=True)]
         super().__init__(*workers)