The reason they haven't developed verbal speach is because they don't have the physical ability to produce the same sounds tha humans can. Breeding apes for intelligence won't ever produce a specimen that is able to speak english or any other human lanaguge, nor would any scientist ever expect it to.
Quite. After a few months of life, the larynx of a human infant descends from just behind the back of the mouth to further down the throat. This change has its ups and downs. One disadvantage is that baby can't breathe and eat at the same time any more, but then again, because their rate of growth is slowing somewhat, the need to do so is less pressing. Other things are becoming more important, so babies move from eatin', shittin' and sleepin' towards trying to take in ever-increasing amounts of information about their environment.
The advantage of the descending larynx is that junior is able to speak as a result of increased flexibility of movement of the various parts of the vocal tract and an enlarged pharynx. That right-angled vocal tract of ours is unique to humans, and that's why we're the only animals who can form sounds in the way we do. No amount of breeding's going to produce this ability in other apesn no matter how intelligent they become (well, not quite no amount, but it'd need a hell of a lot of trial-and-error to get near it...).
Vowels seem to be on the way out anyway, thanks to txt spk. We're gonna end up like Arabic or something (which is no bad thing, because they way you form words in Arabic is fantastically elegant and you nerd types would love it).
Re:U.S. spelling has the original forms
on
Flavor vs. Flavour
·
· Score: 1
Are you "literate" if you talk in 13375p33k and emoticons? 'Cos if not, how the hell can you get a rating that high?
More importantly, a US pint is a bit smaller than a UK one, leading to a general feeling that you're getting subtly fleeced every time you go to the bar...
There's none more closed-minded than the teenybopper weaned on Smash Hits and The Box. You try getting them to listen to something that you know they'd like, because basically it's a far better-realised take on what they "choose" to listen to themselves. It's like trying to get a child to eat their greens. They'll dismiss it because they haven't heard of it, because MTV (or whoever, I know MTV doesn't actually play any music these days) hasn't told them they should like it. How's that any different to some indier-than-thou NME-reading arsehole that insists that anything in the charts must be shite?
Duke Ellington had it right once and for all: There's only two types of music: Good and bad. And sure, it's largely opinion, but there are absolutes in this world: Westlife are undeniably shite, and Ella Fitzgerald is undeniably a genius.
Yeah, moderators, I know it's offtopic. But so what? There's only two types of posts: Good and bad...
...Some Austrian guy threw himself headlong into France, caught them all by surprise with the audacity and speed of it......And there wasn't a white flag of surrender in sight.
Do you think that's what I saw?
on
The Big Kerplop
·
· Score: 1
Now, I'm a sensible, sane individual, most of the time anyway. And I'd only had a couple of beers at the time. But a couple of weeks ago I, and two others in my company, saw something we were initially at a loss to explain.
I was at the legendary Glastonbury Festival (now, I know some of you are rolling your eyes already, but I can assure you, drugs were not involved here), sitting outside my tent in the wee hours of the morning. We were right near the edge of the site, about 50 yards from the fence, sort of on the crest of a gentle hill. To our left was part of the festival, though the main stages were behind us as we looked. To the right was "nowt but fields".
Anyway, from the left, at a low altitude, maybe 100 feet or less, and about 50 yards away at their closest point, came three steadily-glowing orange lights, like dim lightbulbs. They were moving relatively slowly - I'd guess they were going slower than a man could run. They were all individual points of light, rather than joined, and they moved from left to right, straight and level. One was a bit behind the other two, and when they came to about the 11 o'clock position, it started to accelerate, and overtook them 10-15 seconds later. They kept going until seemingly sinking down at the 3 o'clock position - though it was impossible to be sure that they actually did sink down, it might've just looked that way because of the perspective. There was no noise whatsoever from the lights.
Anyway, myself and the other two people I was with were pretty spooked, but after discussing it a bit, we decided that it was probably these paper lanterns with candles in. Thing is, though, there wasn't really any wind to speak of - and how would the wind make one accelerate past the others anyway? The fact that it came from the direction of the Green Fields makes it more likely that people were responsible - if it had come the other way, we really would have been stumped for an explanation.
