FilesExpand file tree

main

/

test_tts_stream.py

Copy path

Blame

More file actions

Blame

More file actions

Latest commit

 

History

History

History

121 lines (93 loc) · 2.97 KB

main

/

test_tts_stream.py

Copy path

Top

File metadata and controls

• Code
• Blame

121 lines (93 loc) · 2.97 KB

Raw

Copy raw file

Download raw file

Open symbols panel

Edit and raw actions

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

import asyncio

from io import BytesIO

import numpy as np

import cv2

import time

from rtmp_streaming import StreamerConfig, Streamer

from queue import Queue

import edge_tts

import soundfile as sf

import resampy

video_file_path = "./test.mp4"

rtmp_server = "rtmp://localhost:1935/live/livestream"

queue = Queue()

fps = 25

sample_rate = 16000

chunk = 320 # 320 samples per chunk (20ms * 16000 / 1000)

input_stream = BytesIO()

voice_name = "zh-CN-YunxiaNeural"

text = "掌握模式识别中涉及的相关概念、算法很重要！真的很重要！"

def create_bytes_stream(byte_stream):

stream, sr = sf.read(byte_stream)

print(f'[INFO]tts audio stream {sr}: {stream.shape}')

stream = stream.astype(np.float32)

if stream.ndim > 1:

print(f'[WARN] audio has {stream.shape[1]} channels, only use the first.')

stream = stream[:, 0]

if sr != sample_rate and stream.shape[0] > 0:

print(f'[WARN] audio sample rate is {sr}, resampling into {sample_rate}.')

stream = resampy.resample(x=stream, sr_orig=sr, sr_new=sample_rate)

return stream

def push_audio(buffer):

input_stream.write(buffer)

if len(buffer) > 0:

input_stream.seek(0)

stream = create_bytes_stream(input_stream)

streamlen = stream.shape[0]

idx = 0

while streamlen >= chunk:

queue.put(stream[idx:idx + chunk])

streamlen -= chunk

idx += chunk

input_stream.seek(0)

input_stream.truncate()

async def edge(_text):

communicate = edge_tts.Communicate(_text, voice_name)

res = None

async for chunk in communicate.stream():

if chunk["type"] == "audio":

if res is None:

res = chunk["data"]

else:

res = res + chunk["data"]

elif chunk["type"] == "WordBoundary":

pass

push_audio(res)

def run_tts(_text):

asyncio.get_event_loop().run_until_complete(edge(_text))

cap = cv2.VideoCapture(video_file_path)

ret, frame = cap.read()

sc = StreamerConfig()

sc.source_width = frame.shape[1]

sc.source_height = frame.shape[0]

sc.stream_width = frame.shape[1]

sc.stream_height = frame.shape[0]

sc.stream_fps = fps

sc.stream_bitrate = 4000000

sc.stream_profile = 'baseline'

sc.audio_channel = 1

sc.sample_rate = sample_rate

sc.stream_server = rtmp_server

streamer = Streamer()

streamer.init(sc)

streamer.enable_av_debug_log()

run_tts(text)

while True:

ret, frame = cap.read()

if not ret:

break

t = time.time()

streamer.stream_frame(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)) # frame.flatten()

if not queue.empty():

for i in range(2):

audio_frame = queue.get()

streamer.stream_frame_audio(audio_frame)

else:

for i in range(2):

audio_frame = np.zeros(chunk, dtype=np.float32)

streamer.stream_frame_audio(audio_frame)

delay = 0.04 - (time.time() - t)

if delay > 0:

time.sleep(delay)

cap.release()

cv2.destroyAllWindows()