🚀 go-pugleaf

In article <14358@borg.cs.unc.edu>, jeffay@jeffay.cs.unc.edu (Kevin Jeffay) writes:
>
> Greetings -- I am trying to determine the requirements on lip synchronization
> for applications such as video conferencing.  Specifically, I am looking for
> some references on human perception of the synchronization of audio and video.
> Ideally I'm looking for published studies in either the human perception or
> perception and psychophysics literature that deal with this topic?  For
> example, can people tell if audio is say 50ms. ahead or behind the video (and
> does it bother them)?

Speaking entirely from personal experience, I'd say people would be pretty
tolerant of discrepancies in synchronization between audio and video. I think
you could have a shift of maybe a tenth of a second without it becoming really
noticeable. Of course, my experience has only been with *constant* shifts
throughout a movie segment; I don't know what the effect would be of, say, a
random jitter of up to that amount.

Accuracy of audio timing is a different matter altogether. I'd say a discrepancy
in timing of music notes on the order of 10-20ms is just on the threshold that
I can detect.

Lawrence D'Oliveiro                       fone: +64-7-856-2889
Computer Services Dept                     fax: +64-7-838-4066
University of Waikato            electric mail: ldo@waikato.ac.nz
Hamilton, New Zealand    37^ 47' 26" S, 175^ 19' 7" E, GMT+12:00

In article <1992Aug7.171104.9943@waikato.ac.nz> ldo@waikato.ac.nz (Lawrence D'Oliveiro, Waikato University) writes:
>In article <14358@borg.cs.unc.edu>, jeffay@jeffay.cs.unc.edu (Kevin Jeffay) writes:
>>
>> Greetings -- I am trying to determine the requirements on lip synchronization
>> for applications such as video conferencing. <deleted stuff>  ... For
>> example, can people tell if audio is say 50ms. ahead or behind the video
>> (and does it bother them)?
>
>Speaking entirely from personal experience, I'd say people would be pretty
>tolerant of discrepancies in synchronization between audio and video. I think
>you could have a shift of maybe a tenth of a second without it becoming really
>noticeable.

Unfortunately, I don't have a documented reference for these figures, but I
have often heard the numbers 30-40 ms used as a worst case for lip sync.

Eric Miller

***************************************************************************
*  The opinions expressed here are mine, and may not reflect those of     *
*    Microware Systems or its other employees.                            *
***************************************************************************