So if it was the lantern things, I can vouch for their efficacy. If it wasn't, well...
Re:Sympathetic vibration is fun.
on
dB Drag Racing
·
· Score: 1
Yeah, but Leftfield also managed to get themselves barred from the Brixton Academy for knocking the plaster off the walls with their sound system. It's not exactly an off-the-shelf job.
Mind you, compared to Dillinja's Valve system, it's like an old hand-powered gramophone...
Re:The REAL problem is...
on
dB Drag Racing
·
· Score: 1
Well, it's legal, like. But a car that weighs 10,000lbs isn't going to go anywhere fast, and it's hardly going to corner like a go-kart either. Plus you'd need a tanker following you around just to keep it in fuel. I think that's their point.
Then again, some people seem to think that sort of thing's clever.
That's more likely to be a consequence of the design of the base stations than anything else. Let's face it, there isn't much demand for network capacity 40 floors up. The transmitters are pointing towards the ground where people in a big city need them.
In more open country, like along the side of a motorway, the transmitters will be expected to operate over a wider area, and so you're much more likely to be able to talk to them from above.
Yes, you're quite right. The US thought sacrificing historical artifacts thousands of years old was a fair price to pay for cheap energy too, so it's perfectly sensible...
All geeks are megalomaniacs at heart. They all love to tinker around with stuff - until it's exactly the way that they want it. Because obviously that's the right way for it to be. So if you were the world's most powerful geek, wouldn't you get a real kick out of forcing your vision of How It Should Be on the world?
I mean, just look at the contempt in which users are held by the average sysadmin - all they do's break his (it's always a bloie) lovely systems which he's handcrafted so painstakingly. The fact that when it comes down to it he's there to serve them is neither here nor there - he sees it as a great sacrifice to let them anywhere near his kit.
Consider this: Call up Sprint, AT&T, MCI, etc. and ask them what their price is for a DS3, including loops. You'll probably end up with something around $500/month. per Mbps. Negotiate a bit and you might get below that a bit - maybe even down around $200/mo. per Mbps if you buy enough capacity. Now, turn around and sell that same sustained Mbps/month for $35-$40 to a cable modem user.
Good business? Don't forget, you've got local transmission, switching/routing, customer support, billing, fixed costs/backoffice, equipment capital & depreciation, etc. So, for $500/mo/Mbps or so, you're making big profits on that $35/mo. customer?
Yeah, but remember this bandwidth that you, as an ISP, pay for and are assured by your backbone provider that you will be allocated at a sustained rate, is usually contended at your customer's end at a minimum of 20:1, and here in the UK you're more likely to be looking at 50:1. So your $35/month. figure gives you an income of at least $700/month if you're feeling reasonably generous towards your customers, and $1750/month on the business model in place here. That should cover your overheads quite nicely.
If you're seeing a sustained throughput anywhere near the maximum possible on your DSL line, then you're a very lucky person, and you should make hay while the sun shines, 'cos you just know that soon enough some local idiot's going to sign up and start hoovering animal porn and bangin' hard house choons, and knock your throughput for six.
And anyway, your wider point that they'll pass on the cost tp bandwidth hogs as soon as they can get away with it's a fair one. They will. But it wouldn't do to put people off just yet, as NTL found out to their cost in the UK lately - they seem to have quietly dropped their bandwidth cap.
Well, I can't claim to know an answer to this, but can tell you this. In a conversation with a guy who worked for one of Britain's biggest ISPS (BT Openworld since you asked), whose role related to network security and the like, I was told that 70% of the traffic across their network was porn.
Now, bear in mind that this is a network catering primarily for residential customers, so this figure would be balanced out by business use when you look at the wider picture for the Internet in general, but there you have it. Also, he didn't seem to be taking P2P into account with that figure - I think he was referring to Web traffic only.
