In a previous post, I talked about the importance of adding captions to the videos that you create, in order for them to be accessible to deaf and hard of hearing individuals.
In that post, I mentioned that YouTube has come up with a new automatic captioning program by which it can automatically add captions to your video by transcribing the audio into text.
Here’s the problem – the Auto-Captioning program leaves something to be desired. While people might say that it is “better than nothing,” I like to think I deserve to be able to watch a video without scratching my head and trying to decipher incorrectly transcribed words.
please, Please, PLEASE folks! Don’t just settle for clicking the auto-captioning feature and leave it at that. Watch your video afterwards with the captioning. Determine for yourself if it’s truly satisfactory. The odds are you’re going to find that it has a couple of errors…especially if you are using specialized vocabulary, a lot of names, or if the sound isn’t totally clear.
You might think that it’s “good enough.” But should anyone have to settle for just “good enough” when there is a way to make it even better?
I strongly recommend that anyone who wants to add captioning to their YouTube video please consider downloading an actual transcript of the audio track on your video, and then request Auto Timing when downloading their video to YouTube. This will allow YouTube to use your transcript as the basis for captioning your video – the auto timing feature will allow for the words of the transcript to be inserted into the video at the proper time based on the audio the program “hears.”
All you have to do is listen to your video and type up a simple transcript of what is said. You do not need to include time codes or frame numbers. You are not manually inserting captions into the video frame by frame. YouTube’s automatic timing program takes care of all of this for you. Just type up the lines of the videos – similar to typing up a simple script for a play or the like. Save this transcript to a file (such as a Word document) and then download the transcript file to YouTube’s auto timing captioning program. YouTube will do all the rest – inserting the words from your transcript document into the video at the proper time, utilizing proper vocabulary that is spelled correctly.
It’s really not that difficult folks…nor is it that time consuming – especially since you’re not manually inserting or typing up the captions frame by frame.
And it will make a world of difference to people like myself. It will open up so many more opportunities for us to when it comes to internet videos.
Why not give it a try?
Need more information on how to do this? Just watch this video!
PLEASE NOTE: There are many different programs which can create captions and text for videos, podcasts, etc. This just happens to be one of them. In my next post (see the link up at the top of this page) I discuss another captioning option that is pretty quick and easy. If you know of others that you’ve used and liked, drop me a line and let me know…and I will pass the info on to my readership. Thanks!
This sculpture shows the sign for "connect"...and that is what Deaf Pagan Crossroads is all about - making connections. Connections between Deafhood and Paganism, connections between the Deaf Community and the Hearing Community, connections between myself as the writer and you as the reader. I hope you will take the time to read my various posts, some of which are listed below. Welcome to the Crossroads, and I hope you make some connections here!
[…] at Deaf Pagan Blogs has a new post about captioning video podcasts for our the deaf and hard-of-Hearing Pagan community. As more and […]
You know, one of the fastest ways I’ve found to get across to people the problem with auto captioning is to show them some really bad auto captioning examples.
We watch TV with captions on a lot so we can keep the sound turned down to hear other things (like ventilator alarms) and because of a good friend who has some hearing loss, and when we’re out at restaurants, watching sports commentary channels with their instant captions is horribly funny….until you realize that people who actually need those captions are going to be at a loss as to what was actually said.
I strongly disagree with your statement “nor is it that time consuming.” Agreed, it is not difficult, but it most certainly is time consuming.
Right now we just do show notes for our show highlighting topics and when they were discussed in an audio show. This takes approximately 1.5 times the length of the show, and this is just topics. Luckily, we can multi-task so this time commitment isn’t so great – but when it comes to captioning, there are no shortcuts that I am aware of (please share shortcuts you are aware of – for example if we could use the low-quality automatic captions as a starting point).
The videos we post to YouTube nowadays are typically 30 minutes to 4 hours. I am sympathetic to those that are hard of hearing – and more practically speaking, I realize those that have no hearing problems would likely make use of captioning (like many of us already do for TV) regardless and thus allow us to reach a larger audience.
