From 49c00a7d6cb9c05d4bd0bb0fc8b59a2eed4b8950 Mon Sep 17 00:00:00 2001
From: huangmingming <huangmingming@deepscience.cn>
Date: 星期一, 13 三月 2023 12:07:11 +0800
Subject: [PATCH] grpc client remove VAD
---
funasr/runtime/python/grpc/grpc_main_client_mic.py | 41 +++++++++++++----------------------------
1 files changed, 13 insertions(+), 28 deletions(-)
diff --git a/funasr/runtime/python/grpc/grpc_main_client_mic.py b/funasr/runtime/python/grpc/grpc_main_client_mic.py
index 8a8fe4d..220e8b5 100644
--- a/funasr/runtime/python/grpc/grpc_main_client_mic.py
+++ b/funasr/runtime/python/grpc/grpc_main_client_mic.py
@@ -1,38 +1,25 @@
import pyaudio
-import scipy.io.wavfile as wav
-import grpc_client
import grpc
import json
-from grpc_client import transcribe_audio_bytes
-from paraformer_pb2_grpc import ASRStub
-import webrtcvad
-import numpy as np
import time
import asyncio
-import datetime
import argparse
+
+from grpc_client import transcribe_audio_bytes
+from paraformer_pb2_grpc import ASRStub
async def deal_chunk(sig_mic):
global stub,SPEAKING,asr_user,language,sample_rate
- sig = np.frombuffer(sig_mic, 'int16')
- if vad.is_speech(sig.tobytes(), sample_rate): #speaking
- SPEAKING = True
- response = transcribe_audio_bytes(stub, sig, user=asr_user, language=language, speaking = True, isEnd = False) #speaking, send audio to server.
- else: #silence
- begin_time = 0
- if SPEAKING: #means we have some audio recorded, send recognize order to server.
- SPEAKING = False
- begin_time = int(round(time.time() * 1000))
- response = transcribe_audio_bytes(stub, None, user=asr_user, language=language, speaking = False, isEnd = False) #speak end, call server for recognize one sentence
- resp = response.next()
- if "decoding" == resp.action:
- resp = response.next() #TODO, blocking operation may leads to miss some audio clips. C++ multi-threading is preferred.
- if "finish" == resp.action:
- end_time = int(round(time.time() * 1000))
- print (json.loads(resp.sentence))
- print ("delay in ms: %d " % (end_time - begin_time))
- else:
- pass
+
+ SPEAKING = True
+ resp = transcribe_audio_bytes(stub, sig_mic, user=asr_user, language=language, speaking = True, isEnd = False) #speaking, send audio to server.
+
+ if "decoding" == resp.action: #partial result
+ print(json.loads(resp.sentence))
+ elif "finish" == resp.action: #final result
+ print (json.loads(resp.sentence))
+
+
async def record(host,port,sample_rate,mic_chunk,record_seconds,asr_user,language):
@@ -92,8 +79,6 @@
language = 'zh-CN'
- vad = webrtcvad.Vad()
- vad.set_mode(1)
FORMAT = pyaudio.paInt16
CHANNELS = 1
--
Gitblit v1.9.1