We have a presntation suite where I work fenced in with those things. If you can picture it, there's a large room with a kind of eye-shaped area in the middle surrounded by glass partitions. Round the outside of the room is a variety of demonstration systems, interactive displays, etc. Anyway, you bring your victims into what appears to be a frosted-glass room in the middle, begin your talk about the future of communications with some sort of "imagine if you could..." spiel, then whack a button, the glass instantly goes transparent and they're surrounded by displays of the very things you've just been talking about. Knocks them dead EVERY time.
God knows what it costs, but it's seriously cool. Gotta get 'em for my house one day, beats blinds and curtains hands down...
I doubt that anything can be 100% successful given that human perception is not totally error-free.
True enough. Maybe I should have said "never be as successful as humans".
Speech recognition HMM models are based on contextual realizations of phonemes (allophonic model). This takes into account coarticulation. Same technique applies for generation: in-context phonemes are used.
Yes, but they're not contextual enough. Triphones, which to my knowledge is the norm, just isn't a wide enough catchment area. Furthermore, I would question the logic of segmenting at all, for the reasons described previously.
I have a master's in linguistics, specialising in speech processing and the like, and I don't really believe in phonemes.
In the beginning, there was the word. And the word was spoken. A long, long time later came writing. Most early forms of writing seem to have been pictographic. Eventually that started to be a bit too complicated for most, and somewhere along the line we switched to trying to represent the sounds of the words that we used. These writing systems had to be sort of retrofitted onto the sounds we used, and so they were never going to amount to a perfect transcription of the sounds used. Huge alphabets quickly become unwieldy, and while there is a great deal of variation between languages in terms of how they deal with these issues, in most cases sounds end up being shoehorned into one category or another - "oh, that's sort of a/t/, I'll write it down like that". You know yourselves how often words in English bear no relation to their spoken forms.
Anyway, a long time after that, people got interested in phonetics. Conditioned as we were into thinking of words as collections of letters, along came the concept of the phoneme, which, as somebody said above, is the smallest individual unit of speech which can be distinguished from other such units. Phoneticists set about mapping all the sounds of all the languages in the world to phonemes, and we got the international phonetic alphabet.
Later still, we managed to invent machines which allowed us to analyse sound spectra. Run a spoken utterance through one of these and what you'll see most certainly isn't a succession of distinct sounds. Truth is, our brain does so much work on the raw sound that our perception of the sounds is entirely different from the reality. "Phonemes" don't just start and end neatly - they overlap massively. A single vowel can affect maybe the preceding four segments and the following six because of the effects of reconfiguring your vocal tract. The next sound might do the same. And the next one... As you can probably imagine, it's a pretty messy picture really. Believe me, I have suffered greatly trying to segment voice spectra by hand.
The point of all this is that when we started speaking yonks ago, we were making use of the vocal tract nature (God, natural selection, take yer pick, I don't want to get into an argument about it) gave us. We weren't thinking of phonemes and stuff, we were just making noises subject to the limitations of the equipment we had. The notion that this is a nice, ordered system of sounds is an artifical one imposed by us in an attempt to make sense of it all, and it amounts to an expanded version of an oversimplified system (the alphabet). Now, we all know what happens with lossy compression...
Simply drawing lines down the spectrogram in the name of making it easier to work with just throws away subtlety, so that when you use a phoneme-based TTS system you get a series of disjointed sounds with perhaps some token effort at coarticulation (i.e. the phenomenon of overlapping sounds described above), and it's always going to sound awful. The consequences for speech recognition are much worse (sure, your hidden Markov model-based systems working with sequences of two or three phonemes are pretty effective, but they'll never be 100% successful in my opinion).
In short, what you have here's an engineer's approach to art. It's like taking a painting by your favourite artist and turning it into a 256-colour bitmap, then analysing the result and trying to make new paintings in the same style.
Slashdot Mirror
Sometimes people just don't want to talk. It's as simple as that.
Sorry, but that sort of illiteracy's bloody annoying.
