Q-Review now has a transcript and caption service, with machine generated transcripts created for any recordings published to a section. Once published, the transcript can be edited and used to provide captions to the viewers. Before editing, the transcripts are approximately 93-95% accurate, so it is important to correct them before adding them as captions – but the transcript editor is very easy to use and the ability to find repeat errors and correct all instances of each error at a time, is a great timesaver.
This page covers:
- Accessing and editing the machine generated transcript
- Downloading and uploading the transcript
It assumes you are an instructor on a particular Q-Review section.
The first thing that you will need to do in order for your recording to receive the machine generated transcript is to publish the video to a Q-Review section. The Q-Review section is typically the area which the students access the recordings from – for example: ESH101 – Shakespeare 2019/20. Once published to the section, the transcript will be available approximately 1 hour later.
Once the transcript is available to you, you can follow the steps below to make the transcript verbatim and then publish the transcript as captions.
Accessing and editing the machine generated transcript
- Navigate to the video which you wish to edit the transcript for. To do so, navigate to your library and click on the • • • icon which appears when you hover the mouse over the recording, then click on the arrow showing ‘more actions’ and then select ‘edit transcript’, as shown below:
Alternatively, if you have already navigated to the recording which you wish to edit the transcript for, you can access the transcript editor via an option beneath the video element:
- Now you can play the video and the text in the transcript will be highlighted as the words are spoken. First click on the play icon to the left of the transcript… …then click on ‘edit transcript in the top right corner if you see any errors.
- One can see in the example below that words which have a lesser ‘confidence level’ of being correct are underlined and there is a search feature to see whether there are multiple instances of this error. In this Instance I can see that instead of ‘cybercrime’, it has found two errors of ‘sabah crime’ and ‘southern crime’. I can either choose to click upon the incorrect words and replace the error with the correct word, or I can click upon the magnifying glass icon to the left of the search field so I change all instances of the error in one go.
- Firstly type the error which you know has occurred somewhere, then select the magnifying glass icon, then type the replacement word & choose whether to correct one instance of the error, or replace all.
- Once you have corrected all the errors, you will see the ‘confidence’ figure at the top of the screen increase and you can then save the changes. Changes can be saved as different versions too. The option to save the version appears in the top right of the screenshot above.
- Now your recording’s transcript is verbatim, you can apply it to closed-captions, as shown below:
The closed captions can be configured so that they display in different colours, sizes, on different background colours and in different areas of the screen.
- To change where they appear, one needs to click on CC in the bottom right corner of the video:
- As you can see in the example below, the slide is grey and black so choosing a faint background for the captions might mean that they do not display clearly to the viewers. One can experiment with the contrast and location of them to ensure that they are presented best.
- Once you have positioned the captions to where they are best, one can turn on the captions via the CC slider shown above.
- If the transcript is replaced and you need to generate new captions or should you wish to remove the captions, then one can do so by clicking on ‘Update’ and then either on ‘delete’ or ‘reprocess’.
Downloading / Uploading the transcript
- To upload or download transcripts one can do so by clicking on ‘update’ against the transcript details and then by selecting the appropriate action that you wish to perform. If you are uploading a transcript then you will need to ensure that it is a VTT file type. There are options to download either the original machine generated transcript or the edited version. If you make any errors, then there is a quick and easy ‘revert’ option, as shown above.
For your information:
From experience, there are often only a few errors which appear in the machine generated transcript, but they usually are repeated several times. A recording which is an hour long and which contains clearly spoken content may have 5-6 words which are incorrect, but there might be 60-100 instances in total. Using the ‘find’ and ‘replace all’ features means that editing can be cut from 15 to 3-4 minutes.
The service operates on AI, so after editing your transcripts a few times you should see the system learn and the accuracy of the machine generated transcripts increase.
The captions are a great benefit to all students, not just those registered with DDS. Please add the captions wherever possible and ensure corrections are applied to the transcripts before they are published as captions
Did this answer your query? If not, you can raise a ticket on the online Helpdesk or email: firstname.lastname@example.org . Alternatively you can also request a particular guide or highlight an error in this guide using our guides request tracker.