Vorsovox - open-source auto-alignment software

Everything except LibriVox (yes, this is where knitting gets discussed. Now includes non-LV Volunteers Wanted projects)
Post Reply
kgullion
Posts: 1
Joined: July 19th, 2021, 6:19 am

Post by kgullion »

Hi, I want to share some software I've written to align an audio recording with a script. It works by transcribing an audio file, splitting up the file into segments, and aligning the best (most recent) take to the script. I hope it can make editing a less daunting process for volunteers.

It's currently hosted at https://kgullion.github.io/vorsovox/ and is run entirely in the browser. I've put an example file of me (very poorly) recording the preamble of the US constitution, along with the script here.

To use, add the files under the select tab, wait for the transcription to finish, and export the highlighted segments. Each segment can be edited by dragging the edges and previewed by double-clicking. For the test data, the output looks something like this:
Image

It does a pretty good job of aligning the file even though the transcriber had some trouble with my slurred pronunciations at times. This is because it first extracts metaphones from both the script and the transcription and aligns those instead. It should work even better on files where the speaker is trying to be legible.

This is still very much beta software. I've got a few more features I'd like to add but you're the subject matter experts, so if you have any ideas, feel free to add them. If you dabble in programming at all, you can find the source code here.
philchenevert
LibriVox Admin Team
Posts: 24587
Joined: October 17th, 2010, 9:23 pm
Location: Basking by the Bayou
Contact:

Post by philchenevert »

Would you be so kind as to make a little video showing how this would work? It seems confusing to me but then most things are confusing to me ! :D A walk through video would be great.
Peruvian owls always hunt in pairs because they are inca hoots.
89 Decibels? Easy Peasy ! https://youtu.be/aSKR55RDVpk
annise
LibriVox Admin Team
Posts: 38542
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

I'm not sure we really are the subject matter experts - we record PD texts read by people for free distribution.

I know some people do use our recordings to improve their nonnative language skills and do read along with the recording so it might help them I don't really know.
So I'm moving this to off topic
And everyone is welcome to discuss it here - :D

Anne
gagiha
Posts: 11
Joined: March 20th, 2022, 12:46 pm

Post by gagiha »

Is it free? Judging by your description, Vorsovox is the bet auto-alignment software I've ever heard about.
Post Reply