All very admirable, but something in your tone tells me it's been some time since you last dipped your wick...
The reason they haven't developed verbal speach is because they don't have the physical ability to produce the same sounds tha humans can. Breeding apes for intelligence won't ever produce a specimen that is able to speak english or any other human lanaguge, nor would any scientist ever expect it to.
Quite. After a few months of life, the larynx of a human infant descends from just behind the back of the mouth to further down the throat. This change has its ups and downs. One disadvantage is that baby can't breathe and eat at the same time any more, but then again, because their rate of growth is slowing somewhat, the need to do so is less pressing. Other things are becoming more important, so babies move from eatin', shittin' and sleepin' towards trying to take in ever-increasing amounts of information about their environment.
The advantage of the descending larynx is that junior is able to speak as a result of increased flexibility of movement of the various parts of the vocal tract and an enlarged pharynx. That right-angled vocal tract of ours is unique to humans, and that's why we're the only animals who can form sounds in the way we do. No amount of breeding's going to produce this ability in other apesn no matter how intelligent they become (well, not quite no amount, but it'd need a hell of a lot of trial-and-error to get near it...).
Vowels seem to be on the way out anyway, thanks to txt spk. We're gonna end up like Arabic or something (which is no bad thing, because they way you form words in Arabic is fantastically elegant and you nerd types would love it).
Are you "literate" if you talk in 13375p33k and emoticons? 'Cos if not, how the hell can you get a rating that high?
More importantly, a US pint is a bit smaller than a UK one, leading to a general feeling that you're getting subtly fleeced every time you go to the bar...
Balls.
There's none more closed-minded than the teenybopper weaned on Smash Hits and The Box. You try getting them to listen to something that you know they'd like, because basically it's a far better-realised take on what they "choose" to listen to themselves. It's like trying to get a child to eat their greens. They'll dismiss it because they haven't heard of it, because MTV (or whoever, I know MTV doesn't actually play any music these days) hasn't told them they should like it. How's that any different to some indier-than-thou NME-reading arsehole that insists that anything in the charts must be shite?
Duke Ellington had it right once and for all: There's only two types of music: Good and bad. And sure, it's largely opinion, but there are absolutes in this world: Westlife are undeniably shite, and Ella Fitzgerald is undeniably a genius.
Yeah, moderators, I know it's offtopic. But so what? There's only two types of posts: Good and bad...
...Some Austrian guy threw himself headlong into France, caught them all by surprise with the audacity and speed of it... ...And there wasn't a white flag of surrender in sight.
You're as big a dickhead as him, then.
Now, I'm a sensible, sane individual, most of the time anyway. And I'd only had a couple of beers at the time. But a couple of weeks ago I, and two others in my company, saw something we were initially at a loss to explain.
I was at the legendary Glastonbury Festival (now, I know some of you are rolling your eyes already, but I can assure you, drugs were not involved here), sitting outside my tent in the wee hours of the morning. We were right near the edge of the site, about 50 yards from the fence, sort of on the crest of a gentle hill. To our left was part of the festival, though the main stages were behind us as we looked. To the right was "nowt but fields".
Anyway, from the left, at a low altitude, maybe 100 feet or less, and about 50 yards away at their closest point, came three steadily-glowing orange lights, like dim lightbulbs. They were moving relatively slowly - I'd guess they were going slower than a man could run. They were all individual points of light, rather than joined, and they moved from left to right, straight and level. One was a bit behind the other two, and when they came to about the 11 o'clock position, it started to accelerate, and overtook them 10-15 seconds later. They kept going until seemingly sinking down at the 3 o'clock position - though it was impossible to be sure that they actually did sink down, it might've just looked that way because of the perspective. There was no noise whatsoever from the lights.
Anyway, myself and the other two people I was with were pretty spooked, but after discussing it a bit, we decided that it was probably these paper lanterns with candles in. Thing is, though, there wasn't really any wind to speak of - and how would the wind make one accelerate past the others anyway? The fact that it came from the direction of the Green Fields makes it more likely that people were responsible - if it had come the other way, we really would have been stumped for an explanation.
