| Author |
Topic: Synthesized Vocals |
Alan Brookes
From:
Brummy living in Southern California
|
Posted 11 Apr 2013 3:41 pm |
|
Sampling and synthesizing of instruments has been done for many years now, but how far is technology from being able to sample and synthesize vocals ?
What you would need would be samples of a singer singing each syllable at different pitches. Some artists such as Frank Sinatra, Elvis Presley, etc., have such a lot of material available that a large proportion of syllables will have been recorded, especially given that in western music the same words crop up many times over. You would need to separate them from the background instruments which are already on the recordings, but that has been done quite a lot in the past.
I'm not looking for reasons why it would be inadvisable and undesireable to do so, only whether the technology exists yet to enable it to be done.
Obviously there are copyright issues. If someone synthesized an album "Al Jolson sings the hits of the Rolling Stones," for instance, the estate of the late Al Jolson might object to the idea.
Just think of the possibilities: "Buddy Holly sings the Hits of the 90s," for instance. |
|
|
 |
Richard Damron
From:
Gallatin, Tennessee, USA (deceased)
|
Posted 12 Apr 2013 6:35 am
|
|
To seperate the voices from the instrumental background, one would need the original true stereo recording. "Vocal Eliminators" are marginal, at best, since true stereo recordings cannot be found.
But - Sinatra singing "Rock Around The Clock"? Ella doing a "Whiter Shade of Pale"? Pavarotti's rendition of "Roxanne"? NOT in MY collection, thank you! |
|
|
|
Jim Cohen
From:
Philadelphia, PA
|
Posted 12 Apr 2013 8:24 am
|
|
I think it would be hard to achieve only by using recordings of deceased singers like Sinatra because, while the same words may crop up over and over again, they crop up in different musical contexts that affect how the vowels are shaped, held, etc, etc. Not all "love"s or "baby"s are alike.
Having said that, it could probably be done, using a very thorough, systematic approach, with a live singer coached by a linguist who knows exactly what phonemes he/she needs to sing and record to be able to cover 80% of western popular music.
The mind boggles (while the body shudders)...
_________________
www.JimCohen.com
www.RonstadtRevue.com
www.BeatsWalkin.com |
|
|
|
Clete Ritta
From:
San Antonio, Texas
|
Posted 12 Apr 2013 9:17 pm
|
|
Hmm... I dont think its such a good idea. We might end up with Stephen Hawking sings William Shatner's Greatest Hits.  |
|
|
|
Alan Brookes
From:
Brummy living in Southern California
|
Posted 13 Apr 2013 9:05 am
|
|
| Now that would be something worth listening to. Two of my lifelong heroes. I'm amazed that Stephen Hawking manages to control his computerized speech with just movement of his cheeks. Most of us would become exasperated by being confined to a wheelchair for nearly fifty years with very little movement, but, far from losing his mind, his mind seems to become clearer every year. It must take the patience of Job, though, to be his assistant day and night. |
|
|
|
David Mason
From:
Cambridge, MD, USA
|
Posted 13 Apr 2013 11:32 am
|
|
I can assure you from direct personal experience that this would be a very bad idea. While a lot of people have the experience of having songs stuck in their heads, I get the cover band versions - and it ain't pretty. If you can imagine Vincent Price singing "Every Breathe You Take", yikes. Actually it's more like a generic bad "Transylvanian" accent - maybe like Count Chocula, perhaps. I think the topper of them is Elmer Fudd singing Prince's big blowzy ballad -
Poiple wain! POI-PLE WAIN!
- You get the picture. Mr. Fudd is a frequent contributor. |
|
|
|
