Captioning – Voices in Text

Philip Wels (Mankato, Minnesota, USA)

Archived discussion

About the presenter

Philip Wels is a Video Production Professional who specializes in video editing and streaming/recording systems for churches. In addition to Peace Devotions, Philip co-produces Canaan Bound Podcast and helps manage the social media accounts for various groups. When not working, he enjoys reading, watching films, or working on his many creative side projects.

[The video version of this article contains a demonstration of the various options that the author describes.]

Closed captions. Open captions. What are they? How do you make them? Do they work with live video? And should you use them?

The video tech surrounding closed captioning has made some major leaps in the last year and with so many things moving online it's only going to continue to grow and change. For the past three years I've been both closed captioning and open captioning Peace Devotion videos and more recently I've started examining technology for captioning live video, specifically, church services.

There are many things to consider and several methods for adding text to the screen. This short article does not exhaust the possibilities, but I hope it will offer some advice and direction for those looking to caption their videos.

First, we should define a few terms.

Closed Captions vs Open Captions. The two main types of captioning are Closed Captions and Open Captions. Closed Captions can be toggled on and off by the viewer¬ while Open Captions are always visible. The version of Peace Devotions which gets posted to Facebook has Open Captions which are "burned in" to the video file. They are part of the video and can't be turned off. However, when uploading Peace Devotions to YouTube I upload a separate .srt file (SubRip Subtitle file) which allows users to toggle the captions on and off.

Closed Captioning typically resides in a separate file from the video itself and is easily editable if changes or corrections need to be made. But if captions are burned into the video (Open Captioning), it's impossible to correct without video editing software. And then you will need to re-export the whole video. That said, with Open Captions you have more control over how captions will look to your audience regarding font styling, color, etc.

Methods for Creating Captions

Method 1: Automatically

Let's start with what can currently happen automatically to your videos. If you're streaming to Facebook or YouTube, or if you are uploading videos to these sites, your videos are likely getting automatic captions added to them as part of the processing.

Facebook and YouTube both have custom user interfaces for reviewing or editing automatic captions. The interfaces vary in usability and finding your way through Facebook's inconstant and everchanging set of creator tools can be annoying, but the tools exist. They allow you to retime phrases and correct the wording and spelling. If you're producing and uploading videos ahead of service time, this might be enough for your needs.

Method 2: Make them yourself (in advance/after the fact)

If you are pre-recording your sermons or entire services you have the option of captioning them ahead of time. This is perhaps one of the most reliable ways to caption videos, but it is also one of the more time-consuming methods. Video editors like Apple's FCPX and Adobe Premiere will allow you to type up captions and time them out as needed but doing captions by hand this way is one of the most tedious, annoying, and boring things you could consider doing with your time. The tools are a little clunky.

One alternative to doing it completely by yourself is to have a machine do most of the work and make corrections as necessary. When captioning Peace Devotions I take the audio from each devotion and run it through a program called SpeedScriber, a service which provides machine transcripts and subtitles for 50¢ per minute and the Mac App provides an interface for making corrections to the text. I can then export as a .srt and .txt file. For my Peace Devotions workflow, I pull the SubRip file into Premiere and use the captioning tool Premiere offers to modify the timing and double check the spelling and grammar during my final editing pass. This also allows me to embed open captions on my export for Facebook. And I usually delete the captions that duplicate the Bible passages I show on screen.

The benefit to this method is that most of the timing is taken care of for me, because timing the words is really the most difficult part. YouTube has a feature where you can upload a transcript file and they will automatically time the words out for you. But that requires a transcript and you still may need to modify the timing at certain sections.

REV.com is another service which can help provide closed captions for your video. They have a number of options to choose from; all of which must be purchased. You can get a machine transcript for 25¢ per minute or have a human type up the text (along with timing) for $1.25 per minute of audio.

Telestream has software specifically made to Close Caption videos. The desktop edition of the software costs $1,898. It looks like it has a lot of great features, but automatic transcriptions require an additional subscription of $99 per month.

All of these options share this downside: you need to have the video done well in advance. Services like REV are fast, but their turnaround guarantee is 24 hours, which means a day of lead time is required in a worst-case scenario. Also with these options, you'll need a budget and likely a lot of time.

Method 3: Doing it Live.

Several ways are available to do Closed Captions live. I'll describe them and then consider the possible problems and downsides to live captioning.

