Set packet_time = 125 msec instead of 20 msec (both side)

[el07694]

Hello i am wondering how can i change the packet time to be 125 msec instead of 20 msec.
Also one more question: Is there any suggested port the aiohttp server will run for aiortc purposes?

[el07694]

As for the main question i fixed by splitting the 125 msec sec to 20 msec (I use buffer variable) and by merging 20mesc packets to 125msec (I also use buffer variable).

As for the first question which port should you suggest me to use for radio broadcasting and live ip-calls?

[Chappie74]

As for the main question i fixed by splitting the 125 msec sec to 20 msec (I use buffer variable) and by merging 20mesc packets to 125msec (I also use buffer variable).

Please can you share some code on how?

[el07694]

Sure, i will share with you tomorrow morning.

Good night!

[el07694]

Here it is the code snippets (i use pydub.AudioSegment in both cases):

'''
code to convert 125 msec packets to 20 msec packets
'''
class Server_Audio_Stream_Offer(MediaStreamTrack):
    kind = "audio"

    def __init__(self,q):
        super().__init__()  # don't forget this!

        self.speackers_deck_queue = q
        self.q = Simple_Queue()

        self.codec = av.CodecContext.create('pcm_s16le', 'r')
        self.codec.sample_rate = 8000
        self.codec.channels = 2

        self.audio_samples = 0

        self.run = True
        self.mp3_q = AudioSegment.empty()
        self.packetize_correct_thread = threading.Thread(target=self.packetize_correct)
        self.packetize_correct_thread.start()

    async def recv(self):
        packet = av.Packet(self.q.get())
        frame = self.codec.decode(packet)[0]
        frame.pts = self.audio_samples
        frame.time_base = fractions.Fraction(1, self.codec.sample_rate)
        self.audio_samples += frame.samples
        return frame

    def packetize_correct(self):
        while (self.run):
            try:
                slice_125 = self.speackers_deck_queue.get()["slice"].set_frame_rate(8000)
                slice_full = self.mp3_q + slice_125
                len_slice_full = len(slice_full)
                desired_slice_len = 20
                packets = int(len_slice_full/desired_slice_len)
                for i in range(0,packets):
                    self.q.put(slice_full[i*desired_slice_len:(i+1)*desired_slice_len].raw_data)
                self.mp3_q = slice_full[packets*desired_slice_len:]
            except:
                print(traceback.format_exc())

    def stop_offering(self):
        try:
            self.run = False
            self.packetize_correct_thread.join()
        except Exception as e:
            pass



'''
code to convert 20 msec packets to 125 msec packets
'''
if len(self.ip_call_1_mp3_q) < self.packet_time:#125msec packet time
	while (len(self.ip_call_1_mp3_q) < self.packet_time):
		chunk = self.ip_call_1_packet_queue.get()["packet"]
		chunk_slice = AudioSegment(chunk, sample_width=2, frame_rate=48000, channels=2)
		self.ip_call_1_mp3_q = self.ip_call_1_mp3_q + chunk_slice
		time.sleep(0.020)
	slice = self.ip_call_1_mp3_q[0:self.packet_time]
	self.ip_call_1_mp3_q = self.ip_call_1_mp3_q[self.packet_time:]
else:
	slice = self.ip_call_1_mp3_q[0:self.packet_time]
	self.ip_call_1_mp3_q = self.ip_call_1_mp3_q[self.packet_time:]