In a nutshell, here’s the problem. I have to get audio voice data from a DVD soundtrack onto paper as a text file. What’s the best/fastest/least-hassle-free method?
Ideally, using a PC, I should be able to process the soundtrack from a DVD thru a voice recognition program and produce a Word DOC file. In merely seconds, right? Since theoretically a processing program doesn’t need to listen to the sound data in realtime, right?
Now in reality, how close can I come? I ran across this free software:
…which purports to allow an optional footpedal to control the stop and start of a WAV file. It’s a start.
My experience with optical character recognition software some years ago suggests that voice recognition (Dragon Naturally Speaking?) might be usable, but need some serious editing. If I have to do it, so be it.
And the DVD will be coming from a government meeting where the voice quality and S/N won’t be studio-quality. Also if I can bypass time-consuming steps like transferring the audio track to a WAV file in realtime, I sure would like to.
And if I have to control the recording by backing up, pausing, playing double-time, etc., for editing, I would like to take advantages of the DVD environment (or an audio/video computer file), since the video can provide useful clues like lip-reading and telling me who the speaker is. Chopping off the video portion would seem to be a step backwards.
Just listen to it and transcribe it. Forget about any kind of voice recognition software, it would take just as long to go through and fix all the inevitable mistakes the software would make.
Was there nobody already transcribing this government meeting? Were they at least taking minutes?
I’d venture a wild-ass guess that to get the same 99% accuracy you get with OCR from a voice recognition system you’d need at least 100x to 1000x the computing power. Then you’d need to train it on each individual voice involved by feeding it samples. At least with current technology.
I asked a while back whether it was worth upgrading my version of MS Office to take advantage of the newer software’s voice recognition capabilities, and everyone who responded seemed to agree that it wasn’t worth the hassle. They said that it’s very hard to work efficiently with it even after you’ve trained it to recognize your own speaking style. I imagine that the difficulty of getting it to recognize multiple voices—some of which are probably louder than others, and some of which probably speak rather indistinctly at times—would make your quest pointless.
I can dig that. So discard the voice rec software idea; I’d still like to bypass the transfer-DVD-to-audio-computer-file if possible. Unless it could be done faster than real time. No theoretical reason why it couldn’t be done.
No local government meetings I know of (and I videotape several per week) are transcribed word for word. Minutes are taken, but since they sometimes differ in significant ways from video or audio tape, we need a verbatim text file.
This is not a one-time thing, but an ongoing project, so you can see efficiency is important.
Is there any software, say a tiny toolbox window floating on top of Word, that will allow easy control of start/stop/backup/play functions on an internal computer DVD player? With a small video window synced to the sound?
But it would be so much neater (hi-techer?) to be able to play the DVD on the same computer you are typing on. Like in a small 1/4 screen window, where you could see and hear right in front of you instead of a separate TV/DVD player to one side. Easier to take your work to a different location, too, than lugging multiple gadgets, even if the player & TV were small (which mine aren’t).
But if I went with the separate DVD player, do any of the higher-end units have jog/shuttle wheels? I only have cheap players and I miss those wheels I used to have on the VHS units. I would imagine they work even better with DVDs as the media.
I can’t think of any commercial player for a PC’s DVD that has an “always on top” mode such that the display screen and play controls are still visible when another app is maximized and has focus. However, what’s to stop you just running Word in a non-maximized window? I realize that you won’t get optimal use of screen real estate, but if you arrange it so that the Word window is on the left half of the screen and the DVD-player window on the right, you can move back and forth relatively easily – certainly with the mouse, and if you’re lucky with the keyboard (depending on DVD player software). I’ve done this before, and didn’t mind the fact that the Word window was small, since transcribing is a very “linear” process (i.e. you don’t need to have much of the document visible at any given time, since you’re just typing straight through from the beginning to the end).
All things considered, that’s probably the way we will be going.
I know there are audio players – and I think audio player software programs – that are optimized for transcribers who need to shuttle & repeat frequently. Their needs are a little different from the average movie viewer’s or CD-listener’s. What I would like to find is the equivalent software for audio/video. Anyone know about something like that? I’m afraid DVD player control software that comes with the computer may not be as easy to use (I haven’t tried any yet, however).
