Extracting timestamps of words from an audio in Python

Question

0.00/5 (No votes)

See more:

I am working on a Speech to Text project in python using Vosk API. I am trying to get the timestamps of certain phrases present in the audio for some data analysis. I need some algorithm or some approach to how I can do the same without using Google Cloud Speech API/IBM Watson Speech API. Any sort of help is welcome.

What I have tried:

I have tried using SimpleAudioIndexer which uses pocketsphinx and Watson Cloud API but its accuracy is not up to the mark.

Posted 2-Oct-20 4:53am

Srijan Biswas

Updated 17-Jan-22 16:40pm

Add a Solution

Comments

[no name] 2-Oct-20 11:50am

If you play back the text (as text to speech), you should be able to do timings because the speech engine should have events for: word (breaks), sentences, etc. that you can hook onto.

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Member 15503411 · Answer 1 · 2022-01-17T16:40:00

You can get the vosk output to include times (at the granularity of a match sequence) by setting SetWords(True).
e.g.
rec = KaldiRecognizer(model, sample_rate)
rec.SetWords(True)
-----
This causes the output to include the result details -
e.g.
{
"partial" : "zero one eight zero"
}
{
"result" : [{
"conf" : 1.000000,
"end" : 6.690000,
"start" : 6.240000,
"word" : "zero"
}, {
"conf" : 1.000000,
"end" : 6.900000,
"start" : 6.690000,
"word" : "one"
}, {
"conf" : 1.000000,
"end" : 7.140000,
"start" : 6.900000,
"word" : "eight"
}, {
"conf" : 1.000000,
"end" : 7.500000,
"start" : 7.140000,
"word" : "zero"
}, {
"conf" : 1.000000,
"end" : 8.010000,
"start" : 7.500000,
"word" : "three"
}],
"text" : "zero one eight zero three"