Barney and I were talking about this the other day. I really wish that somebody could create what he’s looking for, as I know I needed one when I was reporting more (as I’m in the same boat as he):
I’m a reporter. I skipped third grade, so I’m sure I missed some handwriting classes. Ever since, my handwriting has sucked, especially when someone is speaking fast and I’m trying to get down every interesting word. It’s sort of a self-encryption – if I don’t transcribe the notes within 12-24 hours, even I can’t make out what it says. (It’s why I love phone interviews – not out of sheer laziness – well, not just that – but I’m a real fast typist, so those notes are 1,000 times better.)
Sooo…. there’s voice-recognition software out there, right? Getting better all the time? But apparently you still have to train the thing? I’d LOVE to tape interviews, rather than scribble them, then upload the audio file and have a program turn it into text! Heck, I’d settle for 30 percent accuracy, that’d be better than my stinkin’ notes!
Does anybody know if such a thing exists or is even on the horizon? I know it would require a good chunk of CPU power, but those portables/PDAs/etc… are getting pretty dang powerful. It would be a blessing for so many reporters, that’s for sure.
Comments
Something like that would put me out of a job as a medical transcriptionist. And i haven’t heard anything about something like it.
Maybe Barney could find someone in the local college’s MT/LT program who would like to make a little extra cash practicing?
LOL, the problem is, MY handwriting would test ANYone’s ability to read it, or ANY handwriting-recognition program!;-) And to dictate my notes to someone, I’D have to be able to read them! So the voice-to-text thing would be SO helfpul. I know, my pal Jesse tells me what a tall order I’m asking for, technically. But I really believe whoever cracks this nut, even to a 30 or 40 percent degree, will make a fortune. And I also believe it would be useful for FAR more than just reporters.
oh yeah it would barney, like i said, i’d be out of a job if something like that came about :P. I suppose i could switch to proofreading? hehe.
But you could take a note from the doctors (whoa, large pun in there)… conduct your reporting with a minicassette, buy a transcriber machine.. but i and most transcriptionists are on a 24-hour turnaround and i suppose you’d need your bits back much faster.
Of note- a lot of doctors have as much problem with enunciation as they do with writing. heh! I mean, we are talking mouths full of marbles AND crackers.
http://www.dragontalk.com/NATURAL.htm
This should be good…I’ve never actualy tried it, but a friend of mine had an older version (and 30% is pretty low, it should be more like 80-90).
Also, if I remember correctly, they used to have a PocketPC version a while ago…
Have fun..
Well, if you’ve got the sound file, then you can easily just listen to it over and over and get down every single interesting word… That program would be great, but this would be the currently realistic approach, no?
I did a little homework on this in setting up my business and some more recent research in looking into digital call-in systems for my clients who dictate reports and letters. Most of the voice-to-type softwares do require you to build language models (their vocabularies) and invest in editing. I don’t use them in my transcription business because they increase my editing time. Too many sound-alike words, and in working with scientific information, etc, there are chemical compound names, formulae, and the like. They just can’t keep up. The human ear and the human brain, teamed with human fingers outputting the translation, are still the best .
If you need digital sound files transcribed, they can be easiy transferred online or via email to an off-site transcriptionist (like me). No long-term commitments required. I work on an as-needed basis.
If you’re using a Sony or Olympus digital recorder saving compressed files, some are small enough to be emailed. Others have to go through an FTP site, and some larger wave files (recordings 6-10 hrs or more) for long interviews of several hours can be express mailed on CD.
I don’t recommend micro cassette recorders or micro cassettes, although I do still accept those, and standard cassettes too, by mail. The quality can vary greatly and there’s the hassle of making a copy (never send your original) and the possiblity of loss or damage. Digital recorders can be purchased rather inexpensively and have software that allows you to upload directly to your computer. If you need one that’s portable and can be used from anywhere, I’d recommend NCH Express Dictate (http://www.nch.com.au/software/dictation.html).
I also recommend the Olympus because of cost and ease of use. I have one I use for meetings, lectures, and personal dictation, as well as time and expense information.
Hope that helps.
And now that I’m in TV, I know it’d be a GODSEND for TV to come up with an automated way to do speech-to-text. News shows must have dictationists for live stuff, the scripted stuff isn’t a problem of course. So there’s GONNA be an answer, hopefully sooner than later, but no doubt VERY costly at first.
Not sure I was clear on that, so let me add – I’m not referring to transcribing of SOTs (sound on tape) for a reporter, but for the CRUCIAL (as in FCC fines otherwise) closed-captioning of all things said on TV. When it’s live, that’s a big problem/hassle/(opportunity?)