PITCH
SCALER USER’S MANUAL
About
the author
Introduction
Parameters
Explained
Attributes
Explained
Tweaking
Tips : WARNING READ CAREFULLY
Known
Bugs and further development
Technology
& Licences
Acknowledgements
Contact
I’m a European (French) student recently graduated
in computer science applied to business management. I’m member of the Moov’Art’ association which activity is centered
around audiovisual projects. As my passion is around
sound, music and multimedia, I also decided to contribute in the development
of Buzz’s plugins. And that’s it, after months ,
one of my dreams is coming true : my first DSP effect, the Pitch Wizard !
So what can do this machine for you ?
Here’s a short list of what features you can manage
with it :
“ Many old generators couldn’t manage the new
pitch wheel MIDI messages. Plug in your MIDI keyboard to your computer,
connect the Pitch Wizard after a generator, set up the same midi channel for
both generator and Pitch Wizard, set the PW amplitude, play and enjoy !
a Tired of sung or speech samples
that you can’t transpose without getting a Mickey mouse’s or Dark Vador’s voice ? Select the Formant
process in the Pitch Wizard tracks and transpose them as you wish (don’t expect
more than 1 octave up and down : some things still remain utopia)
b Do you feel lazy and have a great CPU : make
your chords dynamically by adding tracks to the Pitch Wizard
and setting the relative transpositions on each track. One key pressed, a
chord out of your speakers !
c Do you want to apply different effects
to each Pitch Wizard track (delay on the higher chord partial, vibrato on
the lower, etc etc): select an AuxBus
channel for each track and connect the desired effects after the AuxReturn machines !
d Tired of recurrent pad presets
: connect Pitch Wizard after a generator, set up the length of scaling slides
and tweak it, you’ll getting more evolutive
pads with moving partials !
Back
to top
Parameters
Explained
u Size of the FFT transform used to process analysis
and synthesis (in some cases this value has to be tweaked to avoid strange
sounds in Formant mode, sorry. I’m still working
on it)
Adjust
this value according to the sample rate of your audio driver and to the “voice”
to process (Higher fundamentals require less FFT size, Lower fundamentals
require more)
Here’s
a informational note for sample rates (for “good” results):
Harmonic
mode Formant mode
<= 44 kHz: 256/512 128/256
48 kHz: 512/1024
256/512
96 kHz : 1024/2048 512/1024
>96kHz interpolate
and try interpolate and try
This
parameter will become an attribute when I fix the problems with Formant recognition.
v Pitch wheel range in semitones
w Position of the pitch wheel
x Initial value of transposition when setting a
scaling slide length, instantaneous transposition level else (doesn’t modify
the sound while a slide is in progress, just sets the next value to use).
y Final value of transposition when setting a scaling
slide length, instantaneous transposition level else (modifies the sound while
a slide is in progress by interpolating values).
z Length of the scaling slide
{ Processing type : Harmonics is good only for instruments,
Formants can be selected for both human voice and
instruments (listen and make your choice)
| Redirection to one or two internal auxiliary bus
channels of Buzz (Mono or stereo mode)
} Volume of this track
Back
to top
Back
to top
Tweaking
Tips : WARNING !!! READ CAREFULLY !!!
-
This version seems to be quite stable, but check your CPU performances
if you decide to increase the overlap rate or the FFT Size: this could lead
to a complete freeze of your system because of the lack of resources.
-
Adapt the size of the FFT according to the
average fundamental frequency of the input signal in Formant mode. See next
section for explanations ...
Known
Bugs and further development.
- The Formant processing
mode has to be considered as an early Alpha version feature : it’s based on
a very simplified algorithm that will be much more enhanced in the future.
- The code is not
optimised yet, I will enhance this among each new release
Back
to top
Pitch Wizard License
Pitch Wizard (c)
Version 0.1 beta (july 2003)
Buzz Effect : Pitch Scaling routines
COPYRIGHT
2003 Marc Alvarez (mad-knight@wanadoo.fr)
Permission to use, copy and distribute this software and its
documentation
for any purpose and FOR FREE is hereby granted without fee, provided that
the above copyright
notice and this license appear in all source copies.