I’ll be more mindful of this for short videos we do, I know we’re planning a very short <20 minute video about a ritual soon. Though any tips and tricks you have for making this easy, let me know.
I plan to embed captions directly in the video, so those on Archive.org can also have the benefit of captioning.
Yes, I am aware of the services in India that can do transcriptions. From experience in my professional podcast ventures, trust me – they still have a long way to go.
I guess it’s all relative.
I have been deaf all of my life…which means 50+ years. So to me, taking one day to type up a transcript and download it to a podcast so it can be accessible to me and other deaf/hard of hearing individuals just doesn’t seem all that time-consuming. I have had to go for most of my life without having Equal Communication Access – and I still don’t have full access in many areas of my life.
Yes, it is getting better…modern technology is bringing us new opportunities that we never had before.
But all the technology in the world is not going to do much good until we can change people’s attitudes and encourage them to want to invest the time and effort into creating greater accessibility.
I once added text to a short video I made using the iMovie program on my MacBook. I acknowledge that it did take some time, but it wasn’t too terribly time consuming, and I felt it was time well-spent.
I do appreciate that when you are making lengthy videos, it becomes more complicated, and requires more time. This is one of the reasons I suggest using YouTube’s program – you don’t have to time code your captions yourself…YouTube does it for you.
If you already have a script of your video that you are working from anyway when you create it…you’re already two-thirds of the way there. Just take that script and download it to YouTube’s auto timing. Then all you will need to do is polish it up as necessary.
When you’re developing unscripted videos such as interviews and the like, then yes…it may take more time. What I would suggest you do is find those folks with both sharp hearing and fast typing skills, and have them sit down and listen to the podcast and just start typing away. Then go back and clean up any typos, etc.
Again…yes, it takes time. And yes, I have heard the argument about how time equals money. And when you’ve got limited resources of both, you might not be all too enthusiastic about jumping on the captioning bandwagon.
But I am going to be brutally honest here – most Deaf people are not going to be very sympathetic to either argument. Remember – we have to live with the alternative…24 hours a day, 365 days a year.
There are still many of us (including myself) who can remember the days when there was NO captioning on television. There are many of us who can remember when we had to pay to buy a special decoder box to hook up to our televisions so we could watch what limited captioning was available. There are many of us who can remember spending hours advocating for Congress to pass the Television Decoder Act, the Americans with Disabilities Act, the Telecommunications Act, the Twenty-First Century Communications and Video Accessibility Act.
Many of us have had to spend extra dollars to purchase equipment that allows us to do the things that many people take for granted. I’m not just talking about expensive hearing aids and the batteries they require (and by the way, most insurance companies will NOT pay for hearing aids.)
I’m talking about visual alert systems for the door, smoke detector, baby cry. Expensive cell phones (and calling packages) to allow us to receive emails, text messages, video calls. High speed cable access to allow for the running of our specialized telecommunication systems. (Granted, many hearing people pay for many of these services also – but for these folks they are an optional upgrade…for us they are a necessity.)
Being deaf is a 24/7 job…and it doesn’t come cheap. I’m not complaining – I’m quite happy with being a Deaf person. But this is the reality of the situation.
So if we in the Deaf Community come across as seeming “unsympathetically demanding”…please try to look at it from our point of view.
Equal Communication Access isn’t just about wishful thinking… it’s about our lives.
The auto captioning feature SUCKS!
I have found a new software that works pretty well — it is fairly intuitive in its interface. Videocritter.org. They have a demo video that shows how to do it. I’d recommend anyone wanting to caption to use that.
Thanks Don! I appreciate your sharing this info. I will have to check it out.
I agree…the auto captioning feature leaves A LOT to be desired!
Dave of PCP, the lack of captioning in your videos and Podcast is what keeps me from following them. You’ve had a great number of topics that I’m interested in, but as the majority of the time, my computer access is during hours when I can’t have sound on (and can’t use headphones because I need to be able to hear other sound), I miss out on them.
Yes, it is extra work. I guess it comes down to whether or not you think the product is worth it.
http://www.google.com/support/youtube/bin/answer.py?answer=100077
Dave,
You must not have looked very hard, or you don’t know how to use Google, but if you own the video on YouTube, you CAN download the auto captions and then fix them. So there. What’s your next excuse?
