[ARCHIVED] Transcriptions for Studio Recordings

[edwardn1]

Since this is not noted in the documentation and is a feature of Studio, Support suggested I create this discussion so others are aware of it.

You can get a transcription of any Studio recording that has added Captions.  It will require some editing but is easy to do.

1.  Publish captions for the recordings.

2.  Open the recording, and select "Captions"

3.  Click on the 3"dots" at the top right and select "download".  This will create a zip file.

4.  Open the zip file and click on either of the "srt" files within.  You will be prompted to select an application to open the file.   You can select Word or Notepad to open it.

5.  The transcription will display.  Edit it to delete all of the "time" notations.

Video demonstration:  Transcript for Studio transcription #studio

[Stef_retired]

 @edwardn1 ‌, thanks so much for providing this useful information. Our Doc Team is reviewing the Studio documentation to determine how best to incorporate it.

[edwardn1]

Great. I was told by Support that it wasn't possible so seems like it needs to be documented.

[acolvin]

I'm confused about deleting the time notations. If you delete those, then how does the video know were to place the text from the SRT?

[edwardn1]

I would copy/paste it onto a Word doc and edit on that. You don’t remove the timings in Studio.

Get Outlook for iOS<https://aka.ms/o0ukef>

[Landmark_Coll]

Thank you for posting this technique for getting video transcripts.

Unfortunately, this is a highly inefficient manual process which leaves out the steps of uploading the edited transcript back to the course and hyperlinking it so users can get a transcript on demand for their screenreader or other purposes. Universal Design practices suggest captions and transcript support different purposes. It is our college's policy to provide both captions and transcripts for all videos we upload.

Since captioning already provides Studio with the text needed for a transcript, Studio should provide a link to a transcript with embedded time codes (as found in VTT or SRT files). This link should appear just below the embedded video automatically without needing to edit the page. The link could just be the word "Transcript" or be an icon (with appropriate ALT text) signifying a transcript is available.

Without this automation it's necessary for users to manually edit a text file, upload it to each course where the video may appear, and manually create a hyperlink to the transcript separately on any page in the course(s) where the video may be embedded.

Hopefully, this suggestion gets due consideration.

Michael Keen

Landmark College

[Landmark_Coll]

I made a typo in my post above. In the second paragraph, the first sentence should read:

Since captioning already provides Studio with the text needed for a transcript, Studio should provide a link to a transcript without embedded time codes (as found in VTT or SRT files).

Sorry.

[Landmark_Coll]

Hi all,

I have documented a technique which I believe is somewhat easier than the OP's post. Please see:

https://etop.landmark.edu/transcript.html

I still feel strongly that Studio should include a "Download Transcript" option directly in the video player, but this gets the job done.

[lalcalay]

Hi. This is my response to the Community Participant from Landmark College who contributed this method for obtaining transcripts from the downloaded str files of the CCs generated by Canvas Studio. Thank you for recognizing this need, and for the need for an efficient way to do this. I haven't tried out your method yet, but I hope it will be effective. I also found a 3rd party app, which I also haven't tried, but which is a Chrome add-on and promises to accomplish this same function. And it was suggested to me that if I am recording my own lecture via an audio recording feature on my smartphone, that I would be better off using a speech-to-text app which would bypass the need to generate CCs. I originally had the idea of running an audio recording through Studio, generating CCs, downloading them, then editing that file into a wrapped text for a readable transcript, which is much more efficient timewise to process than an audio lecture. I did initially try to copy paste the srt file into a Word document, then do a global search and replace function for each digit successively (i.e., 0-1-2, etc.). That eliminated all the numerals, but not the symbols or the punctuation that remained as vestiges of the time stamps. I believe there is a way to remove hard breaks in this search and replace function, and I was able to copy and paste the symbols (horizontal lines of some sort) and eliminate those, but I was reluctant to eliminate the punctuation marks, because that would also eliminate them in the CC text, and they are needed there. Anyway, I wonder how TED.com does it, because they seem to generate readable transcripts. I have also had my ESL students copy and paste transcripts of their own words, as represented by the CCs, from their own video or audio recordings, into the website tophonetics.com, where they can obtain a rendering of their text in IPA (International Phonetic Alphabet), as well as read aloud by an automated native English-speaking voice reader. This is quite empowering and instructive to them, because they have generated their own words from their own thoughts, rather than imitating words produced by others. Once they hear their own words read by a native English speaker, they can adjust their own speech to attain more clarity. Also, I have been looking for a way to produce CCs from a CC generator which does not have a semantic or syntatic processor, which could reproduce more accurately their produced speech on a phonologically processed level alone. This would provide them with visual feedback to help them perceive where their English speech production lacks clarity. I initially thought of this method, and sometimes it does work, but the CC processors now apparently incorporate multiple levels of linguistic processing (i.e., syntactic and semantic) which compensates for lack of clarity at the phonological level. Where could I find CC generators which reproduce the phonological level alone, so that speakers could get visual feedback to help them become more aware of phonological production clarity issues in their own speech?

https://subtitletools.com/convert-subtitles-to-plain-text-online