h a l f b a k e r y
Puts the 'reek' in Eureka
idea:
add, search, annotate, link, view, overview, recent, by name, best, random
meta:
news, help, about, links, report a problem
account:
Browse anonymously,
or get an account
and write.
Login
Create account.
|
|
|
This speech recognition driven text display would try to
represent spoken language as acurately, succinctly and
readably as possible. The text could be colored on a
spectrum from blue to red according to the loudness or
softness, or emotive tone of a voice. The text could climb or
fall --
within a small space on the page according to pitch, and
bunch up or spread out according to meter. The system could
be networked and self-improving and google-like in that it
would decide which attributes were most effective by how
much they were used. The goal of the system would be to
better represent tone of voice in text.
All-so, foe-net-tick spel-ling and an em-fah-sis on rith-mick sill-
lih-ba-za-shun wood bee help-full.
HB archives: Math-modified Language
Math-modified_20Language
Circa Jan 02. Tertially related but much annotated and debated idea from [RayfordSteele]. [bristolz, Mar 16 2005]
[link]
|
| |
I'm pretty sure I've seen some of these ideas on the
halfbakery but I can't find them by subject. |
|
| |
Nah, don't like it. What you describe, I think, is infinitely better accomplished via good writing skills. |
|
| |
Writing and speaking skills are, in my view, far more important than any other educational discipline unless you require a specialised field of study for the purpose of employment / income generation. |
|
| |
"Dumbing down" these vital skills simply serves the will of the mediocrist fifth column afflicting our society. |
|
| |
If you were to remove the last bit about phonetic spelling (unnecessary to the idea, in my opinion), then what you're describing is not quite so much a speech mark-up language, but rather a speech style-sheet language. |
|
| |
Example: "Oh, What a marvellous idea."
Sarcasm, or truth? How would it be marked up, either way? |
|
| |
The text of "Oh, What a marvellous idea." could be
extended or contracted, raised or lowered, increased or
decreased in font, and colored according to variables from
a voiceprint of
someone saying, "Oh, What a marvellous idea." |
|
| |
"OHH, what a M A R V O L O U S idea!" |
|
| |
...HHH........... .............ARRRR...... ........De...
O... ...HHHH...what.a.MAA.... .....VOLOUS...i....a!
|
|
| |
--except more subtley, and without the periods. |
|
| |
The phonetic spelling is a real anchor weight to reading as we people tend to read by word shape, or even phrase shape, recognition (much as those who read music well see the overall shape and pattern of music rather than the individual notes). |
|
| |
Making the language as expressive as musical notation is, in my opinion, a giant reduction in expressiveness. The musical compass is very small compared to that of language, at least English. |
|
| |
How about representing the intonation,
pitch, rhythm and volume of speech as
... (I know this sounds crazy) ....an
audio file? |
|
| |
Only that it's not accessible all at once like text. You can scan a testimony visually better than auditorily. This would be a way of getting lie-detector-like information all in one place at one time. |
|
| |
Actually, THIS idea might solve the problem of getting a computer to read a book in an expressive way! Humans can figure out loudness and tone of voice and pacing from context and punctuation, but a computer needs other, more specific cues. You could have it read blue text in a sad tone of voice, red text in an angry tone, I guess purple could be lustful... a whole new meaning to purple prose! |
|
| |
I like the idea, as a way to standardise vowel sounds and pitch changes. It might allow people to communicate across languages due to the standardisation - however, in many ways - i.e. within speech recognition systems, this is kind of baked in that the recorded sounds are tokenised and stored as discrete syllable patterns which are then compared to a database of known pattern/word sets. It's just that this internal notation has been designed for computers rather than for people. |
|
| |
If memory serves me correctly, there used to be a standard phonetic notation used in the early days of speech synthesisers. I can remember typing strange looking syllables into the machine in order to try and get it so sound like it was from 'up north' etc. |
|
| |
So, as much as I like the idea (providing a way to visually read voice transcripts in a way that keeps as much of the voice patterns intact during the conversion) It's use is reasonably limited, and is just a refinement of something that's been around for 15 years(the first time I remember playing with a speech synthesiser was late 80's) |
|
| |