Wednesday, June 29, 2011

Starting a Video Captioning Business on the Cheap

Do you like to watch movies? Can you type fairly quickly? Congratulations, you are now qualified as a video transcriptionist. And just in time, too, as there will be a gold rush soon for video transcription services.

In October 2010, President Obama signed into law the 21st Century Communications and Video Accessibility Act, which mandates further captioning of video for the Web. The bill will apply to any video that was originally aired on TV. As more and more videos are uploaded to the web, the demand for video captioning services will grow at an exponential rate. Government agencies, schools, museums, and many other types of businesses will need the services of a good video transcriptionist. Your biggest problem may be keeping up with the demand.

For about $150, you can start your own business from the comfort of your living room. Software would cost you less than $130, which would get you a copy of MovieCaptioner and an upgrade to QuickTime Pro. Now all you need is a web page (free from Blogger.com), a domain name ($19.95 per year), and a file upload service such as DropBox.com, where you and your customers can transfer large files over the Internet. Optional items could include headphones and a copy of Dragon Dictate if you really don't like typing that much (as I don't). With Dragon Dictate, you simply repeat what you hear in the headphones into the microphone, and it gets typed into the MovieCaptioner interface automatically! A set of headphones with a built-in microphone would be helpful in doing it this way.

So you have MovieCaptioner and QuickTime Pro, and you've picked out a domain name and have gotten a website put together so you can hang a shingle out for the world to see. A customer inquires about your rates (which you've done your research to see what other companies with similar services are charging). You send them your email address and let them know they can use that to send you their video files via DropBox.com, which gives you up to 2 GB of space for free (you can upgrade to more space for a fee). You share a DropBox with your customer and the files are instantly available to both of you. You import the video into MovieCaptioner, click the Start button, and go to town. After you get your captions done, you export to the format your customer requires, and either email the captioned text file back to them or embed the captions in the movie and throw it back in your DropBox and let them know it's done. An interim step could be to export a transcript with timecode and let them proofread it if they want. That way they can check the spellings of names, technical terms, and so forth. Payment can be done easily via PayPal, Kagi, or another credit card transaction service. You can typically make anywhere from $75 to $150 for a one hour video.

Please be aware that depending on your typing skills and the quality of the audio, a one hour video could take anywhere from 5 to 8 hours to transcribe. This type of work is not for everybody, but some really enjoy it. It is, at least, an easy way to make several hundred dollars or more in a week's time, with a very modest investment to get started. This is a business that a stay-at-home mom or dad can run or even a college student to do over the summer to pay for books and tuition. The potential is there to expand and make a much larger enterprise too if you wanted. Teaming up with others you trust to do good work (see oDesk.com and eLance.com) where you pay them part of the profits to handle your overflow will allow you to handle much more volume.

I hope this may spark some interest and that more people will join in the pursuit of making the world's video accessible to all.

Please try MovieCaptioner to make your movies accessible. You can download a free 14 day demo.

Monday, June 20, 2011

Troubleshooting SCC (Line 21) Caption Files

From time to time you may get an error in your SCC caption files. Here is a (hopefully) helpful strategy for zeroing in on the cause of the problem. Knowing that it is a bad file in the first place is simple. I try to open it in QuickTime Pro. To be able to do this, however, you need to download a free plugin from Apple for importing SCC files into QuickTime. It is called "ClosedCaptionImporter.component". It can be downloaded (currently) from this link: http://developer.apple.com/library/mac/#samplecode/ClosedCaptionImporter/Introduction/Intro.html.

Once you download it, just drop the plugin into your QuickTime folder in your hard drive's Library folder...

Put the closed caption importer plugin into your hard drive's QuickTime folder in the Library

If you have QuickTime 7 open, you'll need to close it to make the plugin available when you re-open it. Note that as of this writing, QuickTime 7 is necessary for all the features of QuickTime Pro to work. QuickTime X, the default QT Player that comes standard with Snow Leopard does not have all the functionality of QuickTime 7. You can install QuickTime 7 from your Snow Leopard install disk, then upgrade it to QuickTime Pro via the Apple website. It only costs $30 and it well worth it for the increased ability to work with QuickTime movie files.

