h a l f b a k e r yIs it soup yet?
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
| |
Ever listened to a modem transcieve data? Ever load a program off a cassette tape? |
|
| |
Don't necessarily need pitch. For example, not too hard to tell when a car needs a new muffler: that's one bit of information right there. |
|
| |
Pretty much baked (Morse code) |
|
| |
rmutt: your car outputs ASCII?? |
|
| |
magnificat: that's pulse, not tone. |
|
| |
Good point, doc... i see what he means now |
|
| |
If the speaker is capable of using other acoustic qualities than just tone - if it has the ability to produce the various sounds that we recognise as phonemes - why limit it to communicating through a one-dimensional medium? Apart from the fact that precisely recognising 256 *discrete* tones within a spectrum of audibility would probably be extremely difficult even for a native speaker of a language such as Chinese, where pitch / intonation is an essential element of semantic meaning, and all but impossible for English speakers used to using tone in a fairly slip-shod manner mainly to add affective colouring, you are also limited by the length of time the sound has to be audible for it to be registered and recognised as a phoneme. And there's a lot of 'filling in the gaps' that takes place in our phonetic perception of sound (which is quite distinct from just 'hearing', according to some researchers) that couldn't be relied on if the sound was a binary data-stream rather than a chunk of a known language. For anyone other than a pitch-perfect, trained musician, I suspect your tonal output would be as indecipherable as a sequence of circles of various shades of red scrolling rapidly across a computer screen, where each shade represents one of 256 words selected arbitrarily as the basic elements of a code. |
|
| |
It seems to me it would be much easier to simply ascribe phonemes, as they are defined systematically within the International Phonetic Alphabet, to the 0-255 values of extended ASCII. There should be enough discrete, immediately-recognisable sounds - /sh/, /ch/ as in 'loch', /th/ as in 'the', /th/ as in 'thin', and innumerable types of clicks and ejectives as used in other languages - to cover at least 0-127. Volume or tonal variation could be used as well, of course, but you'd want big, broad distinctions; 'deep' /sh/ and 'high' /sh/ will be a hell of a lot easier to distinguish than two only very slightly different musical notes. |
|
| |
Anyhoo, my point is just that converting text to speech would be a piece of piss - efficient and correct - if the text was rendered in a phonetics-based character set to begin with. Presuming it is text you want to 'read out' (rather than, say, a sequence of IEEE-format floating single-point numbers), although the IPA gives a different phonetic value to some of the characters in the Roman alphabet, I'd say that "t-hiss way off spee-a-kin-gi" is much easier to understand than "eeeeeee'e'e'e'e'e'e'EEEEeEeEeEeEeEeeeeeeeeeE'E'E'E'E'EE" |
|
| |
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶¶
There you go... |
|
| |