SpinVox "Recognition" Is Often Expensive Human Transcription 226
An anonymous reader writes "SpinVox offers to convert voice messages to text using a system called D2 or 'the Brain.' According to BBC News, said 'Brain' is often of the old-fashioned kind: SpinVox is sending private voice messages to South Africa, the Philippines, and maybe Egypt to be typed by people in a call centre, despite being registered as keeping all private data inside Europe and claiming that the text is somehow anonymised. Insiders say they transcribed 'love messages, secret messages' and everything else from beginning to end, and the company is being bled dry by the cost: SpinVox has been locked out of one of their data centers over a payment dispute. SpinVox refuses to comment further on details — but according to their web page, they're 'enabling the Speech 3.0, Voice 3.0, and Business 3.0 markets,' whatever that means."
O(human) (Score:5, Funny)
Re: (Score:3, Funny)
I haven't tried it, but their video sure is convincing
*facepalm*
Speech 3.0 (Score:5, Funny)
Now with 20% more vowels!
Re:Speech 3.0 (Score:4, Funny)
So it's Japanese? :-)
Re: (Score:2)
Re:Speech 3.0 (Score:5, Interesting)
He's probably referring to the frequency with which vowels appear in any given word. Yes, Japanese has only 5 vowels, but because almost all syllables in the language are simple (1 consonant)(1 vowel) pairs, almost every other letter in a written word is a vowel.
A common tongue twister:
Nama-mugi, nama-gome, nama-tamago (uncooked wheat, uncooked rice, uncooked eggs)
Notice the abundance of vowels.
Re: (Score:2)
Actually that's a poor comparison. English spelling is different from Japanese in that there are lots of unpronounced letters, as well as single sounds spelled with multiple letters. What if we redo that with a more phonetic respelling* (imagine "hard" vowel pronunciation).
Uncukd: 2 vowels, 4 consonants
Wet: 1 vowel, 2 consonants
Ris: 1 vowel, 2 consonants
Egs: 1 vowel, 2 consonants
That's consistently twice as many consonants as vowels. This is generally true because English syllables generally have one vowe
Re: (Score:2)
Speech 3.0: 50% more participants!
Re: (Score:3, Insightful)
Speech 3.0: 50% more hype.
Seriously. I'm just waiting for "Web 3.0. For everyone that got fed up with Web 2.0 and wants more of everything."
Web 2.0 was "You make the content, we make the profit".
Web 3.0 will be "We also make you host the content through P2P, and we'd launch it, we haven't figured out how to make profit of it, though".
Re: (Score:2)
Re: (Score:2)
With 20 percent more consonants, it would certainly be some slavic languate. Most likely Czech.
Seriously. A language that uses N and R as a substitute for vowels has issues.
Re: (Score:2)
Str prst skrz krk.
Vlk prv strhl srst srn, zhltl hrst zrn.
Smrz pln skvrn zvlhl z mlh.
plch zdrhl skrz drn, prv zhltl hrst zrn
chrt pln skvrn vtrh skrz trs chrp v ctvrt Krc
I do admit, I had to look them up. Also, I can see how the use of those sentences is a wee bit limited to very special occasions (read: they're fabricated and don't really have any sensible application). But they show that you can indeed form whole sentences without using a single vowel.
What I'd wonder, do they have to pay for vowels in Czech
Business 3.0? (Score:2, Insightful)
We're not even done with Bubble 2.0 yet!
Re: (Score:3, Interesting)
We're already on Bubble 4.0. The first bubble was Goldman Sachs orchestration of the dot-com bubble (selling worthless websites to stock market speculators). The second was the mortgage bubble. Then Goldman Sachs orchestrated the oil bubble of 2008, and now they're creating another bubble built on money borrowed from China (aka the bailout bubble) which is not real production, but fiat.
That's 4.