Once upon a time I was in the business of transcribing music, and I remember making some serious modifications to the tape machines to make the job easier. It made a significant difference in the time it took to write down a song.
Just brainstorming here…the ideal program would not require copying all data from a disc to HD or RAM, to save startup time – but would buffer data as first read to make shuttling and repeating small sections faster and easier. That would be analogous to shortening the delay time allowed for the reels to come to a stop after rewind on the tape recorder, which I did once.
Also the ability to place annotated bookmarks in data would be useful.
When I transcribe (not very often I admit), I just open up Word, and use the little
Pause/Play button I have on top of my keyboard to control the DVD or CD. I don’t need the DVD player to be always on top to control it. It’s not part of a standard keyboard, but it seems most keyboards these days come with multimedia controls on them.
The best, fastest, and most hassle-free method to transcribe audio would be to hire a stenographer. If you ask around your local courthouse I’m certain you could find a court reporter looking to make some extra money on the side.
There’s no doubt they could provide a full, verbatim transcription of your meeting in real time, which is good since you’ll be paying by the hour for it.
If you contact a stenography school you may be able to find students willing to work much cheaper.
There are two groups involved here. Group #1 is the official clerk/secretary who records to audio all government meetings then condenses that to minutes (or uses the audio to check the minutes she took live) using a special audio playback unit. This unit is for some, a physical tape recorder/playback unit; for others, a computer program. The clerks have considered using the video recording instead, as the audio is an identical feed to both and it seems a waste to duplicate efforts.
The government has no money in the budget to pay for outside help, even if using clerks on the payroll is more costly in the long run. That’s how govt works.
Group #2 is composed of citizens who sometimes find egregious errors in the official minutes and want to make an exact, verbatim transcript available to the public. Their budget is US$0.00 and they will be relying on volunteers with their own computers.
It’s a good idea otherwise, SiouxChief.
Again, no money.
Not at all. I get paid for video recording most govt meetings for permanent record and for cable access TV anyway, and minutes or a steno record cannot be relied upon for accuracy; the video and audio can.
Pulykamell, my computer is too old to have those special, on-keyboard playback controls, but the people who will be doing the transcribing may have such, or I believe they can be purchased separately for desktop computers. That might be the best compromise yet. Still, if special audio playback software is available tailored for this purpose, is the equivalent available for video?
Not to sound too contrary, (and not to hijack this thread) but if “…a steno record cannot be relied upon for accuracy” how come every courtroom in the US relies on them to record proceedings?
IANAL, but I suspect that if a steno record conflicted with a clear video or audiotape record on a critical point, the steno record could be corrected officially.
Unless it could be shown that the video or audio was tampered with, of course.
Our local problem is not with steno records, however, as there are none. The clerks take notes during a meeting on either a laptop or on paper (none know shorthand as far as I can tell), then rewrite or edit those notes later into minutes which are approved at a subsequent meeting. Some clerks claim they refer to the audio tape while editing; others I don’t know about.
I have found blatant disagreement between minutes and videotape on a few occasions. Whether they were deliberate or accidental, I don’t know.
Windows Media Player 10 on Windows XP (and I think 9, and perhaps on other Windows versions) can play DVDs.
It also has a little toolbar control that gives you the main buttons (play, pause, stop, back, forward) on your taskbar. To get this, right-click your taskbar then choose Toolbars and check "Windows Media Player) . When you minimize WMP, you will see the WMP toolbar in your taskbar.
You can also hit CTRL+P to Play/Pause. However, this doesn’t work when it’s out of focus (if you’re in Word, that’ll get you the Print Dialogue box).
You can also go to tools - options - general to mark a checkbox for “always on top”. You would want to have the player in “skin mode” to get a small version of the video. Otherwise, just minimize it to the toolbar.
The best idea yet I think would be to get a “media keyboard” like pulykamell suggested. If $$ is an issue, get one off eBay.
Since theoretically a processing program doesn’t need to listen to the sound data in realtime, right?