Option A:

If you stream to Facebook or YouTube you might already have the option for automatic live captioning. On your churches' Facebook page click on "Live" which should take you to a page where you can begin streaming and change settings for your stream. If you scroll down the page, there should be a "Settings" box with options for Stream, Viewing, and Comments. If you click on Viewing you should have the option to "Turn on auto-generated captions." If you enable this, Facebook should automatically generate captions for all of your live videos. This works whether you stream from Facebook's web interface or through another program.

YouTube is currently rolling out a similar feature, but at first they are offering it only to channels that have over 10k subscribers. Eventually this will likely be available to everyone, but for now it's a limited option.

The benefit to using this option is that it's built in and happens automatically. The downside is that not everyone has access to these features or uses these services.

Option B:

There are several free web services (or browser extensions) which convert speech to text and display it on the screen. One easy way to add Open Captions to your video is to add a webpage screen capture as a video source and layer it over your video. PubNub has a free service they offer to twitch streamers that uses this method. It works well with OBS (Open Broadcast Software) but should also work with other software that support browser or screen capture sources.

The benefit to this option is that it is free, fairly easy to use, and highly customizable. The downside is that the words are part of the video and any mistakes are there to stay. In my limited testing PubNub's system also had a lot more mistakes than other services.

Option C:

Another free web-based captions option is Web Captioner. This service can actually send a closed caption stream to OBS, YouTube, Zoom, and other services. Web captioner claims it supports over 40 languages and has decent documentation for the services that are integrated. It will also let you export transcripts, which could be posted to a website or Sermon Blog. (Although I'd recommend doing a once-over to correct mistakes before uploading.)

The benefits to this option are the integrations with other systems and honestly, I've been impressed with the quality. Web Captioner does a much better job than some of the other systems I've seen.

Caveats and Considerations

Not all captioning services are created equal. In my limited testing, Web Captioner and Facebook seemed to outperform PubNub. And running my script through Speed Scriber (at the cost of a few dollars) gave me pretty good results as well.

But technology is still a barrier in some instances. If you're streaming from OBS, vMix, or using Zoom, it's possible to add closed captioning using something like Web Captioner. But not every piece of software supports captioning in this way. Livestream Studio now owned by Vimeo, supports closed captions, but only when they are embedded in SDI. This would mean having a separate captioning system and a captioning embedder, which could cost several thousand dollars. My church would be better off using Facebook's live captioning option.

I think it's important to consider carefully whether to use automatic subtitles. For religious content, machine translations are typically discouraged. When a person's soul is at stake, getting the correct wording and conveying the correct message is vital. In many, if not most cases discussed here, a machine is translating spoken text into the written word. The results are not perfect. There may be missed words, or the text words may not actually reflect what was spoken. This must be taken into consideration.

A less detrimental effect of incorrect words is that viewers may find them distracting and they may pull attention away from what is being discussed.

For these reasons, I highly recommend using options where captions are "closed" rather than "open" unless you are able to take time to proof them. It may not be worth the time, energy, or effort to go back and correct the automatically generated captions for live videos. But you should know they exist. Perhaps it's good to let the audience know when the captions are automatically generated.

Takeaways & Questions

Adding words to the screen does increase engagement. It helps with accessibility. It helps people who have the volume turned off. It is certainly worth considering, especially for pre-produced programming.

As captioning keeps getting easier, it is worth keeping an eye on the technology. There's a vast array of options out there and with more content moving online, this is an industry which is likely going to grow and expand in the years ahead.

Have you done anything with closed captioning? Do you have tools you like? (Or recommend staying away from?) I'd love to hear your experiences. And if you have questions I didn't answer, please ask! I'll see you in the comment section.


Translate this page into your language
Return to original language with "show original" button at top left.

