Wednesday, February 11, 2009

You Tube's Little Known Speech Synthesizer

Anyone who has spent any amount of time on YouTube can attest that the comments section beneath the videos contains some of the most neanderthal rhetoric you're ever likely to find on the internet (and that's saying a lot). Perhaps because of this (I'm too lazy to look it up, get off my back! UPDATE: Reader Brandon reports that the feature was added in response to this online comic: http://xkcd.com/481/), they added an 'Audio Preview' button that will read your post aloud, allowing you to hear how it sounds when someone says it instead of just typing it. The idea being, I guess, that once you hear it read out loud you'll realize what a petty moron you sound like. But the speech synthesis is actually quite good and could certainly serve as sample fodder if you use an audio recording utility such as Audio Hijack Pro to capture its output. It is definitely grainy and muddy, but that's not necessarily a bad thing, and some clever EQing and time in an audio editing program could probably clean it up quite a bit.
Benny Benassi, eat your heart out!

7 comments:

brandon daniel said...

Believe it or not, they implemented this in response to this XKCD comic: http://xkcd.com/202/

brandon daniel said...

errr, strike that, meant this one:

http://xkcd.com/481/

Simon King said...

If you want to hear some better speech synthesis, ours is not bad:

http://www.cstr.ed.ac.uk/projects/festival/onlinedemo.html

The 'nick' and 'nina' voices using the 'HMM method' are probably the best ones.

The demo sends you back a plain wav file. It's free too.

Simon

Tom said...

Thanks for the info from both of you! I'll update the Blog accordingly!

Viagra Online said...

I agree with you that window is old enough, I think is better doing it in the way you explain us, and the bast part we have the codes to reach it.m10m

puertas metalicas said...

I consider everyone ought to browse on it.

Anna Schafer said...

It is the decision of the vast majority of Speech Easy users!speech recognition software