So invest now in the market. Thanks to Goldman and their buds in the treasury/central bank (former GS employe
Re:Business 3.0? (Score:5, Insightful)
When we repealed the (very good) legislation enacted in response to the Great Depression, we restore to market to its natural boom-bust cycle. We'll keep going through these periods until we restore the safeguards that our great-grandparents wisely created. Even without the dubious benefits of computer models and Chicago economics, these people gave us 50 years of prosperity that we've managed to wreck in a decade. Shouldn't we stop arrogantly assuming that they were wrong, we are right, and accept that we might need regulation after all?
Re: (Score:3, Insightful)
Economic Dogmas (Score:4, Informative)
The real problem is that people have lost their heads in the United States. The return of evangelicals has led to an atmosphere that is literally opposed to science. So, you get exactly what you expect. Opinions that are based on anecdote and wish thinking instead of data. The reason science works is because you start with the assumption that you don't know something until you can prove that you probably know it, with repeatable, verifiable results. When you start trusting the word of pill junkies [rushlimbaugh.com] and homophobic college dropouts [hannity.com] versus the entire scientific community and their reams of data, get ready for some wide-reaching and catastrophic fuckups.
Canada kept the rules. The Canadian banking system is still the most sound. Every time we take cops off the financial beat, we end up with a banking crisis. These realities can be arrived at by simply reading about the last 30 years of panics, and the hundred years of bank panics that existed before the FDIC and sensible Great Depression legislation.
But leave it to the same fuckers from Harvard, who apparently can't even manage a college trust [vanityfair.com] without running it into the ground.
The pro-market propaganda will continue, and probably destroy our economy beyond repair. And then some wise ass will say that it shows that the market does work, by wiping itself out.
Re:Economic Dogmas (Score:4, Insightful)
The cause of economic downfall is almost always plain greed.
change is good (Score:2)
Canada kept the rules. The Canadian banking system is still the most sound.
Think about what does a "sound banking system" actually means. It means that old money stays that way. It means that generation after generation, the same banks gain more and more power and get to call more and more of the agenda. Stable banking systems are good for people who are already wealthy and powerful. Wiping out unwisely invested wealth punishes the greedy and gives the have-nots a new opportunity.
But leave it to the same
Re: (Score:2)
The problem is when we have this regulation, but other countries don't. It is all part of living in a global economy. If you hobble your own companies, then global companies grow bigger and the dominate.
Re: (Score:2)
How good can a transcription be? (Score:4, Interesting)
Re: (Score:2)
Re: (Score:2)
No, it's okay if they're in spane and it's raning.
Re: (Score:2)
Your use of the English language begs for a caning.
Re:How good can a transcription be? (Score:5, Funny)
The more you know (Score:2)
Re: (Score:2)
Re: (Score:2)
Their can bee only won!
Re: (Score:2)
What's the problem?
There's no privacy issues. If the translators don't speak english, they won't learn any secrets.
It's not a bug; it's a feature.
Re: (Score:2)
new text-to-speech algorithm (Score:2, Funny)
By the way I'm releasing a new text-to-speech service; the algorithm makes for a very smooth speech. It does however have a little bit of an accent.
In case you were wondering.. (Score:5, Funny)
From their PDF:
Speech 3.0: Fully-hosted, commercial strength SLAs, proven scale and reliability - no CapEx. Scales on demand to 150m capacity
So Speech 3.0 provides 150 meters of service-level agreements with no experience-point cap.
Voice 3.0: Superior and proven range of voice products. We repeatedly deliver great, mass-market experiences with our expertise in marketing and management of all lifecycle stages.
Voice 3.0 takes you from larva, through pupa, all the way to butterfly, and then you die and get eaten.
Business 3.0: Mature yet flexible business models - designed to adapt to the dynamics of service brands we partner with, from on-demand to full lifecycle revenue strategies
Business 3.0 is apparently a flexible business model where they interact with their partners. So that's new I guess, no one has thought of that yet. It's also where people who write marketing buzzwords go to die.
Re: (Score:2)
Thanks for the laugh.
