Record audio in Colab using getUserMedia({ audio: true })

record.py

# all imports
from IPython.display import Javascript
from google.colab import output
from base64 import b64decode
from io import BytesIO
!pip -q install pydub
from pydub import AudioSegment

RECORD = """
const sleep  = time => new Promise(resolve => setTimeout(resolve, time))
const b2text = blob => new Promise(resolve => {
  const reader = new FileReader()
  reader.onloadend = e => resolve(e.srcElement.result)
  reader.readAsDataURL(blob)
})

var record = time => new Promise(async resolve => {
  stream = await navigator.mediaDevices.getUserMedia({ audio: true })
  recorder = new MediaRecorder(stream)
  chunks = []
  recorder.ondataavailable = e => chunks.push(e.data)
  recorder.start()
  await sleep(time)
  recorder.onstop = async ()=>{
    blob = new Blob(chunks)
    text = await b2text(blob)
    resolve(text)
  }
  recorder.stop()
})
"""

def record(sec=3):
  display(Javascript(RECORD))
  s = output.eval_js('record(%d)' % (sec*1000))
  b = b64decode(s.split(',')[1])
  audio = AudioSegment.from_file(BytesIO(b))
  return audio

record_save.py

# all imports
from IPython.display import Javascript
from google.colab import output
from base64 import b64decode

RECORD = """
const sleep  = time => new Promise(resolve => setTimeout(resolve, time))
const b2text = blob => new Promise(resolve => {
  const reader = new FileReader()
  reader.onloadend = e => resolve(e.srcElement.result)
  reader.readAsDataURL(blob)
})

var record = time => new Promise(async resolve => {
  stream = await navigator.mediaDevices.getUserMedia({ audio: true })
  recorder = new MediaRecorder(stream)
  chunks = []
  recorder.ondataavailable = e => chunks.push(e.data)
  recorder.start()
  await sleep(time)
  recorder.onstop = async ()=>{
    blob = new Blob(chunks)
    text = await b2text(blob)
    resolve(text)
  }
  recorder.stop()
})
"""

def record(sec=3):
  display(Javascript(RECORD))
  s = output.eval_js('record(%d)' % (sec*1000))
  b = b64decode(s.split(',')[1])
  with open('audio.wav','wb') as f:
    f.write(b)
  return 'audio.wav'  # or webm ?

It's long ago, so I'm not sure where the original code come from. It's probably this one and a few others together. I also made many modifications, e.g. to use await instead of then(). @rolambert, if you find the real original source, I will add it then.

Great!, thank you so much.

This is awesome, but is there a way to add a countdown of the time that is left? I've been trying to do that for hours, but I'm not well versed enough using js in python 😔

Nevermind, I just made this solution:

RECORD = """
const sleep = time => new Promise(resolve => {
setTimeout(resolve, time)
}, )
const b2text = blob => new Promise(resolve => {
const reader = new FileReader()
reader.onloadend = e => resolve(e.srcElement.result)
reader.readAsDataURL(blob)
})
var espacio = document.querySelector("#output-area")
var record = time => new Promise(async resolve => {
stream = await navigator.mediaDevices.getUserMedia({ audio: true })
recorder = new MediaRecorder(stream)
chunks = []
recorder.ondataavailable = e => chunks.push(e.data)
recorder.start()
var numerillo = (time/1000)-1
for (var i = 0; i < numerillo; i++) {
espacio.appendChild(document.createTextNode(numerillo-i))
await sleep(1000)
espacio.removeChild(espacio.lastChild)
}
recorder.onstop = async ()=>{
blob = new Blob(chunks)
text = await b2text(blob)
resolve(text)
}
recorder.stop()
})
"""

Hey, I am quite a noob in this but I copied your code to my google colab but the speech_recognition says that it is not an "audio data"
Please help.

Thanks :)

@the-psychedelic Hi, I got a same problem.. Did you ever find a solution to the problem? I hope you found a way

Can I have some help because my problem still there since 2 days I'm trying to figure out a solution,
I did the same code as yours
thank you !

were you able to find a solution?

Hey, I am quite a noob in this but I copied your code to my google colab but the speech_recognition says that it is not an "audio data" Please help.

were you able to find a solution?

ref: https://github.com/facebookresearch/WavAugment/blob/main/examples/python/WavAugment_walkthrough.ipynb

for invalid RIFF-header error:

!pip install ffmpeg-python
import ffmpeg

def fix_riff_header(binary):
  process = (ffmpeg
    .input('pipe:0')
    .output('pipe:1', format='wav')
    .run_async(pipe_stdin=True, pipe_stdout=True, pipe_stderr=True, quiet=True, overwrite_output=True)
  )
  output, err = process.communicate(input=binary)
  
  riff_chunk_size = len(output) - 8
  # Break up the chunk size into four bytes, held in b.
  q = riff_chunk_size
  b = []
  for i in range(4):
      q, r = divmod(q, 256)
      b.append(r)

  # Replace bytes 4:8 in proc.stdout with the actual size of the RIFF chunk.
  riff = output[:4] + bytes(b) + output[8:]
  return riff

def record(sec=3):
  display(Javascript(RECORD))
  s = output.eval_js('record(%d)' % (sec*1000))
  b = b64decode(s.split(',')[1])
  b = fix_riff_header(b)
  with open('audio.wav','wb') as f:
    f.write(b)
  audio = AudioSegment.from_file(BytesIO(b))
  return audio

How can we modify seconds in record(sec=3) to record short as well as long clips?

