get rid of aom obus stuff again

facefusion/apis/stream_helper.py

@@ -10,7 +10,7 @@ from starlette.websockets import WebSocket, WebSocketState
 from facefusion import rtc_store, session_context, session_manager, state_manager
 from facefusion.apis.api_helper import get_sec_websocket_protocol
 from facefusion.apis.session_helper import extract_access_token
-from facefusion.codecs.aom import create_aom_encoder, destroy_aom_encoder, encode_aom_buffer, extract_aom_obus
+from facefusion.codecs.aom import create_aom_encoder, destroy_aom_encoder, encode_aom_buffer
 from facefusion.codecs.opus import create_opus_encoder, destroy_opus_encoder, encode_opus_buffer
 from facefusion.codecs.vpx import create_vpx_encoder, destroy_vpx_encoder, encode_vpx_buffer
 from facefusion.streamer import process_vision_frame
@@ -65,8 +65,7 @@ def run_aom_encode_loop(vision_frame_deque : deque[VisionFrame], session_id : Se
 			frame_buffer = encode_aom_buffer(aom_encoder, yuv_frame.tobytes(), frame_resolution, pts)
 
 			if frame_buffer:
-				for obu_buffer in extract_aom_obus(frame_buffer):
-					rtc_store.send_rtc_video(session_id, obu_buffer)
+				rtc_store.send_rtc_video(session_id, frame_buffer)
 
 		pts += 1
 

facefusion/codecs/aom.py

@@ -1,6 +1,6 @@
 import ctypes
 import struct
-from typing import List, Optional
+from typing import Optional
 
 from facefusion.libraries import aom as aom_module
 from facefusion.types import AomEncoder, BitRate, Resolution
@@ -65,47 +65,6 @@ def collect_aom_buffer(aom_encoder : AomEncoder) -> bytes:
 	return output_buffer
 
 
-# TODO: try to eliminate this
-def extract_aom_obus(frame_buffer : bytes) -> List[bytes]:
-	obu_list : List[bytes] = []
-	offset = 0
-
-	while offset < len(frame_buffer):
-		header_offset = offset
-		header = frame_buffer[offset]
-		obu_type = (header >> 3) & 0x0F
-		has_extension = (header >> 2) & 0x01
-		has_size = (header >> 1) & 0x01
-		offset += 1 + has_extension
-
-		obu_size = 0
-
-		if has_size:
-			shift = 0
-
-			while offset < len(frame_buffer):
-				leb_byte = frame_buffer[offset]
-				offset += 1
-				obu_size |= (leb_byte & 0x7F) << shift
-				shift += 7
-
-				if not (leb_byte & 0x80):
-					break
-
-		payload_offset = offset
-		offset += obu_size
-
-		if obu_type != 2:
-			clean_header = bytes([ header & 0xFD ])
-
-			if has_extension:
-				clean_header += frame_buffer[header_offset + 1:header_offset + 2]
-
-			obu_list.append(clean_header + frame_buffer[payload_offset:payload_offset + obu_size])
-
-	return obu_list
-
-
 def destroy_aom_encoder(aom_encoder : AomEncoder) -> None:
 	aom_library = aom_module.create_static_library()
 

facefusion/libraries/datachannel.py

@@ -258,6 +258,7 @@ def define_rtc_packetizer_init() -> ctypes.Structure:
 			('sequenceNumber', ctypes.c_uint16),
 			('timestamp', ctypes.c_uint32),
 			('maxFragmentSize', ctypes.c_uint16),
+			('nalSeparator', ctypes.c_int),
 			('obuPacketization', ctypes.c_int)
 		]
 	})()

facefusion/rtc.py

@@ -122,6 +122,7 @@ def add_video_track(peer_connection : PeerConnection, media_direction : MediaDir
 	video_packetizer.maxFragmentSize = 1200
 
 	if video_codec == 'av1':
+		video_packetizer.obuPacketization = 1
 		datachannel_library.rtcSetAV1Packetizer(video_track, ctypes.byref(video_packetizer))
 	if video_codec == 'vp8':
 		datachannel_library.rtcSetVP8Packetizer(video_track, ctypes.byref(video_packetizer))

tests/test_codec_opus.py

@@ -36,7 +36,7 @@ def test_encode_opus_buffer() -> None:
 		assert create_hash(encode_opus_buffer(opus_encoder, audio_sample.tobytes(), 960)) == '8abe71cf'
 
 	if is_macos():
-		assert create_hash(encode_opus_buffer(opus_encoder, audio_sample.tobytes(), 960)) == '8ecd1108'
+		assert create_hash(encode_opus_buffer(opus_encoder, audio_sample.tobytes(), 960)) == 'f595020f'
 
 
 def test_destroy_opus_encoder() -> None: