Antares auto-tune-for tdm and vst Manual Do Utilizador

Página de 57
 
Auto-Tune UserÕs Manual
 
15
 
How Auto-Tune Corrects Pitch
 
and Auto-Tune will have difÞculty determining the pitch of this sound.
There is a slider (the Tracking slider, discussed in Chapter 4) that will allow
Auto-Tune to be more casual about what it calls ÒperiodicÓ, which will help
in these cases.
If you have problems processing a particular sound, zoom it up and look at
it. If it is non-periodic, Auto-Tune will not work. If you can correct that
problem, by using EQ or re-recording, then those are options available to
you.
The precision by which Auto-Tune detects pitch is extraordinary. At a fre-
quency of 400 Hz and a sample rate of 44100, the Auto-Tune DSP algorithm
computes the pitch to an accuracy of .0001 samples per cycle, or .0004 Hz.
At this resolution, the very question ÒWhat is pitch?Ó becomes relevant.
That is, as the pitch of typical performances continuously change, the
amount of variation in pitch, even over the time of a few cycles, changes
greatly in comparison to the accuracy by which Auto-Tune computes pitch.
The pitch computed by Auto-Tune is a mathematical estimate of the cycle
period repetition rate over the last two (or sometimes the last four) cycles.
Auto-Tune was designed to detect and correct pitches up to C6. In reality, if
a pitch goes higher than C6, Auto-Tune will often interpret the pitch an
octave lower. This is because it interprets a two cycle repetition as a one
cycle repetition. These pitches can be corrected by the usual means. On the
low end, Auto-Tune will detect pitches as low as A0 (55Hz). This range of
pitches allows intonation correction to be performed on all vocals and
almost all instruments.
 
How Auto-Tune Corrects Pitch
 
Auto-Tune changes the instantaneous pitch and introduces no distortion in
the output. It does this using the same basic computations used in samplers
to re-tuned samples.
This is different from other pitch-shifting algorithms which put noise in
sounds as a result of FFT (Fast Fourier Transform) overlap-and-save algo-
rithm phase errors. The Auto-Tune algorithm is very clean and transparent.
It isnÕt possible to tell that a sound has been processed by Auto-Tune -
except that it is in-tune.