h a l f b a k e r yQuis custodiet the custard?
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
Auto-tune is pretty cool. You give it a list of acceptable pitches
(ie:
the key of the song). It then does a pitch analysis of your (usually
vocal) track and pitch shifts every note to the nearest acceptable
pitch.
The pitch shifting is quite subtle, but you can still hear it (it
sounds
sort of phaser-y). Moreover, the pitch shift removes character from
the original track.
Better would be if you could record several takes of the track (as
is
usually done anyway). Auto-edit would then dynamically cut
between
takes, always choosing the one that is most in tune. The result is
an
in-tune track without any pitch shifting distortion.
While your at it, the algorithm could be given other criteria for
auto-
editing, for example giving negative weight to pops (sometimes
happens when you say b's or p's). It could also take parameters for
when it's allowed to cut (moments of silence, for example).
You could essentially use the algorithm that scores your vocals in
Rock Band (video game).
Pitch Correction wikipedia
http://en.wikipedia...ki/Pitch_correction [calculust, May 14 2009]
Rock Band wiki
http://en.wikipedia...trument_peripherals "For the most part, singers are judged on how closely they match the pitch of the song's vocalist. During "talking parts" that do not judge pitch, a phoneme detector will pick up individual vowels and consonants of the spoken lyrics." [calculust, May 14 2009]
[link]
|
|
Thinking about the fine details of how to get individually dubbed words to sound good when strung together, I could imagine a lot more work for the editor. I could also imagine excellent results at the end of the process. [+]
I liked the additional idea of suppressing pops that result from performers being close to the microphone when pronouncing explosives. Existing audio cleaning algorithms could deal with these. I wonder if we could get an algorithm that fills in a few of Lily Allen's glottal stops with the letter 't'. |
|
| |