Claudio,
I'm not sure I fully agree with your analysis (see details below).
--- In [email protected], "Re Claudio" <[EMAIL PROTECTED]> wrote:
[snip]
> 5) If we put a multiple monochromatic signal in a system with flat
response
> , linear amplitude response and NOT flat group delay , will the
signal be
> distorted ?
> YES .NEW SIGNALS (DISTORTION )WILL BE CREATED FROM THE GROUP DELAY
BEING NOT
> FLAT ,LIKE WHEN THERE IS ONLY NON LINEAR AMPLITUDE RESPONSE .
Actually I don't think there will be new components generated from the
group delay not being flat. For sure the relative phase of all the
signals will be changed, but this will have an effect only on the
(time domain) shape of the overall signal, this is why this is
important for digital signals. If you look at the signal (magnitude)
spectrum, I don't you will see any differences due to the non-flat
group delay.
Moreover, as others have said, it has to be understood how much the
human ear is sensitive to the group delay distortion; you can find an
interesting experiment at
http://www.jhu.edu/~signals/listen-new/listen-newindex.htm where you
can change the phase of all the harmonics of a square wave and listen
the resulting signal; while the time-domain shape changes a lot the
perceived sound does not change (at least for my ears).
> 6) Does the human being perceive audio distortion ?
> YES
I think we should not mix what is usually called "linear distortion",
that is characteristic of a system where the superposition principle
applies, with the "non-linear distortion", where the superposition
principle does not apply.
Common experience is that, of course, we can hear the non-linear
distortion, as new components are generated. Also amplitude linear
distortion can be heard, as some signals components will have a
different amplitude compared to the others.
But regarding the phase distortion I'm a bit skeptical it can be
readily heard: for example, it is well known that if the local
oscillator used for demodulating a SSB signals has not only the exact
frequency but also the same phase as the original carrier oscillator
used for generating it, the demodulated signal will be a mix of the
original signal and its hilbert transform (the signal shifted 90
degrees for all frequencies), yet we can normally listen and
understand SSB signals very well...
Maybe someone on the group who is less programming impaired than me
could synthetically generate an SSB signal demodulated with different
phase shifts to see if the differences can be heard...
Best regards,
Claudio, IN3OTD
