Go Back   Freethought Forum > The Marketplace > Computers & Technology

Reply
 
Thread Tools Display Modes
  #1  
Old 05-30-2021, 01:54 PM
viscousmemories's Avatar
viscousmemories viscousmemories is offline
Admin
 
Join Date: Apr 2004
Location: Ann Arbor, Mi
Gender: Male
Posts: XXMXDCXXXI
Blog Entries: 1
Images: 9
Question Searching Video Content

I've been wrestling with how to create an index for a YouTube channel for a few months and haven't been able to figure out a good solution for it.

Imagine a YouTube channel with hundreds of hours of content spanning multiple topics, not all of which are covered in the title or description. Some of the videos have auto-generated transcripts, many do not.

I want to be able to search the entire channel for all mentions of a specific word or phrase and (for bonus points) get a link to that point in the video(s) where it comes up.

I know the technology exists in YouTube because if you look at the autogenerated transcript for any video, each line of the transcript timestamped and hotlinked to that point in the video. As far as I can tell though they don't expose this functionality through their API, and moreover not all videos have the autogenerated transcript.

I would settle for being able to search across the whole channel for a particular word or phrase and get a list of videos (without links to the specific point in time where the word/phrase is used) but that would be less than ideal.
Reply With Quote
Thanks, from:
Ari (05-30-2021)
  #2  
Old 05-30-2021, 02:07 PM
JoeP's Avatar
JoeP JoeP is offline
Chaos Defrost
 
Join Date: Jul 2004
Location: England/Miisaland
Gender: Male
Posts: XXXMCDXCVI
Images: 18
Default Re: Searching Video Content

Good luck
Sounds like an interesting challenge!

You want to be able search the transcripts (and the titles and descriptions I suppose, but that's not the challenge) of a set of videos (eg a single channel)?

If you can't get access to the YouTube-generated transcripts - or for the ones they haven't created transcripts for:
Are you considering downloading / streaming the videos and running a voice recognition package against the audio? I guess not. Unless it's the only option.
Would you considering paying a few $$$ for a team of Indians to write transcripts? :nope:
__________________

:roadrun:
Free thought! Please take one!

:unitedkingdom:   :southafrica:   :unitedkingdom::finland:       :eur:       :m&ms::m&ms::twix::twix: (rotated 180°):m&ms::m&ms:
Reply With Quote
  #3  
Old 05-30-2021, 04:39 PM
viscousmemories's Avatar
viscousmemories viscousmemories is offline
Admin
 
Join Date: Apr 2004
Location: Ann Arbor, Mi
Gender: Male
Posts: XXMXDCXXXI
Blog Entries: 1
Images: 9
Default Re: Searching Video Content

Quote:
Originally Posted by JoeP
You want to be able search the transcripts (and the titles and descriptions I suppose, but that's not the challenge) of a set of videos (eg a single channel)?
Exactly. I can get part of the way with youtube-dl (specifically, I can download all the transcripts for the videos that have them) but that doesn't help with the videos that don't have them or with hotlinking the text to that point in the video.

Quote:
Are you considering downloading / streaming the videos and running a voice recognition package against the audio?
I have considered playing the video and letting Otter.ai listen to it, but that seems really inefficient for hundreds of hours of video. And again the problem of the linking.

Have you noticed that you can click a sentence of dialogue in the autogenerated transcript and it takes you to that point in the video? I don't know why they don't expose whatever tech that is in the API (or didn't last time I checked).
Reply With Quote
  #4  
Old 05-30-2021, 10:44 PM
JoeP's Avatar
JoeP JoeP is offline
Chaos Defrost
 
Join Date: Jul 2004
Location: England/Miisaland
Gender: Male
Posts: XXXMCDXCVI
Images: 18
Default Re: Searching Video Content

I had not noticed that ... because I've only looked at the transcripts as captions while watching the videos ... where linking to the point you are currently watching would not be very impressive.
__________________

:roadrun:
Free thought! Please take one!

:unitedkingdom:   :southafrica:   :unitedkingdom::finland:       :eur:       :m&ms::m&ms::twix::twix: (rotated 180°):m&ms::m&ms:
Reply With Quote
Thanks, from:
viscousmemories (05-31-2021)
  #5  
Old 05-31-2021, 12:59 PM
viscousmemories's Avatar
viscousmemories viscousmemories is offline
Admin
 
Join Date: Apr 2004
Location: Ann Arbor, Mi
Gender: Male
Posts: XXMXDCXXXI
Blog Entries: 1
Images: 9
Default Re: Searching Video Content

I did a couple of experiments after my last post and it does appear that 'search' in the API also searches the transcripts, but I got some false positives when I tested it.
Reply With Quote
Reply

  Freethought Forum > The Marketplace > Computers & Technology


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

 

All times are GMT +1. The time now is 04:43 AM.


Powered by vBulletin® Version 3.8.2
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.
Page generated in 0.08553 seconds with 13 queries