You might also enjoy learning about the Turbo Encabulator [youtube.com].
Re: (Score:2)
Holy crap, that made my eyes cross.
Re: (Score:2)
It's also where people who write marketing buzzwords go to die.
I thought that was Redmond!?
Was bound to happen (Score:2)
The speeck recognition people have broken their promises for several decades now. Using humans is still the only working speaker-independent way to do it.
What I find surprising is that it is apparently not cost effective. Here is an alternate approach: Have people transcribe it, but let them look at "pictures" as reward. Seems to be working well in breaking catchpas, so why not for this?
Re: (Score:3, Funny)
Okay, humans never screw up their speeck recognition, but that doesn't guarantee that the speeck is correctly transcribed.
Denial from Spinvox here (Score:5, Informative)
Spinvox has a denial here [guardian.co.uk], claiming this is a case of disgruntled employees spreading falsehoods.
Of course one'd expect them to deny it, but they've just upped the stakes. They would be in violation of UK privacy laws *and* lying through their teeth if this denial is false.
Re:Denial from Spinvox here (Score:4, Informative)
Re: (Score:2)
That's akin to saying someone who said he didn't kill someone could be due for murder AND contempt. Do you really care they care for the minuscle transgression if they're found guilty of the grave crime?
Re: (Score:2)
I have had to work with UK privacy laws before, and trust me, violating them [ed.ac.uk] is nothing like murder (see point #1 in the link). It's more like a slap on the wrists and a small fine. Lying and prolonging the media coverage, OTOH, means more customers get to find out that you're lying scumbags.
Which is why IMHO Spinvox is indeed innocent (and is the victim of disgruntled employees) or an especially brazen scumbag.
Re: (Score:2)
They would be in violation of UK privacy laws *and* lying through their teeth if this denial is false.
Wow. Maybe they should consider standing for Parliament.
Automatic Slashdot speech-to-text (Score:5, Funny)
Captcha cracking (Score:2)
Nothing new here (Score:2, Insightful)
Human transcription performed on industrial scale by non-native speakers is nothing new. For example, medical imaging texts are typed up by Cheap Foreign Labour from voice messages recorded by doctors. ;)
So remember this next time you read the analysis of your expensive MRI test.
Re: (Score:2)
Yeah, I mean you can see Benjamin Rand speaking into a medical transciption device as he modifies his will to include Peter Sellers' Chaunce the Gardener in the movie "Being There."
When all you have is a hammer... (Score:2, Insightful)
every problem looks like a nail.
When all you have is six billion, renewable fueled, autonomous, self replicating, self housing, self programing, hundred billion node neural networks...
who the fuck needs an AI for voice recognition?
Bender vs Apu (Score:5, Insightful)
Losing your job to Bender: technological progress.
Losing your job to Apu: outrage.
But really, what's the difference? A service is a service. It's all progress .. sort of.
Re:Bender vs Apu (Score:4, Funny)
bender does the job perfectly over and over for a lower cost.
Apu does a poor job, frequently making mistakes to the point he isn't cheaper in the long run.
THAT is where the outrage comes from.
Re: (Score:2)
bender does the job perfectly over and over for a lower cost.
You obviously haven't seen bender work.
Re: (Score:2)
Wildly OT, but I can't resist...
Bender actually works on several occasions - especially the episode Bendless Love (S03E06) comes to mind, where he's scabbing his way out of sleep-bending. :-)
Re:Bender vs Apu (Score:4, Informative)
I call Raciest on you.
You must not have worked with Indians before, they are just as good if not better then most American Workers, today.
Especially if you have a good management team who can talk the language and know the culture. Sure you will come up a couple of bad eggs or some horror stories. But really you can get those same stories from any group of people. However I find them in general to be very motivated workers and rather quite intelligent and willing to learn new things. They became the american ideal while we have gotten fat lazy and feeling entitled.
