How to send audio and video using socket programming in Python | PyShine
Socket Programming Series

How to send audio and video using socket programming in Python

 · 24 mins read

Hello friends! Today’s tutorial is about sending audio and video streams from a server to a client. These streams will be extracted from an MP4 file. We will use UDP socket for the transfer of video stream and TCP socket for the audio stream. Each MP4 file consists of these streams, with a different frame rate and audio sampling rate. Normally, the audio sampling rate is set to 44100 Hz. However, the video frame rates can vary e.g. 15 frames per second (fps), 20 fps, 24 fps, 30 fps or even more. If we simply use opencv to extract the frames of MP4 and send it (with a delay of 1 ms showing each) then it may possible that your video won’t be playing at its inherent frame rate.

Thanks to our friend Eitan, who suggested about making this tutorial to resolve the issue of varying frame rates in the opencv videos. These variations are caused by the fact that, when we display an opencv image in a while loop, we consider it to play from the perspective delay given in the waitKey() function. The delay in waitKey() function is very important to properly render the video at its correct frame rate. So the question is how to keep the inherent frame rate of video while rendering? The probable answer is to stablize the rendering sample time (TS) in the loop, by checking if the current frame rate is higher or lower than the desired inherent frame rate. Accordingly, increment or decrement the sampling time.

The block of code below shows how to obtain the inherent frames per second (FPS) of an MP4 file.

vid = cv2.VideoCapture(filename)
FPS = vid.get(cv2.CAP_PROP_FPS)

Let’s say fps is the measured frames per second in real-time, so the code below shows how to regulate the TS:

if fps>FPS:
    TS+=0.001
elif fps<FPS:
    TS-=0.001
else:
    pass

Below is the server.py code, which runs three threads concurrently to obtain the synchronization:

executor.submit(audio_stream)  # to generate and send stream of audio data
executor.submit(video_stream_gen) # to generate stream of frames
executor.submit(video_stream) # to obtain and send video stream, with a synchronized fps

Here is the complete server.py code, please follow our previous tutorials about UDP and TCP sockets for more details on finding and running the code properly. Replace the filename = 'count.mp4' with your MP4 file, and also change the host_ip = '192.168.1.21' with yours. At the start, the audio will be extracted from this file and save as ‘temp.wav’. Also note that FFMPEG should be installed in your system path. You may download install FFMPEG from the “Get packages & executable files” section according to your OS. After installing FFMPEG, you can confirm the installation by writing this in terminal:

ffmpeg

and you will see this output like this:

ffmpeg version 3.4.1 Copyright (c) 2000-2017 the FFmpeg developers
  built with Apple LLVM version 9.0.0 (clang-900.0.39.2)
 ...
 ...

You also need PyAudio, if you can’t install it using pip installer, then please go this link and download the .whl according to your Python version. For instance if you are using Python 3.6 then you need to download this PyAudio‑0.2.11‑cp36‑cp36m‑win_amd64.whl. After that go to the location of this download and open up the power shell or terminal and use the command below:

pip3.6 install PyAudio‑0.2.11‑cp36‑cp36m‑win_amd64.whl

Windows OS

server.py

# This is server code to send video and audio frames over UDP/TCP

import cv2, imutils, socket
import numpy as np
import time
import base64
import threading, wave, pyaudio,pickle,struct
import sys
import queue
import os
# For details visit pyshine.com
q = queue.Queue(maxsize=10)

filename =  'count.mp4'
command = "ffmpeg -i {} -ab 160k -ac 2 -ar 44100 -vn {}".format(filename,'temp.wav')
os.system(command)

BUFF_SIZE = 65536
server_socket = socket.socket(socket.AF_INET,socket.SOCK_DGRAM)
server_socket.setsockopt(socket.SOL_SOCKET,socket.SO_RCVBUF,BUFF_SIZE)
host_name = socket.gethostname()
host_ip = '192.168.1.21'#  socket.gethostbyname(host_name)
print(host_ip)
port = 9688
socket_address = (host_ip,port)
server_socket.bind(socket_address)
print('Listening at:',socket_address)

vid = cv2.VideoCapture(filename)
FPS = vid.get(cv2.CAP_PROP_FPS)
global TS
TS = (0.5/FPS)
BREAK=False
print('FPS:',FPS,TS)
totalNoFrames = int(vid.get(cv2.CAP_PROP_FRAME_COUNT))
durationInSeconds = float(totalNoFrames) / float(FPS)
d=vid.get(cv2.CAP_PROP_POS_MSEC)
print(durationInSeconds,d)

def video_stream_gen():
   
