How to -really- normalize ?


    Apr 07 2007 | 6:02 pm
    An oscillator. 4 waveforms. Measured with peakamp~, sine = 1., tri =
    0.49, pulse = 0.49, saw = 0.69... How to boost the signal always to 1. ?
    Normalize~ doesn't work because of the 'reset' message you have to send
    (sometimes it even doesn't change anything), tap.normalize~ is a wahwah
    / lfo style effect, nothing helpfull in the bennies, nor the jimmies,
    nor in any-ies. I made my own dirty kitchen patch to get some results
    but i really wonder if there is a clean way (abstraction, external...)
    to do this job, RMS and / or peak...
    thanks in advance
    best wishes
    f.e
    --
    f.e chanfrault | aka | personal computer music
    > >>>>>> http://www.personal-computer-music.com
    > >>>>>> |sublime music for a desperate people|

    • Apr 07 2007 | 7:04 pm
      Try tl.balance~.
      Best,
      Trond
      f.e wrote:
      > An oscillator. 4 waveforms. Measured with peakamp~, sine = 1., tri =
      > 0.49, pulse = 0.49, saw = 0.69... How to boost the signal always to 1. ?
      > Normalize~ doesn't work because of the 'reset' message you have to send
      > (sometimes it even doesn't change anything), tap.normalize~ is a wahwah
      > / lfo style effect, nothing helpfull in the bennies, nor the jimmies,
      > nor in any-ies. I made my own dirty kitchen patch to get some results
      > but i really wonder if there is a clean way (abstraction, external...)
      > to do this job, RMS and / or peak...
      >
      > thanks in advance
      >
      > best wishes
      >
      > f.e
    • Apr 07 2007 | 8:10 pm
      Isn't this what we call a compressor/expander? With an inifinite ratio in your case..
      Mattijs
      Quote: f.e wrote on Sat, 07 April 2007 20:02
      ----------------------------------------------------
      > An oscillator. 4 waveforms. Measured with peakamp~, sine = 1., tri =
      > 0.49, pulse = 0.49, saw = 0.69... How to boost the signal always to 1. ?
      > Normalize~ doesn't work because of the 'reset' message you have to send
      > (sometimes it even doesn't change anything), tap.normalize~ is a wahwah
      > / lfo style effect, nothing helpfull in the bennies, nor the jimmies,
      > nor in any-ies. I made my own dirty kitchen patch to get some results
      > but i really wonder if there is a clean way (abstraction, external...)
      > to do this job, RMS and / or peak...
      >
      > thanks in advance
      >
      > best wishes
      >
      > f.e
      > --
      > f.e chanfrault | aka | personal computer music
      > > >>>>>> http://www.personal-computer-music.com
      > > >>>>>> |sublime music for a desperate people|
      >
      ----------------------------------------------------
    • Apr 07 2007 | 10:25 pm
      > > Normalize~ doesn't work because of the 'reset' message you have to send
      > >
      well "normalizing" is nonrealtime by definition, if you want
      something similar in realtime you will always need some kind
      of "reset" - or a permanent process such as in a compressor.
    • Apr 08 2007 | 12:01 am
      I think you should just use a hard limiter or compressor and then maybe turn it up.... If you have the bennies I think limit3~ is pretty good. Otherwise, as said before, just record and then do non realtime normalize after.
    • Apr 08 2007 | 12:37 am
      wouldn't it work if you just keep an account of the peak amplitudes of
      your component wave forms and then multiply by a scaler in your way
      out??
      that is if you have three oscilators with peak amplitudes 0.4, 0.5 and
      0.6 (adjusted with a *~ each), total peak amp 1.5, multiply(*~) by
      0.66 and you get 1?? of course phase and frequency differences could
      give you a total peak amp of less than 1.5, but it should work i
      think...
      J
      On 4/7/07, Nicholas C. Raftis III wrote:
      >
      > I think you should just use a hard limiter or compressor and then maybe turn it up.... If you have the bennies I think limit3~ is pretty good. Otherwise, as said before, just record and then do non realtime normalize after.
      >
      >
      > --
      > -=ili!ili=- www.Axiom-Crux.net -=ili!ili=-
      >
      --
      Jaime E Oliver LR
      joliver@ucsd.edu
      www.realidadvisual.org/jaimeoliver
      www-crca.ucsd.edu/
      www.realidadvisual.org
      8693 Via Mallorca No. 19
      La Jolla, CA 92037
      USA
    • Apr 08 2007 | 7:19 am
    • Apr 08 2007 | 7:35 am
    • Apr 08 2007 | 12:51 pm
      This looks like a lovely set of objects but not yet UB?
      On 4/8/07 3:35 AM, "Trond Lossius" wrote:
      > Using the level of one signal to balance another is discussed in
      > "Computer Music" by Jerse and Dodge. My object tl.balance~ is based on
      > the algorithm provided there.
      >
      > Best,
      > Trond
      Cheers
      Gary Lee Nelson
      Oberlin College
      www.timara.oberlin.edu/GaryLeeNelson
    • Apr 08 2007 | 5:59 pm
      No, unfortunately. I get a intel-based computer in a week or so, so the
      motivation for porting to UB is increasing by the day... ;-)
      I hope to have it done in the next two months. I have only a few
      externals ported so far.
      Best,
      Trond
      Gary Lee Nelson wrote:
      > This looks like a lovely set of objects but not yet UB?
      >
      >> Using the level of one signal to balance another is discussed in
      >> "Computer Music" by Jerse and Dodge. My object tl.balance~ is based on
      >> the algorithm provided there.
      >>
    • Apr 08 2007 | 10:51 pm
      even if not exactly what you're looking for, this might be useful as
      it will normalize your signal if it is greater than 1.
      _g
      On 7 avr. 07, at 20:02, f.e wrote:
      > An oscillator. 4 waveforms. Measured with peakamp~, sine = 1., tri
      > = 0.49, pulse = 0.49, saw = 0.69... How to boost the signal always
      > to 1. ? Normalize~ doesn't work because of the 'reset' message you
      > have to send (sometimes it even doesn't change anything),
      > tap.normalize~ is a wahwah / lfo style effect, nothing helpfull in
      > the bennies, nor the jimmies, nor in any-ies. I made my own dirty
      > kitchen patch to get some results but i really wonder if there is a
      > clean way (abstraction, external...) to do this job, RMS and / or
      > peak...
      >
      > thanks in advance
      >
      > best wishes
      >
      > f.e
      > --
      > f.e chanfrault | aka | personal computer music
      >> >>>>>> http://www.personal-computer-music.com
      >> >>>>>> |sublime music for a desperate people|
      >
    • Apr 08 2007 | 11:22 pm
      On Apr 8, 2007, at 1:59 PM, Trond Lossius wrote:
      > No, unfortunately. I get a intel-based computer in a week or so, so
      > the motivation for porting to UB is increasing by the day... ;-)
      Glad to hear it. FWIW, I recompiled a few of your butterworth filter
      objects from the included source. I needed them for an older piece I
      was performing. Worked without a hitch.
      Here's wishing you a speedy re-compile.
      -------------------
      Nathan Wolek, PhD --- nwolek@stetson.edu
      Assistant Professor of Music Technology
      Stetson University - DeLand, FL