How to generate subtitles or captions for video files using Microsoft Stream

This brief tutorial will provide a general overview on how to generate .vtt subtitle files from digital video files, such as .mp4, using Microsoft Stream. It does not not cover the Microsoft Stream application in detail, the .vtt specification, or the trouble shooting video codecs.

You might have found that Pomona College now has the Microsoft 365 Office Suite available . There are many new tools available and one of those tools is a video streaming application called Microsoft Stream. While Streams main function is to share, store, and deliver video content it also has the ability to automatically generate subtitles in the .vtt format. Explained another way, you are able to upload a video file, such as an .mp4, and Stream will use voice recognition technologies to output a text transcript of the video. That’s quite impressive!

Your output will be a .vtt file which is a specially formatted file that will have lines of text, the spoken words, along with the time that it takes place. Please see below for a link which details the .vtt format.

Now let’s do this, follow the steps below! Following the steps  will be a discussion about the realities of this technology and some tips, tricks, tools and list of links.

Stream on Menu
Stream on Menu

First,  find the Microsoft Stream application.  You might find it in the drop down menu that is in your Office 365 web Outlook. You can also try this url: https://www.office.com/ . You may have to “Search all of your apps”.

Second, When you are in Stream you will see “Upload” in a few places.  To upload a video, drag and drop it. When you drag and drop you will see some settings that are available, make sure that the “Autogenerate a caption file” has been checked.

Screen snip of options.
Screen snip of options.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Third, the file will usually upload rather quickly. However, autogenerating the .vtt file can take some time. Here is our experience, sometimes it is 5 minutes and sometimes it is 90 minutes for a 120 minute film. No reason, that is just the way it is, so act and plan accordingly.

Next,  so you want to check to see if your file is done? To find out goto a listing of your videos and click on the “pencil” icon, that will bing up a page that will allow download of the .vtt subtitle file.

File List for MS Stream
Click to enlarge: File List for MS Stream

Finally, lets download the file. Just save it like you would any file. Or, you can just leave it there and it can be used with Microsoft Stream.

Download File
Click to enlarge.

Realities of this technology, tips and tricks, and some useful tools.

Some Expectations

Auto-captioning is not perfect and you should not expect it to be. Some video files work better than others. Here are some things that we have found not to caption well in English.

Musicals do not auto-caption well. Bollywood movies that are in English (or dubbed) do not caption well due to the unique accents. If you have an environment where there is a lot of background noise, like a big noisy crowd at a sporting event, this noise will often drown out the voices that need to be captioned. Finally, any file or media that has poor audio quality does not do well. Overall, a captioning success rate of 90% is generally tolerable. If you need 100% there is a way to modify the .vtt and we will cover that soon.

As of the writing of this article Microsoft Stream also can do Spanish. That means, Spanish audio in then Spanish .vtt text out.  Here at Instructional Technology we have not experimented with the Spanish language captioning. There is no option for translation from English to Spanish or Spanish to English, that’s anther tool. We also don’t think language support would dynamically switch. If it does let us know.

What Video Formats Work?

We have found that most everything works and we did chuck some problematic formats at it like FLV and AVIs with  H263 codecs. It all worked. If you have something that Stream balks at just transcode it to another format using a  tool like Handbrake (see link below) to transform the video to a more common format.

What do I do with the .vtt file and how can I edit it?

You can use Stream to show your videos and Stream will use the .vtt file and show captions in their player. You can search and edit the transcript within Stream itself. Then download the improved file. You can fix those errors and make the transcript 100% accurate.

Screen shot of Stream online edit of transcript.
Example of the Stream online edit of transcript.

Additionally you can  also download the .vtt file and use it somewhere else like on another web site, as a text transcript, play on the VLC player (see link below), or to add subtitles to a DVD. You will notice that the .vtt file is not easily opened with notepad or a general text editor. To edit it you will want to use a tool like Jubbler (see link below) which will allow editing. You can also use Jubbler to translate the file to other formats. Finally, please note that the textual output that is generated from video files will often take a good amount of manipulation for the text to look like it is from a book or that it reads like one.

