A client recently asked me to add subtitles to their webinar archive recording in order to make it accessible for the hearing impaired. I thought I would share the steps involved, along with a marvelous service I used.
The first thing you should know is that you probably won’t be able to add subtitles to your recording if it is in your web conferencing provider’s proprietary format and it lives on their servers. You will need to convert it to a standard video format such as WMV or MP4 or FLV and download it to a place where you can work with the file. (Vendors: If you allow subtitling of webinar recordings on your servers, please add a comment to this post to let us know about it!)
You will need a full electronic file transcript of all narration in your webinar. Don’t worry about formatting it at this point. You just need all the spoken words in text format.
You’ll need to rip the audio track from your video to make it easier to work with. Lots of software applications can do this step. The audio quality is not particularly important, so feel free to use a lower bitrate to reduce the file size of your MP3 audio file.
If you weren’t working from a prepared script during your webinar, you’ll probably use a transcription service to prepare the text transcript file from your audio recording. There are many service providers who offer transcription. Note that if your subject is highly technical, with buzzwords and abbreviations, you will need to review the transcript to clean up things the transcriber doesn’t understand. I don’t have a recommendation for a transcription provider. Just search on “audio transcription” or “transcription services.” You will be surprised how inexpensive audio transcription is nowadays.
Once you have the text file ready, you can move on to subtitling. The most common format for subtitles is .SRT (known as “SubRip” in reference to a software program that reads them). A SubRip file is a simple text file that breaks every phrase into the specific words that appear on screen along with time codes for when they appear and disappear. The time codes are written out to a thousandth of a second.
Writing out time codes is one of the most mind-numbing tasks imaginable, as you keep playing, pausing, and checking times for each few words. This is where I found an amazing service provider. CaptionSync from Automatic Sync Technologies has an automated service that creates a .SRT file from your transcript. You upload the text file, upload the MP3 recording, and it sends back a completely coded .SRT file, automatically broken into time codes that match the audio.
I can vouch from my experience that the algorithms they use are impressive. We had five different speakers on our recording, both male and female. One speaker had a pronounced accent and spoke quickly. One speaker went significantly off script during a paragraph. But the software didn’t give up. It did context matching to keep things flowing at the proper rate to get the words on the screen until it found the next exact match. Really impressive! And the cost was ridiculously low.
I found out after the fact that I could have added some special text codes to my transcript file that would force breaks at designated positions or allow me to add nonverbal subtitle notations such as the speaker’s name. But these edits are fairly easy to make after the fact, since you are playing with a basic text file.
I did find it easier to make my edits using some additional free software. There are a variety of SRT editors available, and you can search for them online. I downloaded Subtitle Workshop from URUWorks. Once I figured out how to use the editing commands efficiently, I was very impressed with the way it let me recombine and resplit subtitle phrases without messing up the time codes. It’s not immediately intuitive, so read the documentation. But the power is great.
And I was able to test how well the assembled video with subtitles looked by using the free online site dotSUB.com. It lets you upload a .SRT file and a video, and play the video with the subtitles overlaid. You can set markers and pause playback if you need to adjust time codes. It gives you the timing down to that thousandth of a second. I don’t like it as an editor, but it’s good for testing and checking times.
Once you have your .SRT file tested and coded to your satisfaction, you use your favorite video editing software to combine them. Most of the major packages on the market give you the ability to create a subtitled video output file. I use Camtasia from TechSmith, which works fine if you are careful about how line breaks are specified in your SubRip file (don’t save it as plain text without hard line breaks or split lines will run together on the video).
If all the post-processing and video production is beyond your means, you can always just pay CaptionSync to produce your captioned video for you. I didn’t use that service so I can’t vouch for it.