The Robot will not do the job perfectly, hence the completive advantage of SpinBox humans can translate human speech better then a computer can. Robots have a lot of hidden costs as well. You change your process you need a full set of new robots and technology. Or you spend a lot of money for more general use robots which preform slower.
The cost of outsourcing isn't as cheap as saying well and American gets paid $25 an hour while an Indian gets paid $5 so it is 5 times cheaper working with India. There is extra management of working with people in different areas and other costs however this is a management issue which can be optimized to work.
I am sure if the work was being outsourced to a country were people speak the same language and look and have a similar culture to us and lighter skin, then there would be less of an outrage. You may deny that fact, and you may believe your denial. However I bet if you honestly looked in yourself you will realize most of the outrage with Indian workers is that they are not you race of people.
Outsourcing to India has many benefits besides cost being halfway around the world allows 24 hour operations. In essence doubling your output. And they are hard workers who do good quality work. Now if your management is stupid then you may get bad results but that is true anywhere.
Re: (Score:2, Interesting)
Re: (Score:3, Insightful)
So where is your outrage against the western corporation that hired them to lie to you?
Re: (Score:2)
Re: (Score:2)
If that is the case, then the jobs will come flooding back in the long run anyway. So alls well that ends well.
Re: (Score:2)
Losing your job to Bender: technological progress.
If you lost your job to Bender it means you're even lazier then the guy who accept packages at the moon amusement park
Re: (Score:2)
>May be you are more lazy _THAN_ the guy that learn English as a second language.
And you must be lazier than the guy that _LEARNED_ English as a second language.
>If you don't even know these two, then you are not qualify to be on slashdot.
I guess you're even less _QUALIFIED_ to be on Slashdot.
Re: (Score:3, Informative)
You are missing the point. Bender stays in the EU, so he's bound by the laws of the EU. Moreover, it's not probable for Bender to go off and steal your data. Okay, he might accidentally burp it all up. But he wouldn't go use the information to extort you. (Well, he would. But I mean a computer wouldn't.) Apu might go and Nigerian spam your ass using the information you were lead to believe was kept highly confidential.
Also, the idea of having a robot transcribe your love messages is far more acceptable to m
Re: (Score:2, Interesting)
Also, the idea of having a robot transcribe your love messages is far more acceptable to many than having a guy listen to your deepest thoughts and giggling while doing so. Who knows? He might even put a few jokes in there.
Heck, you can just see it...that's movie material right there.
Spinvox employee #1 to Spinvox employee #2: "Oh man, will you look at this? How is this guy ever expecting to get laid when this is how he tries to woo a woman?"
Spinvox employee #2 to Spinvox employee #1: "Well, we did just get
Re: (Score:2)
We should use another robot but Bender as the analogy here. He just doesn't fit. Bender WOULD steal your data, he WOULD use it to extort you, and on top of it all he WOULD spam your ass to death.
General purpose voice recognition* doesn't work (Score:2)
If I can't understand a Geordie, let alone a god damn American, how the fuck will a computer, I doubt the Africans/Asians (who despite above claims probably speak the queens English a damn sight better than most of you guys (assuming slashdot is populated by gorram Americans)) will get it spot on, but their internal algorithms have had a data set of at least 18 years to train on, this beats any automated system!. Voice recognition* has its places (e.g the iPhone does it right), but transcription is not one
ridiculous (Score:2)
they dont even need to have speech recognition, they just need to recognize when a few word is spoken and have people listen to individual words.
Virtual Labor (Score:2)
A theme in the film is Virtual Labor - robots of the future will really be remotely operated by cheap overseas labor. SpinVox is doing similar kind of things, but unlike Mechanical Turk has the factore of outsourcing to the low-wage regions.
What about Google Voice transciption? (Score:2)
What about Google Voice transciption? It seems to do such a good job I always suspected it was Google's private version of Amazon Mechanical Turk.
Are image ATMs automatic? (Score:2)
I've been wondering about "image ATMs", which accept checks for deposit, imaging them. I've had one correctly accept a check with the amount handwritten in cursive. I suspect that at least the hard cases are being referred to humans for recognition.
