Quote:
Originally Posted by VatorMan
Live closed caption always looks like it is run through Apple autocorrect. And you know how accurate that is.
|
The thing is, I've seen voice recognition software that works fantastic compared to what I see from their "transcription service" or whatever it is they use. It works in near-real time and I have no idea why it isn't used for stuff like this. And a lot of it is free. You wouldn't believe how good some of the free tools are. I've recently been using an optical character recognition engine called "tesseract" that is friggin' amazing. You can feed it a tiff or pdf and it will not only automatically recognize any text it finds, it will also provide the location of that text so that you can highlight it with a little bit of coding on your part.
You can literally take a picture of the page of an old book, feed it to tesseract, and in less than a couple of seconds have it perfectly turned to text.