Don’s feedback was quite useful. VideoCritter makes captioning a painless process thanks to the keyboard shortcuts. In the past 10 minutes, I’ve already captioned a video using VideoCritter:
No manual inputs of timestamps (or worse, frame numbers) along the painful lines of what Pam was recommending. Just hit enter after a phrase is said, type what was said, press Enter to save, and Enter again to continue playing. It’s actually pretty darn good at not messing up timing too much when I speak quickly.
Now I just need to work down my technique in writing in actual sentences and being more grammatically correct.
As for audio-only podcasts, we will never transcribe the Pagan Centered Podcast itself in a way that is easily available online (though offline printing in non-eBook format is something we may consider due to the difficulties in machine searching through that even after they are pirated into electronic format). We made that decision many years ago. However, we also refuse to stop anyone from transcribing them due to our choice to release under Creative Commons. All of us regular hosts have decided to just not make the raw contents of our often controversial episodes searchable and able to be easily quoted out of context by automated bots doing content scraping.
Though for our other shows (Pagan People, Pagan Comedy Minute etc.), we’ll keep those options open. This is our hobby, we prefer not to devote too many resources to items whose benefit do not outweigh their resource consumption. When we started bleeping naughty words, we nearly doubled our download counts. We have folks on the show with vocal impairments, but their contribution of content is powerful enough to be worth the time to do the editing needed to make it easier to listen to what they are saying. However, these are things that take a few extra minutes of post-production time per recording. Transcriptions is something that will take extra hours per recording (a recording is typically 3 to 4 hours, trimmed to 1 or 2 one-hour-long episodes after post-production), so I have a hard time justifying that effort right now for a hobby. My doctor would much prefer I go outside and get some exercise with that time and I agree that I need to spend less time on my sedentary hobbies.
As for Herb being tied to the computer, I recommend going to an electronics discounter like MicroCenter, Fry’s Electronics, CompUSA or something like that and get a low-end MP3 player, they often run around $10. If you don’t have one of those stores, Wal-Mart often has something in the $20 range. It’s not flashy, doesn’t have cool blue lights but it’ll play mp3s through a headphone or $5 pair of “computer speakers”. I’d recommend burning to plain audio CDs and using a CD player either hanging around or from Goodwill, but burning old-fashioned audio CDs does take a while (relatively speaking).
Thanks, Dave! 🙂
I appreciate your sharing feedback on what works for you. Glad to see that you tried out VideoCritter and found it to be a viable option to using YouTube’s program.
Not being a “geek” myself, I’m not an expert on the technical end of things – I know what I want and I know how to advocate for it, but knowing how to actually make it happen via technology is definitely not my area of expertise.
I am thankful for people like Don G. who shared his knowledge, and people like yourself who have the skills to actually test it out and give us your opinion.
You read it here, folks…if YouTube ain’t your thing, then try VideoCritter.
I can appreciate your not wanting to make transcriptions of audio-only podcasts available on-line. I thought about that myself…the potential for such to be “stolen” by others. I don’t know a good way to get around that other than to suggest that you could make a “blank video” and add the captions to it.
The other possibility is to make transcriptions but don’t post them on-line. Instead, just post a message saying something along the lines of “for deaf and hard of hearing individuals, transcripts of this audio podcast are available upon request” – and then post contact information. This way at least individuals such as myself know they have an option of getting the information in a text format that we can comprehend.
I’m still bothered by the tone that you bring to all of this – you say you are sympathetic to those who are hard of hearing; but it seems your sympathy only extends to the level of which you feel you can justify putting your effort into the whole endeavor. This is evidenced by your statement of “we prefer not to devote too many resources to items whose benefit do not outweigh their resource consumption.”
Yet, you go on to say “we have folks on the show with vocal impairments, but their contribution of content is powerful enough to be worth the time to do the editing needed to make it easier to listen to what they are saying.”