Official response (Score:3, Informative)
South African and English (Score:2)
You guys do know that many, many South African's speak English as their first language, right ?
Most South Africans I know speak better English with fewer weird accents than the UK population. I've never had to try and understand what is meant by 'Arwight mate, innit' while talking to a saffa :)
It was obvious (Score:2)
Re:But it's not crazy (Score:4, Insightful)
What?
No.
Their service says that they keep user supplied data in house. They do not.
Their service says that they use advanced technological means to do the transcription. They do not.
How on earth do you take that to mean 'their service does what it says'?
You are wrong.
Re:But it's not crazy (Score:5, Informative)
But it also knows what it doesn't know and is able to call on human experts for assistance.
http://www.spinvox.com/how_it_works.html [spinvox.com]
Re: (Score:3, Informative)
Re: (Score:2, Funny)
Re: (Score:3, Informative)
Regardless of their lying about it, the actual 'method' itself is technically sound.
Re: (Score:3, Interesting)
And now you're saying that people who barely speak or understand English, let alone the subtlties of the language, being paid to transcribe English, is 'technically sound' and 'the best way to do it'? ...
Frankly, I'm not sure anymore if you're serious, or just being sarcastic.
Next you'll probably tell me "Oh, see that motherboard made of flammable wood? Regardless of it's flammability, it's the best flame-proof way to make a motherboard."
Re:But it's not crazy (Score:4, Informative)
Re: (Score:2)
It's weird watching the moderating of my original post fluctuate - it went form +5 insightful to +1 overated-flamebait in the space of half an hour. I can't for the life of me figure that one out.
I think I'll just stick to typing dictated notes...
Re: (Score:2)
I considered Mr Mista troll-like because of the pattern of constructing strawmen from my points, followed by completely unrelated rhetorical baiting. I could be wrong - he might be earnest; wouldn't be the first time someone was wrong on the internet.
Re: (Score:3, Insightful)
Uhh... Somehow I got confused about who said what - disregard above post or mod it down into obscurity where it belongs...
Re:But it's not crazy (Score:5, Informative)
And now you're saying that people who barely speak or understand English, let alone the subtlties of the language, being paid to transcribe English, is 'technically sound' and 'the best way to do it'? ...
I think it's more likely that these people speak better, more grammatically correct English than the average Brit or American.
I find it likely that the majority of these people who worked in these centers are young, recent college/university graduates who are doing this because they couldn't find another well paying job. This isn't a bunch of Angolans or Indonesians. We're talking about South Africans and Filipinos. The well educated South African and Filipino speaks, reads and writes excellent English.
For that matter, the same is probably true of Egyptians. Though I can't say that with any certainty because I don't know too many Egyptians.
Re: (Score:2)
Nobody has a funny accent when they type. Other than the accent and occasional odd phrasing (which would never come out in a transcription service - english to english is pretty frickin easy), most of these people speak excellent English, particularly since it is a second language for all of them.
Hell, my last college english class was taught by an Indian immigrant. She had a heavy accent, but her English was very precise and correct. The only annoying thing was she liked the word "rubric", and rolled he
Re: (Score:2)
"Able to speak grammatically correct ${LANGUAGE} when in a classroom environment, there's little background noise and everything is being enunciated nice and slowly" is a thousand times removed from "Able to hold a conversation in ${LANGUAGE} as well as any native ${LANGUAGE} speaker".
Re: (Score:2, Insightful)
I think it's more likely that these people speak better, more grammatically correct English than the average Brit or American.
Since when has the average person spoken grammatically correct British/American English? What about slang, regional vocabulary, accents?
I think it'd be hard for a native speaker to translate in a lot of cases e.g. north/south in the UK.
Re: (Score:2)
It still beats Babelfish.
Though... considering the wording of the translations, one has to wonder...
