[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Digi Beta versus D1



>The answer was no. His argument was that all steps had to be D1 from
>beginning to end because Digi beta meant compression therefore loss of
>quality.


The easiest way to spot generational errors is to dif (subtract one picture
from another) the two images. If they are identical the result is zero.
Since digibeta introduces block errors you will see the resulting
differences as groups or block errors. The analysis for the recieving point
is more difficult as they have no "first" generation version to compare the
images. On some content the effects of compression wil be invisible on
other content the differences will stick out. Watch out for truncation
errors when you move from 10 bit DBC land to 8 bit D-1 land.

I am being told about possible picture content induced artifacts but without
>anything being said about which picture would create what artefacts.

The biggest single artifact is noise which has a characteristic different
from FM tape noise or film grain. One of the factors that masks this error
is that the noise is worse in noisy parts of the picture. Look in the
bushes in daylight exteriors. The other type of error is quanization errors
which will occur in smooth shaded areas. The effect of freqency quanization
also produces "ringing" or mosquito noise around high freqency edges (type
or heavily backlit scenes with blinds).

James Fancher           Truth unfolds in time through a communal process
Chief Science Officer
POP
730 Arizona Ave.
Santa Monica, CA 90401

310 458-3300 voice
310 394-6852 fax
310 319-1750 direct line




---
Thanks to Tom McMahon and Microsoft for support in 1998.
No product marketing allowed on the main TIG.  Contact rob at alegria.com
1014 subscribers in 39 countries on Thu Sep 17 14:36:33 CDT 1998 
subscribe/unsubscribe with that Subject: to telecine-request at alegria.com
complete information on the TIG website http://www.alegria.com/tig3/