an automatic cut-up experiment.
using speech recognition i analyzed a video of a speech by barack obama. i then sorted the snippets containing a word each alphabetically.
this video is only an excerpt of the full a-z sequence.
link
 Vienna, Austria
About me
jörg piringer
born in 1974. currently living in vienna/austria. member of the institute for transacoustic research. member of the vegetable orchestra (das erste wiener gemüseorchester). master degree in computer science. sound poet. videos. tinkering. net art stuff etc. my home link the institute for transacoustic research link vegetable orchestra link My Gear
computer, vegetables & other stuff
+Most Recent Blogs
|
Electronic Music promo: sorted speech 1
Store
Written December 29 2009 , Tags: cut-up, modification, obama, speech
Comments
ahh.. .seems to have chopped half of each word off..
if it could be fine tuned it would be really useful for finding samples - being able to scan through it alphabetically or have them perhaps chopped out and the files named as the word. yeah it's still quite rough. but word/sound segmentation is not easy. my script can also write files for each word...
keep working.. seems you've got the bullk of it there and working!
id be curious how it sounds if you took say extra instead of less... just simple like whatever the current finder finds, + x milliseconds on start, + y milliseconds on end? maybe the amount chopped off the start/end is similar and constant for alot of words/word types/starting letter/ending letter/something? another thought is maybe like looping a small bit of the start letter before the word starts, and a small bit of the end letter.. or reversing+envelope like recycle can do to a drum hit to give it some tail Signup to comment
|