THIS SOFTWARE
IS PROVIDED "AS IS" WITHOUT EXPRESS OR IMPLIED WARRANTY OF
ANY KIND. The code is not distributed.
It remains property of the author and should not be recovered from reverse
engineering neither sold or distributed in any way.
This software is based on freeware programs :
The
FFT engine is a freeware code :
FFTReal
Version
1.03, 2001/06/15
Class of Fourier transformation of real data (FFT and IFFT)
Portable ISO C++
(c) Laurent de Soras ldesoras@club-internet.fr
The Pitch Scaling
algorithm is DERIVED from a freeware code :
NAME: smsPitchScale.cp
VERSION: 1.01
HOME URL: http://www.dspdimension.com
KNOWN BUGS: none
SYNOPSIS: Routine for doing pitch scaling while maintaining
duration using the Short Time Fourier Transform.
DESCRIPTION: The routine takes a pitchScale
factor value which is between 0.5
(one octave down) and 2. (one octave up). A value of
exactly 1 does not change
the pitch. numSampsToProcess
tells the routine how many samples in indata[0...
numSampsToProcess-1] should be pitch scaled and moved
to outdata[0 ...
numSampsToProcess-1]. The two buffers can be identical
(ie. it can process the
data in-place). fftFrameSize
defines the FFT frame size used for the
processing. Typical values are 1024, 2048 and 4096. It
may be any value <=
MAX_FFT_FRAME_LENGTH but it MUST be a power of 2. osamp is the STFT
oversampling factor which also determines
the overlap between adjacent STFT
frames. It should at least be 4 for moderate scaling
ratios. A value of 32 is
recommended for best quality. sampleRate
takes the sample rate for the signal
in unit Hz, ie. 44100 for 44.1
kHz audio. The data passed to the routine in
indata[] should be in the range
[-1.0, 1.0), which is also the output range
for
the data.
COPYRIGHT 1999 Stephan M. Sprenger
<sms@dspdimension.com>
The Wide Open License (WOL)
Permission to
use, copy and distribute this software and its
documentation
for any purpose is hereby granted without fee, provided that
the above copyright
notice and this license appear in all source copies.
THIS SOFTWARE
IS PROVIDED "AS IS" WITHOUT EXPRESS OR IMPLIED WARRANTY OF
ANY KIND. See
http://www.dspguru.com/wol.htm for more information.
Back
to top
I wish to thank many people that contributed in
a way to this project :
- Oskari Tammelin for having made Buzz and the API @ www.jeskola.net
- Marc Van Agteren who
keeps the treasures @ www.buzzmachines.com
- Edward Blake for his detailed
guide to build Buzz machines
- Miko Appo for his helpful machine template @ http://web.hibo.no/~mva/dev/ch_template106.zip
- Stephen Sprenger
for his useful article about pitch scaling and a piece of his code @ www.dspdimension.com
- Steven W. Smith for his wonderful DSP book @
www.dspguide.com
-
Laurent de Soras which FFTReal
routines are the core of this plugin (I don’t remember
where I got his package but here’s his mail : ldesoras@club-internet.fr
-
Matteo Frigo’s team who is in charge
of the FFTW project (ultra fast FFTs) @ www.fftw.org (I wish I could use FFTW
routines but I would have to distribute Buzz sources to comply with the GNU GPL, and as Buzz
is not open source I used FFTReal)
- My girlfriend who supported me along this long
and hard trip (no mail neither picture, sorry ...)
- all Buzz community members that help keeping
this virtual studio the best as it can be
You can send a mail
to mad-knight@wanadoo.fr for comments,
suggestions, bug reports, songs made with this plugin, or even commercial
adaptation requests and job proposition ...
Back
to top
