So now that all the shows are on YouTube and we've got some transcripts for those shows, I did a little experiment to see how YT's auto-sync feature works. Basically, you can upload a text file, tell it to sync and you'll have proper subtitles for the show. While YT's auto-transcribe feature is not bad, but it's also not great and apparently it doesn't help with search ranking (and possibly penalize you since sometimes it spits out gibberish which search engine interpret as spam). In other words, it's better and recommend to use proper transcriptions for the shows.
However, if we are going to start using the transcribed texts, we might want to re-evaluate how we transcribe. For audio only transcriptions there was more leeway in condensing speech, removing extra 'umms' and 'aahss', rewording sentences etc. However with video, where the text is rolled out live with the speaker, the rules are a bit different. What I'm wondering is if we can find a happy medium, where a bit more of the live speech is captured, while at the same time keeping the text fairly readable for those just reading.
Anyway, for those who are interested, here's what I did to convert the transcript into a YT friendly format and sync it to the video. It does a fairly good job for the most part but it still does require some intervention on part of the user to fix areas that aren't perfectly synced.
***********
1. Open a plain text editor and copy/paste into it from the SoTT article. Delete all instances people's names and stuff in brackets (ie, [laughter] - see red circles). Save the text file.
2. Go into creator studio in YT. Select the video you want to edit and then go to the "Subtitles/CC" tab.
3. Click on the button "Add new subtitles or CC". A drop down menu appears. Select "English". At the same time, you might also want to turn ON community contributions. A link will appear under the video. This link will allow others to edit the time stamps and text. I think what we could start doing is for videos that have a transcript uploaded, this link can be put into the spreadsheet. At some point it can be reviewed and since the auto-sync isn't perfect it could do with someone tweaking the captions that are a bit off.
4. Select upload file. A dialogue box will appear. Browse to where you have saved the text file and upload.
5. After it uploads you will see the text from your upload in the text box. Review the text quickly by scrolling through looking for weird symbols. Sometimes characters don't save correctly in text format (for example, a quotation mark copied from word doc is not the same as the one copied from a text doc). If all looks okay, click "Set timings".
6. It will take a while (depending on length of video) but you can check back periodically to see if it has finished by clicking on the greyed out button. If it is not ready you'll get a message telling you to check again.
7. Once done you'll be taken to the editing screen. Here you can review the text and fix parts that are off. It is pretty easy to edit and remove/add captions if necessary. Once finished hit "Publish".
And that's it! The other thing that is really good about using proper transcriptions is the the auto-translate feature also works better. You'll get better translations since it reads off your transcript file instead of the auto-caption.