    WIDTH=400
    while(vid.isOpened()):
        try:
            _,frame = vid.read()
            frame = imutils.resize(frame,width=WIDTH)
            q.put(frame)
        except:
            os._exit(1)
    print('Player closed')
    BREAK=True
    vid.release()
	

def video_stream():
    global TS
    fps,st,frames_to_count,cnt = (0,0,1,0)
    cv2.namedWindow('TRANSMITTING VIDEO')        
    cv2.moveWindow('TRANSMITTING VIDEO', 10,30) 
    while True:
        msg,client_addr = server_socket.recvfrom(BUFF_SIZE)
        print('GOT connection from ',client_addr)
        WIDTH=400
        
        while(True):
            frame = q.get()
            encoded,buffer = cv2.imencode('.jpeg',frame,[cv2.IMWRITE_JPEG_QUALITY,80])
            message = base64.b64encode(buffer)
            server_socket.sendto(message,client_addr)
            frame = cv2.putText(frame,'FPS: '+str(round(fps,1)),(10,40),cv2.FONT_HERSHEY_SIMPLEX,0.7,(0,0,255),2)
            if cnt == frames_to_count:
                try:
                    fps = (frames_to_count/(time.time()-st))
                    st=time.time()
                    cnt=0
                    if fps>FPS:
                        TS+=0.001
                    elif fps<FPS:
                        TS-=0.001
                    else:
                        pass
                except:
                    pass
            cnt+=1
            
            
            
            cv2.imshow('TRANSMITTING VIDEO', frame)
            key = cv2.waitKey(int(1000*TS)) & 0xFF	
            if key == ord('q'):
                os._exit(1)
                TS=False
                break	
                

def audio_stream():
    s = socket.socket()
    s.bind((host_ip, (port-1)))

    s.listen(5)
    CHUNK = 1024
    wf = wave.open("temp.wav", 'rb')
    p = pyaudio.PyAudio()
    print('server listening at',(host_ip, (port-1)))
    stream = p.open(format=p.get_format_from_width(wf.getsampwidth()),
                    channels=wf.getnchannels(),
                    rate=wf.getframerate(),
                    input=True,
                    frames_per_buffer=CHUNK)

    client_socket,addr = s.accept()

    while True:
        if client_socket:
            while True:
                data = wf.readframes(CHUNK)
                a = pickle.dumps(data)
                message = struct.pack("Q",len(a))+a
                client_socket.sendall(message)
                

from concurrent.futures import ThreadPoolExecutor
with ThreadPoolExecutor(max_workers=3) as executor:
    executor.submit(audio_stream)
    executor.submit(video_stream_gen)
    executor.submit(video_stream)

Usage

python server.py

After running the server.py, a queue of frames will store 10 frames, halt the execution of code and wait for the first frame to be taken out of the queue and then the process of adding and removing frames in queue goes on. Please note that windows of imshow() are placed at predefined locations, if we drag these windows, the operation of displaying frames will affect the synchronization. If the codes are running, please keep the window showing images at their original locations. This is just the proof of concept that video and audio data can be sent from sockets using Python. At any time if you want to quit, click on the image and press q. Below is the client side code. Once the the server is setup, then run the client code below and you will see how the frame rate is stabilizing and keeping the synchronization of audio and video.

client.py


# Welcome to PyShine
# This is client code to receive video and audio frames over UDP/TCP

import cv2, imutils, socket
import numpy as np
import time, os
import base64
import threading, wave, pyaudio,pickle,struct
# For details visit pyshine.com
BUFF_SIZE = 65536

BREAK = False
client_socket = socket.socket(socket.AF_INET,socket.SOCK_DGRAM)
client_socket.setsockopt(socket.SOL_SOCKET,socket.SO_RCVBUF,BUFF_SIZE)
host_name = socket.gethostname()
host_ip = '192.168.1.21'#  socket.gethostbyname(host_name)
print(host_ip)
port = 9688
message = b'Hello'

client_socket.sendto(message,(host_ip,port))

def video_stream():
	
	cv2.namedWindow('RECEIVING VIDEO')        
	cv2.moveWindow('RECEIVING VIDEO', 10,360) 
	fps,st,frames_to_count,cnt = (0,0,20,0)
	while True:
		packet,_ = client_socket.recvfrom(BUFF_SIZE)
		data = base64.b64decode(packet,' /')
		npdata = np.fromstring(data,dtype=np.uint8)
	
