Using Sphinx in a different way

[Maroon Sun]

Hi,

I've got a question about an uncommon potential use of Sphinx.

I've got an audio file (.wav, or similar) and I have the transcribed text of this audio.
Is it possible to use sphinx to retrive the word possitions of the text in the audio file?

Here is a sample:
Transcription Textfile looks like this:
"Paul goes to the cinema to see the latest movie of Tim Robbins"
The audio file contains the spoken text from above.

Now I want to feed the text and the audio into Sphinx and retrieve the position information of the words:

Paul: 0.2s
goes: 1.1s
to: 1.4s
the: 1.8s
cinema: 2.1s
to: 2.7s
see: 2.9s
......

Could this be something that could be done with Sphinx?

Regards
Gerwin

[Nickolay V. Shmyrev]

There is nothing uncommon here. Such question is asked every two month. You can just search this forum. It will require some trick to deal with a logn texts but in general its perfectly doable.

[Maroon Sun]

Thanks for the quick answer.
How accurate will this be?

Gerwin

[Nickolay V. Shmyrev]

20-30 ms

[gurucharan]

Hi Gerwin,

Can you let me know that have you used Sphinx 4 or any other tool to transcribe audio file into text.
If you have used Sphinx 4 please let me know the configuration you have used.

Thanks,
Gurucharan