I forwarded the preloader info. to Puppyflowers.
The poems are actually one of two sets written with MS Reader avatars back... when they came out, in a longer project that is or was based on 1960s children's pre-reading perception exercises (I have some great old school mimeo workbooks, etc.).
So those at Puppyflowers were written, then translated using babelfish. I recorded them being read by the MS Reader avatars in their language, and let the voice recognition software homolinguistically translate. Then I made them into dialogs with the male and female avatars in each language (which translate slightly differently). The longer part has the stages in it.
The others were me reading tongue twisters into the voice recognition software, then reading the results into the voice recognition software, then reading the results.... Our pet parrot was going wild so that got in there too.
The avatars section really reminded me of post ww2 Europe, because the avatars were French and German.
The thing about most of the results of the perception exercises (which are based on nouns) is the way they conjure commercialism. This was the surprising thing about the untrained voice recognition as well -- it kept throwing in brand names, aka "Eddie Bauer" etc.
Thus, in your voice recognition poems from tv, movies, etc., I can't help but notice that self-reflexive quality that these machine written texts seem to have -- mechanical themselves, they bring out the technical aspects of the material -- editing, scene setting / voice over in film; types of rhetorical gestures in speechmaking, etc.
The other project is of course vast. Of course, I gravitated to
Inside Rosario’s two bedroom home (bought with her husband, who after two years of marriage disappeared while swimming in the ocean) are hundreds of thousands of photographs pasted, glued, tacked to the walls and ceiling and floor. Each photograph was connected to another, lines of yarn or thin nylon cord connecting the pair. Rosario could read photographs as texts, she recognized that each snapshot is half an equation, that the moment and place captured has another match somewhere in an attic or basement. And within these matched photographs were unexplainably connected lives.
which seemed metaphorical for the process of writing hypertext, and
But during shape recognition tests, Cain could not identify dogs as dogs or oranges as circles. His brain interpreted all two dimensional shapes not as objects but as strings of text. However the text did not correspond with the shapes directly, they were poetic translations of objects and geometry.
Cain continued to live and work with the nuns until he died at the age of forty-seven. He is said to have father two children by two different nuns. And published fourteen books of poetry under fourteen different names.
because it is so much like my Phylum project.
_____
Cahterine,
I like the work. Did you get longer files from the
speech to text programs and then load the brief text
into the creation? Or have are you connected the
sounds to text in terms of thematics?
Also..I noticed the work takes a while to load. You
most likely already know this...but you might check
out some of these preloaders ...some can load external
swfs which keep you from hassling with changing the
keyframes...
http://actionscript.org/movies-results.php?searchTerm=preloader&Submit=Search%21
http://www.flashkit.com/movies/Scripting/Preloaders/index.php
laslty...if you dont mind...can I use you as a test
person...I jsut finished this project last night and
cant see it straight...
http://www.heliozoa.com/gene/bomargene.htm
cheers, Jason
--- Catherine Daly wrote:
> Jason --
>
> Puppyflowers took mine for their sound issue.
>
> http://www.puppyflowers.com/4/musiccc.swf
>
_______
> I've recently finished the first few poems of a
> speech
> to text poetics project. And the resulting poems are
> PRINT based so I'm attempting a return to paper. The
> website below has a description of the project and
> links to the first few poems. But briefly: I take
> media files (speeches, movies, ambient noise, radio
> etc.) and record the audio through a speech to text
> program. The program, being trained to my voice,
> reads
> everything wrong and spits out these long single
> spaced files of awkward text and phrases loosely
> based
> on the original. Think of it as a translation of a
> translation of a translation of a translation..the
> code beneath our language. I then reform those text
> files into poetry (using certain rules).
>
> The URL:
> http://www.heliozoa.com/speech/speechtext.html
>
> I would love to hear thoughts on if these poems (and
> the project is compelling), do they work for you,
> just
> nay thoughts really. I know this isn't really a new
> idea, but my first few attempts turned out such
> interesting poems, I wanted to take it further. And
> lastly, anyone know of any print mags or journals
> that
> might be interested in something like this?