This script is a basic example of how to continuously record audio until the user has finished speaking and then process that speech. You can customize and expand upon it for your specific use case or application.

# all imports
from IPython.display import Javascript
from google.colab import output
from base64 import b64decode
from io import BytesIO
!pip -q install pydub
from pydub import AudioSegment

RECORD = """
const sleep = time => new Promise(resolve => setTimeout(resolve, time))
const b2text = blob => new Promise((resolve, reject) => {
  const reader = new FileReader()
  reader.onloadend = e => resolve(e.target.result)
  reader.onerror = e => reject(new Error("Failed to read blob"))
  reader.readAsDataURL(blob)
})
var recordUntilSilence = time => new Promise(async (resolve, reject) => {
  let stream, recorder, chunks, blob, text, audioContext, analyser, dataArr, silenceStart, threshold = 50, silenceDelay = 2000
  try {
    stream = await navigator.mediaDevices.getUserMedia({ audio: true })
  } catch (err) {
    return reject(new Error("Failed to get media stream"))
  }
  audioContext = new AudioContext()
  const source = audioContext.createMediaStreamSource(stream)
  analyser = audioContext.createAnalyser()
  analyser.fftSize = 512
  dataArr = new Uint8Array(analyser.frequencyBinCount)
  source.connect(analyser)
  recorder = new MediaRecorder(stream)
  chunks = []
  recorder.ondataavailable = e => chunks.push(e.data)
  recorder.onstop = async () => {
    blob = new Blob(chunks)
    try {
      text = await b2text(blob)
      resolve(text)
    } catch (err) {
      reject(new Error("Failed to convert blob to text"))
    }
  }
  recorder.onerror = e => reject(new Error("Recorder error"))
  recorder.start()
  const checkSilence = () => {
    analyser.getByteFrequencyData(dataArr)
    const avg = dataArr.reduce((p, c) => p + c, 0) / dataArr.length

    if (avg < threshold) {
      if (silenceStart === null) silenceStart = new Date().getTime()
      else if (new Date().getTime() - silenceStart > silenceDelay) {
        recorder.stop()
        audioContext.close()
        return
      }
    } else {
      silenceStart = null
    }
    requestAnimationFrame(checkSilence)
  }
  silenceStart = null
  checkSilence()
})
console.log("JavaScript code executed successfully.")
"""

def record_until_silence():
  try:
    display(Javascript(RECORD))
    s = output.eval_js('recordUntilSilence()')
    b = b64decode(s.split(',')[1])
    audio = AudioSegment.from_file(BytesIO(b))
    return audio
  except Exception as e:
    print(f"An error occurred: {e}")
    return None

# from https://gist.github.com/korakot/c21c3476c024ad6d56d5f48b0bca92be

from IPython.display import Javascript
from google.colab import output
from base64 import b64decode

# RECORD = """
# const sleep = time => new Promise(resolve => setTimeout(resolve, time))
# const b2text = blob => new Promise(resolve => {
#   const reader = new FileReader()
#   reader.onloadend = e => resolve(e.srcElement.result)
#   reader.readAsDataURL(blob)
# })
# var record = time => new Promise(async resolve => {
#   stream = await navigator.mediaDevices.getUserMedia({ audio: true })
#   recorder = new MediaRecorder(stream)
#   chunks = []
#   recorder.ondataavailable = e => chunks.push(e.data)
#   recorder.start()
#   await sleep(time)
#   recorder.onstop = async ()=>{
#     blob = new Blob(chunks)
#     text = await b2text(blob)
#     resolve(text)
#   }
#   recorder.stop()
# })
# """
RECORD = """
const sleep = time => new Promise(resolve => {
setTimeout(resolve, time)
}, )
const b2text = blob => new Promise(resolve => {
const reader = new FileReader()
reader.onloadend = e => resolve(e.srcElement.result)
reader.readAsDataURL(blob)
})
var espacio = document.querySelector("#output-area")
var record = time => new Promise(async resolve => {
stream = await navigator.mediaDevices.getUserMedia({ audio: true })
recorder = new MediaRecorder(stream)
chunks = []
recorder.ondataavailable = e => chunks.push(e.data)
recorder.start()
var numerillo = (time/1000)-1
for (var i = 0; i < numerillo; i++) {
espacio.appendChild(document.createTextNode(numerillo-i))
await sleep(1000)
espacio.removeChild(espacio.lastChild)
}
recorder.onstop = async ()=>{
blob = new Blob(chunks)
text = await b2text(blob)
resolve(text)
}
recorder.stop()
})
"""
def record(sec, filename='audio.wav'):
    display(Javascript(RECORD))
    print("before s")
    s = output.eval_js('record(%d)' % (sec))
    print(s)
    b = b64decode(s.split(',')[1])
    with open(filename, 'wb+') as f:
        f.write(b)

audio = 'audio.wav'
second = 5
print(f"Speak to your microphone {second} sec...")
record(1, audio)
print("Done!")


import librosa
import librosa.display
speech, rate = librosa.load(audio)



librosa.display.waveshow(speech, sr=rate)

import matplotlib.pyplot as plt
plt.show()

import pysndfile
pysndfile.sndio.write('audio_ds.wav', speech, rate=rate, format='wav', enc='pcm16')

from IPython.display import display, Audio
display(Audio(speech, rate=rate))

I am executing the above code. But the cell doesn't stop executing. And the output till now is as follows:

Speak to your microphone 5 sec...

before s

It doesn't change even if I use the commented section instead.

I am executing this in jupyter notebook locally. Is that the problem? or is it something else.

+1