Topic: How creepy is this? :)
http://www.youtube.com/watch?v=ZWiCUUfI7M4
Exploration of spectrum profile of human voice resynthesized in realtime by using piano keys. Marvellous.
Hard work and guts!
- EvilDragon
- Member
- Offline
- Beto-Music
- Member
- Offline
Re: How creepy is this? :)
Would be quite fun if pianoteq could get some feature like that.
Just for fun. A talking piano :-)
- EvilDragon
- Member
- Offline
Re: How creepy is this? :)
Theoretically you could get this.
You need to spectrally analyze the incoming sound, you would ust FFT for that, break it into windows, then assign amplitude of each band to MIDI velocity, and tie frequencies to MIDI Note On events. Then you'd send the stream of MIDI data to Pianoteq.
Hard work and guts!
- doobya
- Member
- Offline
Re: How creepy is this? :)
That was kind of creepy indeed... But a great tech demonstration!
Good find, ED!
- Beto-Music
- Member
- Offline
Re: How creepy is this? :)
How far will take to modeled technology be able to emulate a specific human voice?
Can you imagine emulate Elvis in songs he never recorded???
Last edited by Beto-Music (18-10-2009 06:52)
- benjamind
- Member
- Offline
Re: How creepy is this? :)
Beto-Music wrote:
How far will take to modeled technology be able to emulate a specific human voice?
Can you imagine emulate Elvis in songs he never recorded???
Couldn't understand why anyone would want to emulate another's voice.
Even so, if it can be done then it should be possible. Voices, to me, are unique and musical in their own organic and special way.
Emulating those voices isn't going to have the same impact as a good original voice.
Ben
- Beto-Music
- Member
- Offline
Re: How creepy is this? :)
Supose someone wants to do a great movie in homage to a artist from the 30's or 40's.
The sound recording back then was horrible compared to today's standarts.
For me it would be more moral to emulate the artists voice, recreating his records, using the vintage recording a base, in order to create a high quality sound, than just hire somebody to impersonate hin.
benjamind wrote:
Beto-Music wrote:How far will take to modeled technology be able to emulate a specific human voice?
Can you imagine emulate Elvis in songs he never recorded???
Couldn't understand why anyone would want to emulate another's voice.
Even so, if it can be done then it should be possible. Voices, to me, are unique and musical in their own organic and special way.
Emulating those voices isn't going to have the same impact as a good original voice.
Ben
- NeilCraig
- Member
- Offline
Re: How creepy is this? :)
It's already been done. Several years ago, some software/security company (I forget who, probably Dyncorp or one of those outfits) demonstrated speech synthesis software when Colin Powell (who was in the room) professed to be unhurt and being treated well by his captors.
If I get time I'll hunt out the original details.
Best//Neil