RNNoiseFilter: don't filter silence frames

Fixes #3259

changelog/3309.fixed.md

@@ -0,0 +1 @@
+- Fixed a `RNNoiseFilter` issue that would cause a "[Errno 12] Cannot allocate memory" error when processing silence audio frames.

src/pipecat/audio/filters/rnnoise_filter.py

@@ -115,6 +115,9 @@ class RNNoiseFilter(BaseAudioFilter):
         if self._sample_rate != 48000 and self._resampler_in:
             in_audio = await self._resampler_in.resample(audio, self._sample_rate, 48000)
 
+        if len(in_audio) == 0:
+            return audio
+
         # Convert bytes to numpy array (int16)
         audio_samples = np.frombuffer(in_audio, dtype=np.int16)