		frame = cv2.imdecode(npdata,1)
		frame = cv2.putText(frame,'FPS: '+str(fps),(10,40),cv2.FONT_HERSHEY_SIMPLEX,0.7,(0,0,255),2)
		cv2.imshow("RECEIVING VIDEO",frame)
		key = cv2.waitKey(1) & 0xFF
	
		if key == ord('q'):
			client_socket.close()
			os._exit(1)
			break

		if cnt == frames_to_count:
			try:
				fps = round(frames_to_count/(time.time()-st))
				st=time.time()
				cnt=0
			except:
				pass
		cnt+=1
		
			
	client_socket.close()
	cv2.destroyAllWindows() 


def audio_stream():
	
	p = pyaudio.PyAudio()
	CHUNK = 1024
	stream = p.open(format=p.get_format_from_width(2),
					channels=2,
					rate=44100,
					output=True,
					frames_per_buffer=CHUNK)
					
	# create socket
	client_socket = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
	socket_address = (host_ip,port-1)
	print('server listening at',socket_address)
	client_socket.connect(socket_address) 
	print("CLIENT CONNECTED TO",socket_address)
	data = b""
	payload_size = struct.calcsize("Q")
	while True:
		try:
			while len(data) < payload_size:
				packet = client_socket.recv(4*1024) # 4K
				if not packet: break
				data+=packet
			packed_msg_size = data[:payload_size]
			data = data[payload_size:]
			msg_size = struct.unpack("Q",packed_msg_size)[0]
			while len(data) < msg_size:
				data += client_socket.recv(4*1024)
			frame_data = data[:msg_size]
			data  = data[msg_size:]
			frame = pickle.loads(frame_data)
			stream.write(frame)

		except:
			
			break

	client_socket.close()
	print('Audio closed',BREAK)
	os._exit(1)
	


from concurrent.futures import ThreadPoolExecutor
with ThreadPoolExecutor(max_workers=2) as executor:
	executor.submit(audio_stream)
	executor.submit(video_stream)

Usage

python client.py

MAC OS

MAC OS users please increase the size of UDP datagram using below in the terminal:

 $sudo sysctl -w net.inet.udp.maxdgram=65535

If the above code does not work for MAC OS users, please use the code below:

server.py

# This is server code to send video and audio frames over UDP/TCP

import cv2, imutils, socket
import numpy as np
import time
import base64
import threading, wave, pyaudio,pickle,struct
import sys
import queue
import os
# For details visit pyshine.com
q = queue.Queue(maxsize=10)

filename =  'av.mp4'
command = "ffmpeg -i {} -ab 160k -ac 2 -ar 44100 -vn {}".format(filename,'temp.wav')
os.system(command)

BUFF_SIZE = 65536
server_socket = socket.socket(socket.AF_INET,socket.SOCK_DGRAM)
server_socket.setsockopt(socket.SOL_SOCKET,socket.SO_RCVBUF,BUFF_SIZE)
host_name = socket.gethostname()
host_ip = '192.168.1.102'#  socket.gethostbyname(host_name)
print(host_ip)
port = 9699
socket_address = (host_ip,port)
server_socket.bind(socket_address)
print('Listening at:',socket_address)

vid = cv2.VideoCapture(filename)
FPS = vid.get(cv2.CAP_PROP_FPS)

TS = (1/FPS)
BREAK=False
print('FPS:',FPS,TS)
totalNoFrames = int(vid.get(cv2.CAP_PROP_FRAME_COUNT))
durationInSeconds = float(totalNoFrames) / float(FPS)
d=vid.get(cv2.CAP_PROP_POS_MSEC)
print(durationInSeconds,d)

def video_stream_gen():
    
    WIDTH=400
    while(vid.isOpened()):
        try:
            _,frame = vid.read()
            frame = imutils.resize(frame,width=WIDTH)
            q.put(frame)
        except:
            os._exit(1)
    print('Player closed')
    BREAK=True
    vid.release()
	


def audio_stream():
    s = socket.socket()
    s.bind((host_ip, (port-1)))

    s.listen(5)
    CHUNK = 1024
    wf = wave.open("temp.wav", 'rb')
    p = pyaudio.PyAudio()
    print('server listening at',(host_ip, (port-1)))
    stream = p.open(format=p.get_format_from_width(wf.getsampwidth()),
                    channels=wf.getnchannels(),
                    rate=wf.getframerate(),
                    input=True,
                    frames_per_buffer=CHUNK)

    client_socket,addr = s.accept()

    while True:
        if client_socket:
            while True:
                data = wf.readframes(CHUNK)
                a = pickle.dumps(data)
                message = struct.pack("Q",len(a))+a
                client_socket.sendall(message)
                

