wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: rename audio_packet_bytes to fixed_audio_packet_size

Browse Source
This commit is contained in:
Varun Pratap Singh
2026-01-12 22:11:39 +05:30
parent 14a115f372
commit 3e982f7a4a
2 changed files with 7 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
changelog/3410.changed.md
View File

@@ -1 +1,2 @@
- Enhanced `FastAPIWebsocketTransport` with optional protocol-level audio packetization to support media endpoints requiring strict framing and real-time pacing. - Enhanced `FastAPIWebsocketTransport` with optional protocol-level audio packetization to support media endpoints requiring strict framing and real-time pacing.
- Renamed `audio_packet_bytes` to `fixed_audio_packet_size` for clearer audio framing semantics.

12
src/pipecat/transports/websocket/fastapi.py
View File

@@ -56,14 +56,14 @@ class FastAPIWebsocketParams(TransportParams):
add_wav_header: Whether to add WAV headers to audio frames. add_wav_header: Whether to add WAV headers to audio frames.
serializer: Frame serializer for encoding/decoding messages. serializer: Frame serializer for encoding/decoding messages.
session_timeout: Session timeout in seconds, None for no timeout. session_timeout: Session timeout in seconds, None for no timeout.
audio_packet_bytes: Optional fixed-size packetization for raw PCM audio payloads. fixed_audio_packet_size: Optional fixed-size packetization for raw PCM audio payloads.
Useful when the remote WebSocket media endpoint requires strict audio framing. Useful when the remote WebSocket media endpoint requires strict audio framing.
""" """
add_wav_header: bool = False add_wav_header: bool = False
serializer: Optional[FrameSerializer] = None serializer: Optional[FrameSerializer] = None
session_timeout: Optional[int] = None session_timeout: Optional[int] = None
audio_packet_bytes: Optional[int] = None fixed_audio_packet_size: Optional[int] = None
class FastAPIWebsocketCallbacks(BaseModel): class FastAPIWebsocketCallbacks(BaseModel):
@@ -366,7 +366,7 @@ class FastAPIWebsocketOutputTransport(BaseOutputTransport):
# Buffer for optional protocol-level audio packetization. # Buffer for optional protocol-level audio packetization.
# Some serializers may emit arbitrarily sized raw PCM payloads, while # Some serializers may emit arbitrarily sized raw PCM payloads, while
# certain downstream transports or media endpoints require audio to be # certain downstream transports or media endpoints require audio to be
# sent in fixed-size frames. When `params.audio_packet_bytes` is set, # sent in fixed-size frames. When `params.fixed_audio_packet_size` is set,
# this buffer accumulates outgoing audio until a full packet can be # this buffer accumulates outgoing audio until a full packet can be
# emitted, preserving any remainder for subsequent sends. # emitted, preserving any remainder for subsequent sends.
self._audio_send_buffer = bytearray() self._audio_send_buffer = bytearray()
@@ -429,7 +429,7 @@ class FastAPIWebsocketOutputTransport(BaseOutputTransport):
if isinstance(frame, InterruptionFrame): if isinstance(frame, InterruptionFrame):
# Drop any partially buffered audio to avoid replaying stale PCM # Drop any partially buffered audio to avoid replaying stale PCM
if self._params.audio_packet_bytes: if self._params.fixed_audio_packet_size:
self._audio_send_buffer.clear() self._audio_send_buffer.clear()
await self._write_frame(frame) await self._write_frame(frame)
@@ -497,8 +497,8 @@ class FastAPIWebsocketOutputTransport(BaseOutputTransport):
if payload: if payload:
# Optional protocol-level audio packetization: # Optional protocol-level audio packetization:
# If a downstream WebSocket media endpoint requires fixed-size PCM frames, # If a downstream WebSocket media endpoint requires fixed-size PCM frames,
# configure params.audio_packet_bytes (e.g. 640 for 20ms @ 16kHz PCM16 mono). # configure params.fixed_audio_packet_size (e.g. 640 for 20ms @ 16kHz PCM16 mono).
packet_bytes = self._params.audio_packet_bytes packet_bytes = self._params.fixed_audio_packet_size
if packet_bytes and isinstance(payload, (bytes, bytearray)): if packet_bytes and isinstance(payload, (bytes, bytearray)):
self._audio_send_buffer.extend(bytes(payload)) self._audio_send_buffer.extend(bytes(payload))