So, as long as you have the plugin installed you should be able to see captions in an SCC file that you open in QuickTime Pro. It should look like this...

A working SCC file in QuickTime Player

It should have a white background and if you scrub through the movie, you should see captions. A file that has errors will open like this...

A non-working SCC file in QuickTime Player

It will have a black background, it will display smaller than a working file, "Scenarist_SCC V1.0" will show at first, and the binary code will display instead of the transcribed captions if you scrub through it. If you see that, then you know you have troubles in River City. Sometimes, though, an SCC file can have a garbage character in it that may allow it to play in QuickTime properly, but will cause an error in other applications.

To troubleshoot this, I'll open the SCC file in BBedit or the free version called Text Wrangler, available from Barebones.com. Here I'll do a quick scan to look for certain types of anomalies. Here are a few types of things I might look for.

A blank caption after the timecode. All lines must have at least some formatting applied, even if there is no text in the caption.

SCC line with a blank caption

Another thing to look for is unpaired binary code. Every set of alphanumeric digits should be in sets of four usually, although they can come in twos, I believe. MovieCaptioner always creates them in sets of four digits. If you see anywhere where a line ends with a 3 digit number instead of 4, there's probably trouble there. Sometimes this could be caused by a character in the captions that is not supported by the SCC format which can goof up the sequence of numbers.

a 3 digit SCC line with an unsupported character

In the example above, an unsupported character is showing up as a bullet. Instead of it being a 2 digit number to represent the character, it is only one digit, which throws off the rest of that line. Note the last number sequence is only 3 characters long as a result. This will throw an error when it goes to display in QuickTime. The line should have ended with " 942f", but instead it is " 42f".

The other more difficult thing to find is what's known as a timecode overrun. In the SCC format, each caption needs about 1 frame of buffer time per every 2 letters of a caption. If you add all that up and subtract if from the timecode for that line, it should not be less than the timecode preceding it. So the caption "I may be going now" might need 18 frames of buffer time to display. Subtracting 18 frames from its timecode should not run into the timecode of the previous caption. This seems to vary, however, with different captions, so I've not found it to be a hard and fast rule. If someone would verify what this is exactly, I'd love to know myself. Before you can do anything about this, however, you need to find the caption (or captions) that have timecode issues. I do this by opening the SCC file in my text editor and using cut and paste, I test out sections of the file. I will typically select and cut all the captions from the midpoint of the file till the end and save it. I will leave the text editor open while I test the saved SCC file in QuickTime Pro. That way I can undo my cut and try again. I'll drag the SCC file onto the QT icon and see if it opens properly. If it does, I know that the error lies in the last half of the file. I will keep splitting the file in this manner until I get it to work again and then not work again, so I will eventually home in on the line that's giving me the problem. Once I identify the offending line, I will look it over for anomalies as above. If I can't find any, I look at the timecode and count the characters to give me an idea of whether or not the timecodes may be too close. If they do look close, I will start to add frames to the offending line's timecode and saving, then testing in QuickTime Player until I get it to work. I'll then make sure there are no other lines in the SCC file that are causing such problems. These timecode overruns typically are caused when 2 or more captions are long and come in rapid succession. I will sometimes have to abbreviate words or divide the captions more evenly to iron out this problem.

But how do you find the captions in your project when all you can read is binary code??? Well, I've got an answer for that. It's called SCC Caption Decoder, a lightweight application that will decode an SCC file and save it as a text transcript that is human readable.



This really helps. After running the file through SCC Caption Decoder I can look for that same timecode where the problem was and know what the captions says so I can find it in my caption project. You can't just go by timecode necessarily because for the SCC format, the buffer time is built into the caption's timecode, so even though you set a caption at 00:04:21:07, the buffer time might change it to 00:04:20:01, for example. SCC Caption Decoder can also convert the time to what the original time was if you select the optional radio button. If you don't need the timecode, it can also save it in paragraph text form as well. It's a handy little tool for people who work with SCC captions a lot, and is really great if you just want a text transcript of what an SCC file contains. Give it a try. There are both Windows and Mac versions and you can try it out for 7 days. Cost to purchase is only $19.95. A convenient tool to have in your arsenal.

decoded SCC file as readable text

Please try MovieCaptioner to make your movies accessible. You can download a free 14 day demo.