t1 = threading.Thread(target=audio_stream, args=())
t2 = threading.Thread(target=video_stream_gen, args=())
t1.start()
t2.start()


fps,st,frames_to_count,cnt = (0,0,1,0)
cv2.namedWindow('TRANSMITTING VIDEO')        
cv2.moveWindow('TRANSMITTING VIDEO', 10,30) 

while True:
	msg,client_addr = server_socket.recvfrom(BUFF_SIZE)
	print('GOT connection from ',client_addr)
	WIDTH=400
	while(True):
		frame = q.get()
		encoded,buffer = cv2.imencode('.jpeg',frame,[cv2.IMWRITE_JPEG_QUALITY,80])
		message = base64.b64encode(buffer)
		server_socket.sendto(message,client_addr)
		frame = cv2.putText(frame,'FPS: '+str(round(fps,1)),(10,40),cv2.FONT_HERSHEY_SIMPLEX,0.7,(0,0,255),2)
		if cnt == frames_to_count:
			try:
				fps = (frames_to_count/(time.time()-st))
				st=time.time()
				cnt=0
				if fps>FPS:
					TS+=0.001
				elif fps<FPS:
					TS-=0.001
				else:
					pass
			except:
				pass
		cnt+=1
		
		
		
		cv2.imshow('TRANSMITTING VIDEO', frame)
		key = cv2.waitKey(int(1000*TS)) & 0xFF	
		if key == ord('q'):
			os._exit(1)
			TS=False
			break

client.py

# Welcome to PyShine
# This is client code to receive video and audio frames over UDP/TCP

import cv2, imutils, socket
import numpy as np
import time, os
import base64
import threading, wave, pyaudio,pickle,struct
# For details visit pyshine.com
BUFF_SIZE = 65536

BREAK = False
client_socket = socket.socket(socket.AF_INET,socket.SOCK_DGRAM)
client_socket.setsockopt(socket.SOL_SOCKET,socket.SO_RCVBUF,BUFF_SIZE)
host_name = socket.gethostname()
host_ip = '192.168.1.102'#  socket.gethostbyname(host_name)
print(host_ip)
port = 9699
message = b'Hello'

client_socket.sendto(message,(host_ip,port))




def audio_stream():
	
	p = pyaudio.PyAudio()
	CHUNK = 1024
	stream = p.open(format=p.get_format_from_width(2),
					channels=2,
					rate=44100,
					output=True,
					frames_per_buffer=CHUNK)
					
	# create socket
	client_socket = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
	socket_address = (host_ip,port-1)
	print('server listening at',socket_address)
	client_socket.connect(socket_address) 
	print("CLIENT CONNECTED TO",socket_address)
	data = b""
	payload_size = struct.calcsize("Q")
	while True:
		try:
			while len(data) < payload_size:
				packet = client_socket.recv(4*1024) # 4K
				if not packet: break
				data+=packet
			packed_msg_size = data[:payload_size]
			data = data[payload_size:]
			msg_size = struct.unpack("Q",packed_msg_size)[0]
			while len(data) < msg_size:
				data += client_socket.recv(4*1024)
			frame_data = data[:msg_size]
			data  = data[msg_size:]
			frame = pickle.loads(frame_data)
			stream.write(frame)

		except:
			
			break

	client_socket.close()
	print('Audio closed',BREAK)
	os._exit(1)
	


t1 = threading.Thread(target=audio_stream, args=())
t1.start()

cv2.namedWindow('RECEIVING VIDEO')        
cv2.moveWindow('RECEIVING VIDEO', 10,360) 
fps,st,frames_to_count,cnt = (0,0,20,0)
while True:
	packet,_ = client_socket.recvfrom(BUFF_SIZE)
	data = base64.b64decode(packet,' /')
	npdata = np.fromstring(data,dtype=np.uint8)

	frame = cv2.imdecode(npdata,1)
	frame = cv2.putText(frame,'FPS: '+str(fps),(10,40),cv2.FONT_HERSHEY_SIMPLEX,0.7,(0,0,255),2)
	cv2.imshow("RECEIVING VIDEO",frame)
	key = cv2.waitKey(1) & 0xFF
	
	if key == ord('q'):
		client_socket.close()
		os._exit(1)
		break

	if cnt == frames_to_count:
		try:
			fps = round(frames_to_count/(time.time()-st),1)
			st=time.time()
			cnt=0
		except:
			pass
	cnt+=1
	
		
client_socket.close()
cv2.destroyAllWindows()