Some Links to Software Mentioned Above

Jubbler Subtitle Editor: http://jubler.org/

Handbrake Video Transcoder: https://handbrake.fr/

VLC: https://www.videolan.org/vlc/

WebVTT:  https://en.wikipedia.org/wiki/WebVTT

Introducing the LUNAVersal Viewer

Over the summer of 2018 the LUNA Imaging suite was upgraded to version with much new functionality. Perhaps the  most notable addition is the LUNAVersal viewer. In short, this new viewer is a “mobile”  viewer.

LUNA Imaging says:  Our new responsive LUNAversal Viewer provides an alternative option to the LUNA Viewer.  While theLUNA Viewer is full-featured and ideal for researchers, the LUNAversal Viewer is responsive and ideally suited to sharing access to collections on mobile devices. “

Responsive – UI built for expanded delivery to audiences on mobile platforms
Aims to meet WCAG 2.0 Level A web accessibility standards and includes standard keyboard navigation
Utilizes IIIF technology for image delivery
Open code provides an opportunity to further customize the LUNAversal Viewer
Customizable font and background color
Supports Google Analytics

Search and Display Features included:  Collection selection, Media Group viewing, Facets, Related Media, Keyword Search, Thumbnail size selection,
Image zooming, thumbnail navigation and rotation, Data Display,
Browse media results by paging ( desktop), swiping (mobile) and
Share and Embed (includes URL to IIIF mainfest).

How do I take the LUNAVersal Viewer for a test drive?

There are two ways to test it out.

If you are using a traditional computing device then simply us this URL: http://luna.pomona.edu/ll/index.html  and your browser will be redirected to the LUNAVersal viewer.

For phones or tablets,  simply point your mobile devices browser to: http://luna.pomona.edu and LUNA will magically default to the new viewer.

 

 

New LUNA Collection: Pomona Theatre Garment Collection

A new collection has become available on the LUNA imaging system. The collection is titled the “Pomona Theatre Garment Collection” and can be viewed at  the following URL from any on campus IP address at the 7c’s without authentication : http://luna.pomona.edu/luna/servlet/POMONA~7~7

The collection features detailed photographs of select  historic Theatre garments from the Theatre Department’s collection. Most of the items in this collection are well over 100 years old. The photography was taken over the summers of 2017 and 2018. This faculty project was funded with a grant from  the Hahn Teaching with Technology program and was under the direction of Sherry Linnell, Resident Designer and Professor of Theatre.

Credit is to be give to the following individuals: Photography and Metadata 2017 & 2018 – Michael Mao, Theatre Student; Photography 2017 – Eve Wang, Theatre Student;  Photography, Post editing, and LUNA Management 2017 & 2018 – Jason Smith, Instructional Technologist.

 

Video47 : Server Upgrades & Fun Facts

Video47 is a poplar video hosting tool many faculty and students have come to love over the years. It is a tool that can be found within all Pomona College hosted Sakai courses and used by Pomona Faculty to create playlists from a centralized video repository. All viewers (typically students or community auditors) need to do is be members of the Sakai course to view content.

Over the Sumer of 2018 a new more robust machine was created for the service. Installed was the most up-to date and  fresh operating system. Disk space for the media was made larger and optimized. The streaming engine was updated and some new automation and monitoring tools have been initialized. Finally, all the media and data was synchronized to the new machine.

Now for some fun facts about Video47:

  • Video47 made its debut in the Fall of 2009. It’s 9 years old!
  • The average number of Pomona courses that use Video47 playlists are now over 100 per semester.
  • The video repository contains over 4000 titles.
  • The Video47 Unicorn is named Rainbow.

 

SAX Server Updated: Summer 2018

This summer the server known as sax.pomona.edu was rebuilt and rejuvenated. The content and applications from the old server were moved to the new install. Improvements were made to the disk partitions so that continued media growth will be able to continue without issue. The most current OS available has been installed and secured.

What exactly is SAX and what does it host? Sax hosts a variety of media items for instruction as well as custom applications built by the Instructional Technology Group.

 

Two New Scripps College LUNA Collections

There are two new collections within LUNA that may be viewed from  any on campus IP address or via VPN. These collections are: Scripps Asian Art collection  and the Scripps Gardens collection.  

One might ask: Why does Scripps College have collections on the LUNA server? This is because LUNA has been a collaborative initiative between  Pomona College and Scripps College which dates all the way back to 2005 when the first image collections were created.

Finally, you may learn more about what LUNA Software is by visiting the LUNA Imaging Software  website.