Slashdot Mirror
Typingpool: Human Audio Transcription Parallelism
theodp writes "Silly rabbit, parallel processing is not just for Big Data! Building on techniques outlined by Andy Baio back in 2008, Wired writer and 20% Doctrine evangelist Ryan Tate has released Ruby-based software called Typingpool to make audio transcriptions easier and cheaper. 'Typingpool chops your audio into small bits and routes them to the labor marketplace Mechanical Turk,' Tate explains to his reporter pals, 'where workers transcribe the bits in parallel. This produces transcripts much faster than any lone transcriber for as little one-eighth what you pay a transcription service. Better still, workers keep 91 percent of the money you spend.' Remember to Use the Force for Good, Tate adds."
Seems an interesting idea for long recordings needed in a hurry, but the transcriber will be losing possibly important contextual cues by reducing the length of the utterance. Also, there is a great overhead per-person in terms of manual labour (waiting for audio to buffer, HCI interaction, etc). On the upside, it might be less dull to listen to shorter, more varied recordings than one long one. But it would suck to transcribe half a murder-mystery.
The biggest market for audio transcription I'm aware of is in medical transcription, followed by legal. Many of the terms used are not normal english; In fact without a basic understanding of medicine, you could easily confuse one thing for another, with potentially tragic results. This is fine for everyday english, but not in industries where terminology is used that isn't. And that's a lot of it. This would be more useful for something like Siri -- no doubt Apple has humans to process some of the unknowns in the background, and would find a service like this useful, if they don't have something similar already.
#fuckbeta #iamslashdot #dicemustdie
Five or six years, I transcribed podcasts for Mechanical Turk. Their audio files were already split into shorter passages (3 or 5 minutes, for example). Split them even further, and the transcriber might miss out on the context, which is often vital to knowing what exactly is being jabbered about.
That said, I'm not sure who is the target demographic for this kind of work anymore. Many of the podcasts were on subjects of interest to me, and I was getting about $10/hour from Mechanical Turk, which wasn't bad considering that I was often doing the work from backpacker beach havens around the globe where a couple of hours of work would pay all my expenses for the day. But the last time I had a look at Mechanical Turk, the amount they pay had been heavily reduced. Who wants it now? Even if you are in a cheap third world country, if you have the English skills for such transcriptions, you can surely find better and more dependable work elsewhere.
This reminds me eerily of the term, "zipheads". This is not a racial slur. Those who have read "A Deepness in the Sky," will likely recognize it too.
How are the fees calculated?: "Amazon Mechanical Turk collects a 10% commission on top of the reward amount you set for Workers. For example, if a HIT reward is set to $0.20, Amazon Mechanical Turk collects $0.02 for each assignment." So, the worker gets 91% of the total amount paid ($0.20/($0.20+$0.02).
Maybe I just lack imagination, but how could you use it for ungood?
Examples?
The volunteer process by which Japanese anime is subtitled within hours of release works a lot like this.
And a new group of people expected to go on unemployment. Good job.
What about when the whole is greater than the sum of its parts? This sounds great if what you want is a text dump transcription of a pod cast or some such. But, what of formatting and assembly of the whole document?
This sounds only slightly more useful/effective than automated speech to text transcription.
I transcribe professionally, and this sounds like a really bad idea. Like others have said, context is really important. A lot of what I do is hard to understand. Often I will go back and fill in stuff that I couldn't understand in the beginning but am able to understand after hearing the speakers for a while. So if I were just doing random snippets from a bunch of different recordings, all with different speakers and audio qualities, accents, etc., I would do a horrible job. This also would make it more difficult to have a consistent style in terms of editing for nonverbatim transcription. Maybe this would be OK for crappy subtitles or quick podcast transcriptions or whatever, but it sure wouldn't replace any of the stuff that I get paid for.
Human Computers describes the (mostly) women who did the mathematical calculations the engineers handed them to do for the war effort, freeing up valuable engineering time.
The article doesn't say how much parallelism there was in these "human computer pools" but I suspect there was a lot of it.
I guess it goes to show that even today, some problems are, for the moment at least, done cheaper, faster, and/or better (pick any two) with a person than an electronic computer.
See also: Logopolis, a Dr. Who serial from 1981 which features people (not humans) doing mathematical calculations rather than having computers do them.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
Like any tool, Typingpool could probably be used that way. Please don’t use it that way!
You're using a service that generally does not attract people from the developed (First World) nations. It *is* being Used That Way.
Typingpool defaults to paying $0.75 a minute, and I often offer $1.00/minute, which produces transcripts very quickly, tends to attract better workers and is still roughly half the best rate I’ve seen for high-speed professional transcription. I have had success completing transcripts at lower rates, and Baio four years ago was able to find plenty of workers at $0.40/minute. But lower prices generally translate to slower transcription and lower quality.
You get what you pay for, but the service would have to have a builtin "No Third World" option defaulted to "on" for it to work well.
Also, bear in mind that just because you pay Mechanical Turk workers a lower rate through Typingpool than you’d pay a service doesn’t mean the workers are actually getting paid less or are worse off
However, the likeliness favors the conclusion that they are getting paid less/worse off.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
He thought that he was too good for US citizenship.
Perhaps the US should pursue on that additional $62 million by returning him and said assets to US jurisdiction. In this case, the military and intelligence used to effect his return would be a force used for good.[
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.