>
> The URL again (just to be cheeky):
> http://www.heliozoa.com/speech/speechtext.html
>
> Cheers and thanks... Jason Nelson
The poems are actually one of two sets written with MS Reader avatars back... when they came out, in a longer project that is or was based on 1960s children's pre-reading perception exercises (I have some great old school mimeo workbooks, etc.).
So those at Puppyflowers were written, then translated using babelfish. I recorded them being read by the MS Reader avatars in their language, and let the voice recognition software homolinguistically translate. Then I made them into dialogs with the male and female avatars in each language (which translate slightly differently). The longer part has the stages in it.
The others were me reading tongue twisters into the voice recognition software, then reading the results into the voice recognition software, then reading the results.... Our pet parrot was going wild so that got in there too.
The avatars section really reminded me of post ww2 Europe, because the avatars were French and German.
The thing about most of the results of the perception exercises (which are based on nouns) is the way they conjure commercialism. This was the surprising thing about the untrained voice recognition as well -- it kept throwing in brand names, aka "Eddie Bauer" etc.
Thus, in your voice recognition poems from tv, movies, etc., I can't help but notice that self-reflexive quality that these machine written texts seem to have -- mechanical themselves, they bring out the technical aspects of the material -- editing, scene setting / voice over in film; types of rhetorical gestures in speechmaking, etc.
The other project is of course vast. Of course, I gravitated to
Inside Rosario’s two bedroom home (bought with her husband, who after two years of marriage disappeared while swimming in the ocean) are hundreds of thousands of photographs pasted, glued, tacked to the walls and ceiling and floor. Each photograph was connected to another, lines of yarn or thin nylon cord connecting the pair. Rosario could read photographs as texts, she recognized that each snapshot is half an equation, that the moment and place captured has another match somewhere in an attic or basement. And within these matched photographs were unexplainably connected lives.
which seemed metaphorical for the process of writing hypertext, and
But during shape recognition tests, Cain could not identify dogs as dogs or oranges as circles. His brain interpreted all two dimensional shapes not as objects but as strings of text. However the text did not correspond with the shapes directly, they were poetic translations of objects and geometry.
Cain continued to live and work with the nuns until he died at the age of forty-seven. He is said to have father two children by two different nuns. And published fourteen books of poetry under fourteen different names.
because it is so much like my Phylum project.
_____
Cahterine,
I like the work. Did you get longer files from the
speech to text programs and then load the brief text
into the creation? Or have are you connected the
sounds to text in terms of thematics?
Also..I noticed the work takes a while to load. You
most likely already know this...but you might check
out some of these preloaders ...some can load external
swfs which keep you from hassling with changing the
keyframes...
http://actionscript.org/movies-results.php?searchTerm=preloader&Submit=Search%21
http://www.flashkit.com/movies/Scripting/Preloaders/index.php
laslty...if you dont mind...can I use you as a test
person...I jsut finished this project last night and
cant see it straight...
http://www.heliozoa.com/gene/bomargene.htm
cheers, Jason
--- Catherine Daly wrote:
> Jason --
>
> Puppyflowers took mine for their sound issue.
>
> http://www.puppyflowers.com/4/musiccc.swf
>
_______
> I've recently finished the first few poems of a
> speech
> to text poetics project. And the resulting poems are
> PRINT based so I'm attempting a return to paper. The
> website below has a description of the project and
> links to the first few poems. But briefly: I take
> media files (speeches, movies, ambient noise, radio
> etc.) and record the audio through a speech to text
> program. The program, being trained to my voice,
> reads
> everything wrong and spits out these long single
> spaced files of awkward text and phrases loosely
> based
> on the original. Think of it as a translation of a
> translation of a translation of a translation..the
> code beneath our language. I then reform those text
> files into poetry (using certain rules).
>
> The URL:
> http://www.heliozoa.com/speech/speechtext.html
>
> I would love to hear thoughts on if these poems (and
> the project is compelling), do they work for you,
> just
> nay thoughts really. I know this isn't really a new
> idea, but my first few attempts turned out such
> interesting poems, I wanted to take it further. And
> lastly, anyone know of any print mags or journals
> that
> might be interested in something like this?
>
> The URL again (just to be cheeky):
> http://www.heliozoa.com/speech/speechtext.html
>
> Cheers and thanks... Jason Nelson
Comments