#!/usr/bin/env python3 # -*- coding: utf-8 -*- """ 语音识别诊断工具 用于测试和诊断语音识别功能的具体问题 """ import asyncio import json import gzip import uuid import numpy as np import wave import os from typing import Optional class ASRDiagnostic: """ASR诊断工具""" def __init__(self): self.api_config = { 'asr': { 'appid': "8718217928", 'cluster': "volcano_tts", 'token': "ynJMX-5ix1FsJvswC9KTNlGUdubcchqc", 'ws_url': "wss://openspeech.bytedance.com/api/v2/asr" } } def generate_asr_header(self, message_type=1, message_type_specific_flags=0): """生成ASR头部""" PROTOCOL_VERSION = 0b0001 DEFAULT_HEADER_SIZE = 0b0001 JSON = 0b0001 GZIP = 0b0001 header = bytearray() header.append((PROTOCOL_VERSION << 4) | DEFAULT_HEADER_SIZE) header.append((message_type << 4) | message_type_specific_flags) header.append((JSON << 4) | GZIP) header.append(0x00) # reserved return header def parse_asr_response(self, res): """解析ASR响应""" print(f"🔍 解析响应，原始大小: {len(res)} 字节") if len(res) < 8: print(f"❌ 响应太短，无法解析") return {} try: message_type = res[1] >> 4 payload_size = int.from_bytes(res[4:8], "big", signed=False) payload_msg = res[8:8+payload_size] print(f"📋 消息类型: {message_type}, 载荷大小: {payload_size}") if message_type == 0b1001: # SERVER_FULL_RESPONSE try: if payload_msg.startswith(b'{'): result = json.loads(payload_msg.decode('utf-8')) print(f"✅ 成功解析JSON响应") return result else: print(f"❌ 响应不是JSON格式") except Exception as e: print(f"❌ JSON解析失败: {e}") except Exception as e: print(f"❌ 响应解析异常: {e}") return {} async def test_asr_with_audio_file(self, audio_file_path: str): """使用音频文件测试ASR""" print(f"🎵 测试ASR - 音频文件: {audio_file_path}") if not os.path.exists(audio_file_path): print(f"❌ 音频文件不存在: {audio_file_path}") return try: # 读取音频文件 with wave.open(audio_file_path, 'rb') as wf: channels = wf.getnchannels() width = wf.getsampwidth() rate = wf.getframerate() frames = wf.readframes(wf.getnframes()) print(f"📊 音频信息: 采样率={rate}Hz, 声道={channels}, 位深={width*8}bits") print(f"📊 音频大小: {len(frames)} 字节") # 如果是立体声，转换为单声道 if channels > 1: audio_array = np.frombuffer(frames, dtype=np.int16) audio_array = audio_array.reshape(-1, channels) audio_array = np.mean(audio_array, axis=1).astype(np.int16) frames = audio_array.tobytes() print(f"🔄 已转换为单声道") return await self._test_asr_connection(frames) except Exception as e: print(f"❌ 音频文件处理失败: {e}") return None async def test_asr_with_silence(self): """测试静音音频""" print(f"🔇 测试ASR - 静音音频") # 生成3秒的静音音频 (16kHz, 16bit, 单声道) duration = 3 # 秒 sample_rate = 16000 silence_data = bytes(duration * sample_rate * 2) # 2 bytes per sample return await self._test_asr_connection(silence_data) async def test_asr_with_noise(self): """测试噪音音频""" print(f"📢 测试ASR - 噪音音频") # 生成3秒的随机噪音 duration = 3 # 秒 sample_rate = 16000 noise_data = np.random.randint(-32768, 32767, duration * sample_rate, dtype=np.int16) noise_data = noise_data.tobytes() return await self._test_asr_connection(noise_data) async def _test_asr_connection(self, audio_data: bytes): """测试ASR连接""" try: import websockets # 构建请求参数 reqid = str(uuid.uuid4()) request_params = { 'app': { 'appid': self.api_config['asr']['appid'], 'cluster': self.api_config['asr']['cluster'], 'token': self.api_config['asr']['token'], }, 'user': { 'uid': 'asr_diagnostic' }, 