Discussion

-
David Norris 2020-10-19 10:44:39pm
This was great Phil! I've been using Rev.com in post production for video with great success, but it's really helpful to know there are some other services out there—especially for live captioning. It was great to see how well the different services you covered did when you were speaking with such articulation. I'm sure the tech will continue to improve.
-
John Roebke (One Africa Team) 2020-10-20 9:37:19am
Thanks for the tip on using PubNub to generate subtitles for live streaming. It is helpful for non-native English speakers to view the English text simultaneously while listening to a presentation
-
Philip 2020-11-07 12:37:23pm
Good point John, having text on screen would be a good feature for any church with an ESL program or a multi-lingual congregation. There's been a few times when I've wished we could translate Peace Devotions and provide captioning in other languages but haven't had the chance to do that yet.
-
Tom Kuster (Christ in Media Institute) 2020-10-24 7:15:27pm
Phil, I can understand your reluctance to recommend live captioning for religious messages like sermons, because of the imperfections at this stage of the technology. As a "senior" I find myself relying more and more on captioning with any screen viewing. I would not consider live-captioning of a sermon to be a distraction, but a means of getting every word the pastor speaks, even though he sometimes says things too fast or too softly. Do you anticipate improvement in the technology to a point where you would recommend live captioning of sermons? If so, how will we know when that time has arrived?
-
Philip Wels 2020-10-25 6:02:33pm
Hey Tom,
I do think the current state of technology can be utilized to offer some good accessibility enhancements. I think I'd actually give it a limited recommendation now, or at least suggest that it be considered. My main caution is just that it's not perfect and both producers and users need to keep that in mind. (Most the time it seems to do pretty well, though this morning I was testing out Web Captioner for a bit and the word "Zombie" popped up during one of the scripture readings.) It's those types of errors which can be distracting and as those become less, it'll be easier to recommend.

The bigger issues are when little words like "not" are accidentally omitted. As long as people are listening and the captioning is used for assistance, I think this is a "safe" technology to use. But I don't think right now you could watch with the sound off and get the same level of detail. And if I had to make up a benchmark, I would make it that: Can you get an equivalent understanding from reading the captioning as you could by listening. (And right now we're actually pretty close.)

Of course, as you say, sometimes the pastor speaks too fast, or too quiet, and words, ideas, meaning are lost. As sinful people, there will always be *something* lost in translation. (This brings back memories of classes in Speech and Communication at Bethany.)

But I also believe this technology is going to continue to improve. This is a useful technology for SO many industries. Dictation software is a big market. Smart phones, smart speakers, voice assistants all need to properly and correctly understand human speech for commands. Twitch, YouTube, Facebook, and other online video platforms would love to offer these assistive technologies. Audio and video companies are continually working to eliminate background noise and isolate vocals for recording. All the different pieces are improving and coming together. I'm not sure it'll ever be "perfect" but neither is our human hearing.
-
Mike Hartman (WELS World Missions - Academia Cristo) 2020-10-28 11:56:24pm
Hi Phil, Thanks for sharing this. Very interesting stuff. We do use captioning on some of our videos. I completely agree with your comment that it increases engagement, especially on many social media platforms. Many videos on Facebook are read first and then heard. In my opinion, audio is secondary for videos posted on Facebook.

Another benefit of captioning is that it can help people understand a speaker. Here is a recent video we created. (https://vimeo.com/472512751) The speaker is a non-native English speaker. We added captioning to help the viewer understand what was said:

I have a question for you. Allow me to explain the situation. Then I'll close with my question.

Two weeks ago, I attended a number of sessions of the Zoomtopia online conference. One thing that struct me is how they set up that conference so that a global audience could participate. As an attendee I could chose what language I wanted to hear the presentation. The presenter spoke in English, but I could choose to hear a person give a live running translation in German, Spanish, French, Chinese, Japanese, among other options.

As a person who serves in world missions, this is a very intriguing. The CELC (celc.info) provides a forum for Lutherans around the world. When you gather people from around the world, language is a big issue.

Are you aware of any program that would allow someone to do either live captioning in a different language from the speaker's language or allow a person to actually play a live translation over the presenters audio as they did in the Zoomtopia conference?

Thanks!


-
Philip W 2020-11-01 2:27:21pm
Hi Mike, Thanks for sharing!

I know Apple just released a translation app as part of iOS 14, which is supposed to allow people to speak into it and hear translations in other languages, but I haven't had a chance to try it out yet. Translation apps continue to improve. I know Google is also working to improve their translation software.

As for live translations, I'm not aware of any technology that is able to do it live. (Though it might exist.) I can say that this is the "dream" solution. I think that is the type of "futuristic tech" that many organizations are working towards. I wouldn't be surprised if we start seeing these types of solutions start showing up over the next several years.
-
James Aderman 2020-11-01 11:31:26pm
Thanks for sharing some of the wonderful things you do to share the gospel online, Philip. Give some thought to letting the WELS Intersections community know about this.
-
Jeff Lemke (Bethany Lutheran College and MLC) 2020-11-03 10:32:00pm
Phil, This is a great topic, yes, I also love how it increases engagement throughout the social channels and I've watched how some of the live caption ones improve as they "get use to" the content that is being covered. Great stuff, thanks for the article.