Create your own interactive transcript

Interative transcripts are where text appears next to a video or audio and the words being spoken are highlighted on the text as the video or audio plays.

This video-post is about making your own. I assume you 1) have your own website, 2) have already transcribed your media and 3) use OSX Mac or Linux.

The programs in order of use are – Gentle forced aligner (https://github.com/lowerquality/gentle), Hyperaudio convertor (https://hyperaud.io/converter/) and Hyperaudio Lite (https://github.com/hyperaudio/hyperaudio-lite).

Note that if you don’t use OSX/Linux then, in order to get an appropriate file to feed into the Hyperaudio convertor, you can use one of the online transcription services that have free minutes such as Maestra (https://maestrasuite.com/).

Or there is an online demo of the Gentle forced aligner https://lowerquality.com/gentle/ though not sure what file size limit is for that.

I apologise for the noise that appears later in the video! Thanks for watching and don’t hesitate to ask me any questions.

10 thoughts on “Create your own interactive transcript

  1. Really keen to try this Mura but I’m not getting very far on my Linux (Ubuntu 20.04) setup!

    I’ve tried installing gentle with Docker (it hangs on ‘INFO:root:listening’; or maybe it’s not hanging but I don’t know what I’m supposed to do next); and also via downloading the source code and running the python script, but in this case I get “ModuleNotFoundError: No module named ‘twisted'” (I installed twisted for Python3 but I get the same issue).

    If anyone can furnish me with an idiot’s guide I’d be really grateful!

      1. Yeah, tried a few things but still struggling. However, the demo page works even for quite large files, so I’ve managed to get started on your process!

        Using command lines seems the most promising avenue right now but I’m not managing to get gentle to see my mp3 and txt files.

      2. Hi Neil thought i would try the command line route and whelp(!) a lot of issues to resolve! did you manage to compile all the dependcies? it takes some time!

        i have forked project here https://github.com/muranava/gentle and i think i have added all the issues e.g. note the instructions to –recurse-submodules to git clone and have added missing path –fst-root=../tools/openfst-1.6.5 to ext/install_kaldi.sh and added gfortran,python2.7 & libatlas-base-dev to install_deps.sh. Also note instructions to make openfst.

  2. That’s great, thanks to all that I’ve got it installed and working – both using the browser and the command line (the former is easier)! Cheers!

  3. Hey again Mura. I wonder if you know how I could make the search jump to the first instance of that word in the transcript? At the moment the search will highlight the results in the text, but you have to scroll down to find them. Cheers.

    1. Hi Neil, you would have to re-code this file I guess https://github.com/hyperaudio/hyperaudio-lite/blob/main/js/hyperaudio-lite-wrapper.js

      If I have time I might tinker with it but dont hold your breath – but just spent a few hours trying to do a labelling exercise in HotPotatoes that will no doubt be used only once!

      I see why you asked when you posted your full intervew with Fulcher, I have only used the tool for short recordings so search function works sufficiently for me so far.
      ta
      mura

  4. Hi Mura

    This is a great idea, and thanks for sharing this. After a lot of messing around I finally got this to work and it’s really great, nice and simple. I don’t have a Mac, but with Windows I used a subtitle creator to create a subtitle file for a video clip I had, then uploaded that to Hyperaudio, then copied and pasted the code as you say. It worked fine in the end (perhaps you need to emphasize for dummies like me that the media file should be in the same location as the index.html file),

    But my question is – I work with a Moodle so I’m not allowed to upload files to my institutional web server. I’m just wondering if this could be incorporated into moodle in any way, or I guess the students could just download the entire file and play it in their browser (if the source file were hosted on the web, and not on my computer).

    Paul

Penny for your thoughts

This site uses Akismet to reduce spam. Learn how your comment data is processed.