如何在VOSK语音识别中使用Wave文件作为输入

我手头的项目需要获取录音文件，经代码处理后提取文件中的文本内容，再将提取得到的文本与其他文本进行匹配校验。当前遇到的问题是：代码无法正常读取、使用已录制的音频文件。

代码核心逻辑说明：

• init函数是代码的核心基础逻辑，负责参数解析、模型加载、配置初始化
• verify函数用于确认语音与文本的匹配结果

import argparse
import json
import os
import queue
import random
import sys
from difflib import SequenceMatcher
import numpy as np
import sounddevice as sd
import vosk

q = queue.Queue()

def int_or_str(text):
    """Helper function for argument parsing."""
    try:
        return int(text)
    except ValueError:
        return text


def callback(indata, frames, time, status):
    """This is called (from a separate thread) for each audio block."""
    if status:
        print(status, file=sys.stderr)
    q.put(bytes(indata))


def init():
    parser = argparse.ArgumentParser(add_help=False)
    parser.add_argument(
        '-l', '--list-devices', action='store_true',
        help='show list of audio devices and exit')
    args, remaining = parser.parse_known_args()
    if args.list_devices:
        print(sd.query_devices())
        parser.exit(0)
    parser = argparse.ArgumentParser(
        description=__doc__,
        formatter_class=argparse.RawDescriptionHelpFormatter,
        parents=[parser])
    parser.add_argument(
        '-f', '--filename', type=str, metavar='FILENAME',
        help='audio file to store recording to')
    parser.add_argument(
        '-m', '--model', type=str, metavar='MODEL_PATH',
        help='Path to the model')
    parser.add_argument(
        '-d', '--device', type=int_or_str,
        help='input device (numeric ID or substring)')
    parser.add_argument(
        '-r', '--samplerate', type=int, help='sampling rate')
    args = parser.parse_args(remaining)
    try:
        if args.model is None:
            args.model = "model"
        if not os.path.exists(args.model):
            print("Please download a model for your language from vosk official model page")
            print("and unpack as 'model' in the current folder.")
            parser.exit(0)
        if args.samplerate is None:
            device_info = sd.query_devices(args.device, 'input')
            # soundfile expects an int, sounddevice provides a float:
            args.samplerate = int(device_info['default_samplerate'])

        model = vosk.Model(args.model)

        if args.filename:
            dump_fn = open(args.filename, "wb")
        else:
            dump_fn = None

    except KeyboardInterrupt:
        print('\nDone')
        parser.exit(0)
    except Exception as e:
        parser.exit(type(e).__name__ + ': ' + str(e))

    return model, args
def verify(random_sentence, model, args):
    num, T_num, F_num, num_word = 0, 0, 0, 1
    with sd.RawInputStream(samplerate=args.samplerate, blocksize=8000, device=args.device, dtype='int16',
                           channels=1, callback=callback):
        rec = vosk.KaldiRecognizer(model, args.samplerate)
        print("{}) ".format(num_word), random_sentence, end='\n')
        print('=' * 30, end='\n')
        run = True
        while run:
            data = q.get()
            if rec.AcceptWaveform(data):
                res = json.loads(rec.FinalResult())
                res['text'] = res['text'].replace('ي', 'ی')             
                if SequenceMatcher(None, random_sentence, res['text']).ratio() > 0.65:                    
                    T_num, num, num_word += 1
                    
                else:
                    F_num, num, num_word += 1
                    
                run = False

    print('=' * 30)
    print('True Cases : {}\n False Cases : {}'.format(T_num, F_num))


if __name__ == "__main__":
    model, args = init()
    verify(random_sentences, model, args)

原有代码仅支持从麦克风实时采集音频做识别，没有实现本地已录制音频文件的读取逻辑，所以无法直接使用本地音频文件做校验。

前置依赖安装 #

先安装音频文件读取库：

pip install soundfile

核心修改点 #

1. 移除原有实时流相关的队列、callback逻辑，新增本地音频文件读取逻辑
2. 修改参数配置，新增本地音频文件路径必填参数
3. 自动适配音频格式为vosk要求的16位单声道格式，兼容不同采样率的音频文件

修复后的完整代码 #

import argparse
import json
import os
import sys
from difflib import SequenceMatcher
import soundfile as sf
import vosk
import numpy as np

def int_or_str(text):
    """参数解析辅助函数"""
    try:
        return int(text)
    except ValueError:
        return text

def init():
    parser = argparse.ArgumentParser(add_help=False)
    parser.add_argument(
        '-l', '--list-devices', action='store_true',
        help='显示音频设备列表并退出')
    args, remaining = parser.parse_known_args()
    # 保留原有设备列表查询能力，不需要可删除
    if args.list_devices:
        import sounddevice as sd
        print(sd.query_devices())
        parser.exit(0)
    parser = argparse.ArgumentParser(
        description=__doc__,
        formatter_class=argparse.RawDescriptionHelpFormatter,
        parents=[parser])
    parser.add_argument(
        '-a', '--audio-path', type=str, metavar='AUDIO_PATH', required=True,
        help='待识别的本地音频文件路径')
    parser.add_argument(
        '-m', '--model', type=str, metavar='MODEL_PATH',
        help='Vosk模型路径')
    parser.add_argument(
        '-r', '--samplerate', type=int, help='采样率，不填则自动读取音频文件的采样率')
    args = parser.parse_args(remaining)
    try:
        if args.model is None:
            args.model = "model"
        if not os.path.exists(args.model):
            print("请下载对应语言的Vosk模型，解压后放置在当前目录下命名为model文件夹")
            parser.exit(0)
        # 读取音频文件获取采样率
        if args.samplerate is None:
            with sf.SoundFile(args.audio_path) as f:
                args.samplerate = f.samplerate
        model = vosk.Model(args.model)
    except KeyboardInterrupt:
        print('\n已退出')
        parser.exit(0)
    except Exception as e:
        parser.exit(type(e).__name__ + ': ' + str(e))
    return model, args

def verify(random_sentence, audio_path, model, args):
    T_num, F_num = 0, 0
    rec = vosk.KaldiRecognizer(model, args.samplerate)
    rec.SetWords(True)
    # 逐块读取音频文件
    with sf.SoundFile(audio_path) as f:
        # 自动转成单声道16位整型
        if f.channels > 1:
            data = np.mean(f.read(dtype='int16'), axis=1, dtype='int16')
        else:
            data = f.read(dtype='int16')
        # 分块送入识别器
        block_size = 8000
        for i in range(0, len(data), block_size):
            block = data[i:i+block_size].tobytes()
            if rec.AcceptWaveform(block):
                break
        # 获取最终识别结果
        res = json.loads(rec.FinalResult())
        res['text'] = res['text'].replace('ي', 'ی')
        # 匹配校验
        if SequenceMatcher(None, random_sentence, res['text']).ratio() > 0.65:
            T_num += 1
            print("识别结果匹配：", res['text'])
        else:
            F_num += 1
            print("识别结果不匹配：", res['text'])
    print('=' * 30)
    print('匹配成功数: {}\n匹配失败数: {}'.format(T_num, F_num))

if __name__ == "__main__":
    model, args = init()
    # 这里替换成你实际要匹配的目标文本
    target_sentence = "你要匹配的目标文本内容"
    verify(target_sentence, args.audio_path, model, args)

使用方式 #

运行时通过参数指定音频路径即可：

python 脚本名.py -a 你的音频文件路径.wav

支持wav、flac等无压缩音频格式，若是mp3等压缩格式需要先转成wav格式再使用。

内容的提问来源于stack exchange，提问作者Ebrahim Mohammadi