Captioning Services

There are a number of captioning options for your content provided by UCL and a few approved external services.

Creating captions and transcripts for pre-recorded content

We can recommend automatic speech recognition (ASR) captioning and/or transcribing tools and services for prerecorded media that are either free or low-cost and produce fairly accurate results.

Paid options

The use of automated speech recognition (ASR) is now prevalent in the creation of captions, particularly where cost or urgency in production time means the use of a manual captioning process is considered uneconomic. This is generally the situation where there is a high volume or significant length in time of video content and especially relevant when delivering live broadcasts, where in essence, the use of ASR technology may enable captioning that previously would be unviable.

Generally speaking, the use of ASR systems that are underpinned by different AI systems these are constantly learning and improving, so it is thought that over a period of time their performance will become enhanced. Unfortuantely, there is much confusion as to the average level of accuracy whne creating captions using ASR, and it is probably less than 80% in most cases, therefore using this method you should always allow time to review and correct as necessary.

UCL has a number of suppliers who can offer captioning through the ASR process. UCL staff have tested the following products. You will generally need to sign up for an account and it is noted that most of the offers are for personal rather than business or educational accounts. 

Note 1: For captioning there is no UCL preferred service or software as every user will have different requirements. For translation and transcription there is a UCL Framework which can be found here... (UCL credentials required to login).

Note 2: Users who are short on time and/or who require a high level of accuracy should strongly consider paying for human created / corrected captions and transcripts. The process uses people rather than software, and 'paid for' captions and 'transcripts' are typically 99% accurate and can have quick turnaround times.

Note 3: UCL Educational Media offer a 100% corrected 'human created' captions service through a third party supplier for short form videos. Undertaking this can be costly but solves the issue where videos are embedded on websites and viewed by an external audience.

Product	URL	Comments
AI-Media	https://www.ai-media.tv	"Perfect the accessibility of your recorded content with captions for recorded media. Recorded captions come from our deeply established position in the industry, informed by our in-house practice guidelines and standards, and our deep roots in accessibility." Note: this is not a DIY service.
Verbit:	https://verbit.ai	UCL Education Media has access to Verbit and offers users a low cost "paid for" option within Mediacentral. This service delivers Automatic Speech Recognition captions, these will need correcting by the uploader, but are generally available within a couple of hours after the initial request  We are currently in discussion with Verbit to be make this available upon request with other applications. However for the time being, if you’d like to use Verbit, please contact us directly.

Free options using UCL provided systems

Free options make use of ASR  technology and although the technology is continually improving, ASR produced captions and transcripts will normally contain errors, and will need reviewing and editing by the user.

Be aware that for media files which have low audio or poor speech quality which is hard to detect, accuracy levels can drop significantly, meaning the user might have to heavily edit the files. It is reccommended that for each mimute of video you should allow at least 2 minutes to correct any errors.

Internal platforms / applications

Service	Information
Lecturecast:	Any media uploaded into UCL Lecturecast which meets a 90% confidence score threshold ( very good audio) will now have ASR-produced transcripts and captions applied automatically. Files can then be edited and downloaded using the built-in editor. See instructions here.
Microsoft Stream	Best for closed captions. Uploading a media file to Stream allows you to create, edit and download .SRT and .VTT files, which are the two most widely used closed caption file types. See instructions here  Note: if you need both captions and transcripts via Microsoft, we would recommend first creating the captions file via Microsoft Stream and then, once checked and edited, copying the contents into Microsoft Word to format into a transcript.
Microsoft Word Online:	Best for transcripts. All UCL users are now able to make unlimited usage of the in-built 'Transcribe' feature and edit the results directly in Word. Please note that this feature is available in the online version of Word only. See instructions here.

External sites (YouTube, etc.)

If your media file is already hosted on an external site, it is worth checking to see what captioning and transcribing options the site might offer. For example, YouTube allows users to generate ASR-produced captions for any video, which can then be edited using YouTube Studio. See instructions here.

Captioning for Live Events

Please note UCL Educational Media do not caption live events, if you require this kind of service then staff across UCL have found that commissioning the services of an external provider trialed by the Digital Accessibility team is very successful.

Paid options

Supplier	URL	Details
121 Captions	https://www.121captions.com	121 Captions can be found in MyFinance, please contact them directly or discuss with the ISD Digital Accessibility team. Also ensure you give plenty of notice to ensure they have availability on the day of your event and request a quotation. For further information please see their website.  121 Captions provide live captioning service who are expert providers of verbatim realtime captioning onsite and remotely to your tablet or laptop. This uses a stenographer service which will capture every word as it is spoken with 99% accuracy. This makes events more accessible to those with hearing loss or second language speakers with a high-quality service. Users can get a transcription of the event. The company clients include conference organisers and students with hearing loss and they caption events all around the world in different time zones. All that is needed is an internet and audio connection to you.

Free options for online events using MS Teams or Zoom

Service / Application	Information
MS Teams	Allows users to produce live captions during meetings and generates a post-meeting transcript. See instructions here. If recorded, a video of the meeting will then be saved to Microsoft Stream. So if you need to edit the captions afterwards, see the instructions above for Microsoft Stream.
Zoom	Allows users to produce live captions during meetings, either manually or automatically. Post-meeting transcripts are available for users with the appropriate license. See instructions here. If recorded, a video of the meeting will then be saved to Lecturecast. So if you need to edit the captions afterwards, see the instructions above for Lecturecast

Human created captions and transcripts

There are various options for delivering 99% correct captions, these are generally produced by using a third party supplier who employ staff to create the caption files. 

The UCL Educational Media ( part of Digital Education) can offer a charged-for bespoke captioning service. These captions are produced via human transcription and are 99.5% correct, therefore more accurate than using automatic speech recognition software. Although at a cost, this will save staff (or students) time rather than having to do the work themselves.

Available services through UCL Educational Media

We primarily offer the following services:

• Captioning of video content
• Dubbing and Voiceover either in English or different languages
• Interlingual Subtitling - Subtitles for the Deaf and the Hard-of-Hearing (SDH)
• Audio Description for the Blind and Partially Sighted (AD)

Contact the team via email for further information and costings.

 External services for creating (almost) perfect captions

External suppliers on MyFinance

Product	URL	Comments
AI-Media	https://www.ai-	"Perfect the accessibility of your recorded content with captions for recorded media. Recorded captions come from our deeply established position in the industry, informed by our in-house practice guidelines and standards, and our deep roots in accessibility." Please note that this is not a DIY service.
Verbit:	https://verbit.ai	UCL has partnered with Verbit to offer users a "paid for" option within Mediacentral, this service offers the option of 'Human Created' which are 99% accurate or ASR captions which will need editing. We are woring to provide Verbit services in other applications upon request. However for the time being, if you’d like to trial Verbit, please contact us directly.