'request': { 'reqid': reqid, 'nbest': 1, 'workflow': 'audio_in,resample,partition,vad,fe,decode,itn,nlu_punctuate', 'show_language': False, 'show_utterances': False, 'result_type': 'full', "sequence": 1 }, 'audio': { 'format': 'wav', 'rate': 16000, 'language': 'zh-CN', 'bits': 16, 'channel': 1, 'codec': 'raw' } } print(f"📋 ASR请求参数:") print(f" - AppID: {request_params['app']['appid']}") print(f" - Cluster: {request_params['app']['cluster']}") print(f" - Token: {request_params['app']['token'][:20]}...") print(f" - RequestID: {reqid}") # 构建请求 payload_bytes = str.encode(json.dumps(request_params)) payload_bytes = gzip.compress(payload_bytes) full_client_request = bytearray(self.generate_asr_header()) full_client_request.extend((len(payload_bytes)).to_bytes(4, 'big')) full_client_request.extend(payload_bytes) # 设置认证头 additional_headers = {'Authorization': 'Bearer; {}'.format(self.api_config['asr']['token'])} print(f"📡 连接WebSocket...") # 连接WebSocket async with websockets.connect( self.api_config['asr']['ws_url'], additional_headers=additional_headers, max_size=1000000000 ) as ws: print(f"✅ WebSocket连接成功") # 发送请求 print(f"📤 发送ASR配置请求...") await ws.send(full_client_request) res = await ws.recv() result = self.parse_asr_response(res) print(f"📥 配置响应: {result}") # 发送音频数据 chunk_size = int(1 * 2 * 16000 * 15000 / 1000) # 1秒 chunks total_chunks = 0 for offset in range(0, len(audio_data), chunk_size): chunk = audio_data[offset:offset + chunk_size] last = (offset + chunk_size) >= len(audio_data) payload_bytes = gzip.compress(chunk) audio_only_request = bytearray( self.generate_asr_header( message_type=0b0010, message_type_specific_flags=0b0010 if last else 0 ) ) audio_only_request.extend((len(payload_bytes)).to_bytes(4, 'big')) audio_only_request.extend(payload_bytes) await ws.send(audio_only_request) res = await ws.recv() result = self.parse_asr_response(res) total_chunks += 1 if last: print(f"📨 发送最后一块音频数据 (总计{total_chunks}块)") # 获取最终结果 print(f"🎯 等待最终识别结果...") if 'payload_msg' in result and 'result' in result['payload_msg']: results = result['payload_msg']['result'] print(f"📝 ASR返回结果数量: {len(results)}") if results: text = results[0].get('text', '识别失败') print(f"✅ 识别结果: {text}") return text else: print(f"❌ ASR结果为空") else: print(f"❌ ASR响应格式异常: {result.keys()}") print(f"📋 完整响应: {result}") return None except Exception as e: print(f"❌ ASR连接异常: {e}") import traceback print(f"❌ 详细错误:\n{traceback.format_exc()}") return None async def run_diagnostic(self): """运行完整诊断""" print("🔧 ASR诊断工具") print("=" * 50) # 1. 测试静音 print("\n1️⃣ 测试静音识别...") await self.test_asr_with_silence() # 2. 测试噪音 print("\n2️⃣ 测试噪音识别...") await self.test_asr_with_noise() # 3. 测试录音文件（如果存在） recording_files = [f for f in os.listdir('.') if f.startswith('recording_') and f.endswith('.wav')] if recording_files: print(f"\n3️⃣ 测试录音文件...") for file in recording_files[:3]: # 最多测试3个文件 await self.test_asr_with_audio_file(file) else: print(f"\n3️⃣ 跳过录音文件测试 (无录音文件)") print(f"\n✅ 诊断完成") def main(): """主函数""" diagnostic = ASRDiagnostic() try: asyncio.run(diagnostic.run_diagnostic()) except KeyboardInterrupt: print(f"\n🛑 诊断被用户中断") except Exception as e: print(f"❌ 诊断工具异常: {e}") if __name__ == "__main__": main()