Infinite clipping

Discussion in 'Analog & Mixed-Signal Design' started by KL7AJ, Jul 15, 2016.

  1. KL7AJ

    Thread Starter AAC Fanatic!

    Nov 4, 2008

    Back in the dark ages, ATT (Bell labs) and others did some studies on "infinite clipping", that is, differentiating only the zero crossings of an audio waveform. The results were amazing...not just from an acoustic standpoint...but by how well the human brain can replace all the missing parts.....something still poorly understood.

    Cool stuff. They're sort of re-opening this investigation.
  2. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    This technique of extreme clipping was used in many early secure voice systems. I remember using one old system in tech school. To say the voice quality was bad is an understatement.

    In the early system the human voice was separated into several frequency bands and the amplitude levels of each were compressed to a very narrow range to encrypt them before transmission. This was reconstructed with another vocoder on the receive end.
    Last edited: Jul 15, 2016
  3. wayneh


    Sep 9, 2010
    I may be missing the point. I can hear the difference between a 440Hz sine wave and a 440Hz square wave, but I can use either to tune my guitar and I suppose it would not be hard to understand a voice composed of square waves. So what?
  4. nsaspook

    AAC Fanatic!

    Aug 27, 2009
    It was an early form of data compression before complex DSP electronics. Separating the speech into bands and then hard limiting the amplitude to zero crossing (samples of the signal) was a rough approximation of an FFT transform with simple filters and circuits. Outmoded with modern electronics that can easily have better performance and intelligibility using the same transmission bandwidth.

    atferrari likes this.
  5. GopherT

    AAC Fanatic!

    Nov 23, 2012
    @nsaspook, This image is great, thanks for posting!