如何让 vosk 通过语音识别加载文本

Question

我收到的是这个{“text”：“hello”}，而不是我说的“你好”。我怎样才能摆脱剩下的东西

stuff = {"text" : ''"} 部分

我尝试导入 JSON。我试图让 pip 语音来修复它。我不知道我是否遗漏了一些小东西 我使用语音识别来完成大部分工作。

# Get rid of ALSA lib error messages in Linux
import platform

if  platform.system() == "Linux":
    from ctypes import *

    # Define error handler
    error_handler = CFUNCTYPE(None, c_char_p, c_int, c_char_p, c_int, c_char_p)
    # Don't do anything if there is an error message
    def py_error_handler(filename, line, function, err, fmt):
      pass
    # Pass to C
    c_error_handler = error_handler(py_error_handler)
    asound = cdll.LoadLibrary('libasound.so')
    asound.snd_lib_error_set_handler(c_error_handler)

# Now define the voice_to_text() function for all platforms
import speech_recognition as sr
import vosk
import json
vosk.Model('folder_model for vosk')
speech = sr.Recognizer()
vosk.SetLogLevel(-1)
def newvtt():
    voice_input = ""
    with sr.Microphone() as source:
        speech.adjust_for_ambient_noise(source)
        try:
            audio = speech.listen(source)
            voice_input = speech.recognize_vosk(audio)
        except sr.UnknownValueError:
            pass
        except sr.RequestError:
            pass
        except sr.WaitTimeoutError:
            pass
    return voice_input

print(newvtt)

Answer 1

您得到的

voice_input

字符串是一个 JSON 字符串。因此，要获取实际的语音字符串，您必须解析 JSON。

实施示例：

import speech_recognition as sr
import vosk

# Import json library to parse JSON
import json

vosk.Model('folder_model for vosk')
speech = sr.Recognizer()
vosk.SetLogLevel(-1)

def newvtt():
    voice_input = ""
    with sr.Microphone() as source:
        speech.adjust_for_ambient_noise(source)
        try:
            audio = speech.listen(source)
            # Recognize speech using Vosk and get the result as JSON
            result = speech.recognize_vosk(audio)
            # Parse the JSON result to extract the text
            result_dict = json.loads(result)
            voice_input = result_dict.get("text", "")  # Get only the "text" part
        except sr.UnknownValueError:
            pass
        except sr.RequestError:
            pass
        except sr.WaitTimeoutError:
            pass
    return voice_input

print(newvtt()) # This will ensure that only 'Hello' string is returned instead of JSON

希望这有帮助。