So if it was the lantern things, I can vouch for their efficacy. If it wasn't, well...
Yeah, but Leftfield also managed to get themselves barred from the Brixton Academy for knocking the plaster off the walls with their sound system. It's not exactly an off-the-shelf job.
Mind you, compared to Dillinja's Valve system, it's like an old hand-powered gramophone...
Well, it's legal, like. But a car that weighs 10,000lbs isn't going to go anywhere fast, and it's hardly going to corner like a go-kart either. Plus you'd need a tanker following you around just to keep it in fuel. I think that's their point.
Then again, some people seem to think that sort of thing's clever.
OS XXIII will probly be hitting the shelves, so what's the point?
That's more likely to be a consequence of the design of the base stations than anything else. Let's face it, there isn't much demand for network capacity 40 floors up. The transmitters are pointing towards the ground where people in a big city need them.
In more open country, like along the side of a motorway, the transmitters will be expected to operate over a wider area, and so you're much more likely to be able to talk to them from above.
Yes, you're quite right. The US thought sacrificing historical artifacts thousands of years old was a fair price to pay for cheap energy too, so it's perfectly sensible...
All geeks are megalomaniacs at heart. They all love to tinker around with stuff - until it's exactly the way that they want it. Because obviously that's the right way for it to be. So if you were the world's most powerful geek, wouldn't you get a real kick out of forcing your vision of How It Should Be on the world?
I mean, just look at the contempt in which users are held by the average sysadmin - all they do's break his (it's always a bloie) lovely systems which he's handcrafted so painstakingly. The fact that when it comes down to it he's there to serve them is neither here nor there - he sees it as a great sacrifice to let them anywhere near his kit.
Consider this: Call up Sprint, AT&T, MCI, etc. and ask them what their price is for a DS3, including loops. You'll probably end up with something around $500/month. per Mbps. Negotiate a bit and you might get below that a bit - maybe even down around $200/mo. per Mbps if you buy enough capacity. Now, turn around and sell that same sustained Mbps/month for $35-$40 to a cable modem user.
Good business? Don't forget, you've got local transmission, switching/routing, customer support, billing, fixed costs/backoffice, equipment capital & depreciation, etc. So, for $500/mo/Mbps or so, you're making big profits on that $35/mo. customer?
Yeah, but remember this bandwidth that you, as an ISP, pay for and are assured by your backbone provider that you will be allocated at a sustained rate, is usually contended at your customer's end at a minimum of 20:1, and here in the UK you're more likely to be looking at 50:1. So your $35/month. figure gives you an income of at least $700/month if you're feeling reasonably generous towards your customers, and $1750/month on the business model in place here. That should cover your overheads quite nicely.
If you're seeing a sustained throughput anywhere near the maximum possible on your DSL line, then you're a very lucky person, and you should make hay while the sun shines, 'cos you just know that soon enough some local idiot's going to sign up and start hoovering animal porn and bangin' hard house choons, and knock your throughput for six.
And anyway, your wider point that they'll pass on the cost tp bandwidth hogs as soon as they can get away with it's a fair one. They will. But it wouldn't do to put people off just yet, as NTL found out to their cost in the UK lately - they seem to have quietly dropped their bandwidth cap.
Well, I can't claim to know an answer to this, but can tell you this. In a conversation with a guy who worked for one of Britain's biggest ISPS (BT Openworld since you asked), whose role related to network security and the like, I was told that 70% of the traffic across their network was porn.
Now, bear in mind that this is a network catering primarily for residential customers, so this figure would be balanced out by business use when you look at the wider picture for the Internet in general, but there you have it. Also, he didn't seem to be taking P2P into account with that figure - I think he was referring to Web traffic only.