That’s my point. You obviously are creating podcasts that people DO think are powerful enough to be worth listening to…otherwise why would you be investing the time and effort to be creating them at all? And if the rest of world can listen to their contents and be educated, entertained, enlightened, and so forth… why can’t I?
Am I not worth it? Or am I just a statistic to be juggled against your resource consumption? A person of lesser importance whose needs don’t take priority in the greater scheme of things?
It seems like what you are basically saying here is “sure… I will support Equal Communication Access as long as it doesn’t require a lot of extra effort on my part, and brings in additional download counts.”
Otherwise, you’re gonna go for a walk in the park…and I’m gonna be shit out of luck.
Dave, thank you for following up on this and for the mp3 suggestion, which I use already for podcasts that don’t require a lot of attention (music programs, for example, that I listen to at night).
But anything with more than a small amount of dialog is difficult to understand. While there have been significant advances in assistive hearing devices (hearing aids, etc.), they are still far from perfect, and as with any electronic device, they have their own issues.
But people who don’t already have hearing problems don’t understand that.
I think people are still missing the point here.
AUTOMATIC CAPTIONING and AUTOMATIC TIMING on YouTube’s captioning program…
ARE NOT THE SAME THING!!!
Yes, I agree with those who say the AUTOMATIC CAPTIONING program sucks – this is what a lot of people are doing…they are just clicking the program for YouTube to automatically caption their videos…which it does by using voice to text software that converts the sounds of speech into readable text – often with inaccurate results which can sometimes be humorous, but more often just puzzling and frustrating.
The AUTOMATIC TIMING program is what you want. This allows you to download a simple typed up transcript of your video… all you have to do is type it up in sentences sort of like a script for a play, and download it as instructed in the video above.
YouTube then automatically inserts your transcript into the video, using the words you have typed to match up to the sound. This helps to reduce/prevent the bloopers that often appear when the software attempts to decipher speech on its own. It will automatically insert the proper words from your downloaded transcript into their proper place in the video.
YOU DO NOT HAVE TO MANUALLY INSERT TIME CODES OR FRAME NUMBERS.
The program takes care of that for you.
All you have to do is just type up a script like the following (quotation marks added for clarity):
“Good evening.
Thank you for inviting me to speak to you tonight.
As you know, I recently returned from a trip to China.
Tonight I am going to share with you some pictures and stories about my adventures while traveling there.”
THAT’S IT, FOLKS!
Don’t be turning this into something more complicated or more time-consuming that it actually is, or has to be. All you need to do is listen to your video, type up your script (if you don’t have one already), download it to YouTube, click the Auto Timing, and presto! It adds it and does all the work for you.
Truthfully I don’t understand why people are sitting around bitching and moaning about how it’s too complicated, too much work, too time-consuming, too… whatever.
JUST DO IT.
Right now we don’t script anything so we don’t have transcriptions hanging around to just drop into our shows like Caption Geek assumes. All our show notes are generated as part of the post-production process, that’s why it takes time for us to generate.
Though it is good to know that automatic timing on transcripts is not nearly as bad as automatic captioning.
Other podcasters script their shows, but it is very noticeable in the audio quality (there’s a near lack of cadence which, to me, makes a show sound boring). For them, the automatic timing capabilities of YouTube would be very easy and not at all time consuming.
As for ocean, I have 24 hours in my day just like you. There are many things I would love to do in the Pagan community like being more involved with the PNC, devoting more time to the PPP etc. However, my resources are limited (not unlimited) and I must make choices about how to use my limited resources. The requirement to make choices means that I will not always be happy and those influenced by how I allocate my resources may not be fully satisfied either.
Luckily I can responsibly rely on technology to augment what I am capable of doing so I can get more done in less time. When we started podcasting 6 years ago, the tools to automate the removal of stuttering did not exist (more accurately, it would take 15 minutes of manual editing to remove each stutter), now it’s essentially a menu option in Audacity that takes a second or two, then you just cut out all the repeating audio.
All the stuff I mentioned above takes minutes over the course of a 4 hour recording. All the stuff you are asking me to do currently takes hours. This isn’t a “I don’t care about the deaf community” – heck, it’s not entirely a “accommodating the deaf community requires disproportionate effort at this time” issue either.