Re: (Score:2)
It's the difference between "what" and "how". It does WHAT it says (transcribes voice mail to text), but not HOW they say (employees instead of algorithms, cheap countries instead of home).
It's quite silly actually. I personally don't care much about the "how" (as long as my "data" is indeed anonymized), so I don't understand why they'd lie about it.
Re: (Score:2)
And here's exactly the problem: Privacy, secrecy and anonymity.
A computer does not care what he translates. It doesn't 'read' (or 'listen to') your message. It picks it apart logically on semantics and assembles a message in another language based on this.
Humans read, understand, translate and write. And the 'understand' part is the one with the security and confidentality issue. And don't tell me "they're some dumb idiots in Backwater Nowhere, they don't understand what's written and they don't care". Says
Re: (Score:2)
Except that b and v have a different sound..
Stuff like I scream/ice cream would be a problem though.
Re: (Score:2)
Their service says that they use advanced technological means to do the transcription. They do not.
Any sufficiently advanced AI is indistinguishable from a Pakistani sweat shop worker.
Re: (Score:2)
What?
No.
Their service says that they keep user supplied data in house. They do not.
Their service says that they use advanced technological means to do the transcription. They do not.
How on earth do you take that to mean 'their service does what it says'?
You are wrong.
You seemed to have forgotten to attach Proof.zip before hitting send.
You say "They do not" as if one could tell from the situation that they do or do not.
Knowing how companies lie, you clearly can not choose the wine in front of them.
Knowing how disgruntled ex-employees lie, you can clearly not choose the wine in front of you.
Re: (Score:2)
It may be they lied about keeping user supplied data in house, and they may have implied that they used advanced technological means to do the transcription, but if their service does what it says I can't blame them for using human labour to do the transcription.
I don't know... an unethical service, an unscrupulous company, a management with the lack of business sense to realize this is a public relations disaster, a fiscally untenable platform, and (possibly) opens them up to legal action... I'd call that crazy.
Sure, the technology works, but the whole idea is preposterous. Who transcribes the workers' voice mail? Or is voice mail transcription reserved for the upper class? Surely it's not such an elite service to warrant that treatment, which indicates tha
Re:But it's not crazy (Score:5, Funny)
>>>Human brains remain the only high performance computer manufactured with unskilled labour.
I object! It takes a lot of skill to satisfy today's demanding women. And what happens if you lack that skill? They'll just jump ship to some other guy's bed. Unskilled labor indeed. It takes a lot of skill to convince Miss Prissy to let her guard down, bribe her with a 50,000 dollar wedding, remove the diaphragm, and let you impregnate her.
No I'm not bitter.
Although I do have this gnawing pain in my gut until I can taste the bile rising up my throat and into my mouth. Well. Maybe I'm a little bitter. Or else I just have heartburn; anybody have a TicTac?
Re: (Score:3, Insightful)
A brain comes from moron woman even if it has the IQ of a contraceptive sponge.
Re: (Score:2)
You're doing it wrong. I had my wedding paid for by my wife, and nevertheless 3 kids. ,-)
Re: (Score:2, Funny)
*sigh* What happened to good ol' rape?
Ok, snide and tasteless comments aside. The only thing it takes a lot of is time. Women can get quite desperate when they hear their biological clock tick away. Unfortunately that happens after their best before date...
Re:But it's not crazy (Score:4, Insightful)
Where are we as a species, if making babies is a fetish?
Re: (Score:3, Insightful)
From a purely evolutionary point of view? We're quite sane, if you ask me. If the idea of impregnating your woman gives you a boner, I'd say it's about as close to the original idea behind sex as it can be.
Don't tell anyone but, hey, getting her pregnant was the idea behind fucking. I know, it kinda changed in the meantime, but originally, that was the plan.
Re: (Score:2, Interesting)
Re: (Score:2)
Controlling who has access to the private messages in a particular way is part of what they said they would do for the people using the service, so, no, if they said they were doing automated transcription and keeping all personal data inside Europe
Re: (Score:2)