We have a presntation suite where I work fenced in with those things. If you can picture it, there's a large room with a kind of eye-shaped area in the middle surrounded by glass partitions. Round the outside of the room is a variety of demonstration systems, interactive displays, etc. Anyway, you bring your victims into what appears to be a frosted-glass room in the middle, begin your talk about the future of communications with some sort of "imagine if you could..." spiel, then whack a button, the glass instantly goes transparent and they're surrounded by displays of the very things you've just been talking about. Knocks them dead EVERY time.
God knows what it costs, but it's seriously cool. Gotta get 'em for my house one day, beats blinds and curtains hands down...
And hearing John Inman camping it up inside their head in an "Are You Being Served?" stylee?
"Iiiiii'm freeeeeeeeee!......"
The plural's "genera". Just for future reference, like.
I doubt that anything can be 100% successful given that human perception is not totally error-free.
True enough. Maybe I should have said "never be as successful as humans".
Speech recognition HMM models are based on contextual realizations of phonemes (allophonic model). This takes into account coarticulation. Same technique applies for generation: in-context phonemes are used.
Yes, but they're not contextual enough. Triphones, which to my knowledge is the norm, just isn't a wide enough catchment area. Furthermore, I would question the logic of segmenting at all, for the reasons described previously.
I have a master's in linguistics, specialising in speech processing and the like, and I don't really believe in phonemes.
/t/, I'll write it down like that". You know yourselves how often words in English bear no relation to their spoken forms.
In the beginning, there was the word. And the word was spoken. A long, long time later came writing. Most early forms of writing seem to have been pictographic. Eventually that started to be a bit too complicated for most, and somewhere along the line we switched to trying to represent the sounds of the words that we used. These writing systems had to be sort of retrofitted onto the sounds we used, and so they were never going to amount to a perfect transcription of the sounds used. Huge alphabets quickly become unwieldy, and while there is a great deal of variation between languages in terms of how they deal with these issues, in most cases sounds end up being shoehorned into one category or another - "oh, that's sort of a
Anyway, a long time after that, people got interested in phonetics. Conditioned as we were into thinking of words as collections of letters, along came the concept of the phoneme, which, as somebody said above, is the smallest individual unit of speech which can be distinguished from other such units. Phoneticists set about mapping all the sounds of all the languages in the world to phonemes, and we got the international phonetic alphabet.
Later still, we managed to invent machines which allowed us to analyse sound spectra. Run a spoken utterance through one of these and what you'll see most certainly isn't a succession of distinct sounds. Truth is, our brain does so much work on the raw sound that our perception of the sounds is entirely different from the reality. "Phonemes" don't just start and end neatly - they overlap massively. A single vowel can affect maybe the preceding four segments and the following six because of the effects of reconfiguring your vocal tract. The next sound might do the same. And the next one... As you can probably imagine, it's a pretty messy picture really. Believe me, I have suffered greatly trying to segment voice spectra by hand.
The point of all this is that when we started speaking yonks ago, we were making use of the vocal tract nature (God, natural selection, take yer pick, I don't want to get into an argument about it) gave us. We weren't thinking of phonemes and stuff, we were just making noises subject to the limitations of the equipment we had. The notion that this is a nice, ordered system of sounds is an artifical one imposed by us in an attempt to make sense of it all, and it amounts to an expanded version of an oversimplified system (the alphabet). Now, we all know what happens with lossy compression...
Simply drawing lines down the spectrogram in the name of making it easier to work with just throws away subtlety, so that when you use a phoneme-based TTS system you get a series of disjointed sounds with perhaps some token effort at coarticulation (i.e. the phenomenon of overlapping sounds described above), and it's always going to sound awful. The consequences for speech recognition are much worse (sure, your hidden Markov model-based systems working with sequences of two or three phonemes are pretty effective, but they'll never be 100% successful in my opinion).
In short, what you have here's an engineer's approach to art. It's like taking a painting by your favourite artist and turning it into a 256-colour bitmap, then analysing the result and trying to make new paintings in the same style.
Got you to read it, though, didn't it? Would you have been as quick to click if the article said "really small webserver-on-a-chip unveiled"?