Efforts to accommodate the deaf do translate into practical benefits for those who are without hearing issues as well. For example, these efforts make subtitling to other languages a fully automated thing on YouTube, and Google’s translations are much better than their speech to text. For shows where the hosts don’t object to web-based transcripts, it would improve search engine rankings for those pages (search engines index text, not audio), giving their shows a larger audience regardless of the hearing capabilities of their listeners.
To overcome this disproportionate effort currently required to add captioning, we are (as of last night) investigating the use of more powerful speech to text utilities to see if we can automatically generate transcripts for our videos and just polish up the areas where the transcript obviously made mistakes (which is likely in parts where there is Pagan terminology).
I don’t mind allocating time to a project to tackle this from a technical perspective. Taking a few hours a week for a project that benefits everyone to save hundreds of hours of mindless effort that only benefits our shows seems like a productive use of time. Currently, it seems many speech to text programs just fail miserably when there’s as few as 3 people speaking in the same audio, let alone the 15 or so we frequently have on the Pagan Centered Podcast.
If you have any preferred projects I can join or any software I can try out, let me know — otherwise I’m just going to work down the list of “open source speech to text” on Google.
As for evaluating existing solutions (commercial and otherwise), right now we’re just looking at the usual speech to text programs like Dragon etc. However, any input on on tools you find work well for you in general would help us exert our efforts on this more efficiently.
Hi Dave ~
The fact that we are having this on-going conversation about this subject and looking at ways to resolve what has been an on-going problem for deaf and hard-of-hearing individuals tells me that you are sincere in trying to juggle limited resources with obvious technical knowledge and a desire to come up with a solution that can be mutually beneficial for everyone.
For that, I thank you. Let’s keep the dialogue going.
You have made some excellent points here – that captioning or transcribing podcasts can have benefits that go beyond being accessible for the deaf and hard-of-hearing community. Such efforts can be worthwhile in many other ways as well.
As I stated earlier, I am not a technical expert so I would not know where to begin in finding the software or programs that you speak of.
You look like you are taking the right route here. What you may also want to do is check out Disability Programs in your area (such as at universities or Vocational Rehabilitation or Centers for Independent Living, etc.) to see what they might know about.
I’m going to post my updates under the Twitter tag #pcp-cc so I don’t flood the comments thread here, just search for pcp-cc on Twitter to catch the updates:
http://twitter.com/#!/search/pcp-cc
There’s lots of nifty cool things that can be done once I figure out how to get this stuff working.
Once I figure out how to get CMU Sphinx working (they seem to be the best maintained project with the most resources), I’ll probably want to slap an easier to use interface on top of all this stuff so other podcasters don’t go through the whole scratching-their-head thing I’m doing right now.
If you’re interested, you can learn more about the CMU Sphinx project at http://cmusphinx.sourceforge.net/ – just be warned that much of the content isn’t intended for the average person.
CMU Sphinx isn’t perfect, but in live demos it does much better that most commercial and open source stuff out there right now when it comes to transcribing recordings made in real world environments (not just someone talking into a headset).
Looks like you are doing some great stuff here, Dave.
Once you get something working that you would like to share, just give me a poke and we can figure out how to write something up that can be posted here at the Crossroads (with links or at least credit given to you).
That way folks looking for ways to create greater accessibility (and I do get those requests for information on such) will have some options.
Thanks!
Thank you. I am adding captions to all my animated episodes at the Reformed Druids using these tips. My hearing was damaged by bagpipes in College, so I am sensitive to the issue and studied ASL and raised guidedog puppies.
Have you heard about The Amara Project (www.amara.org) yet? It’s a neat website that allows users to submit YouTube videos they want to see subtitled, and there are volunteers who use their time to get it done. More videos are getting subtitled there every day, so you might want to check it out to see if there’s something that you like. I also run a Facebook page called Subtitle Youtube. I really enjoy subtitling videos and would like to know what kind of videos you’d like to see subtitled.
Regards,
Sara