More interlacing woes in FCP 6.0.2
More interlacing woes in FCP 6.0.2
- Subject: More interlacing woes in FCP 6.0.2
- From: Peter Litwinowicz <email@hidden>
- Date: Tue, 11 Dec 2007 20:04:54 -0800
- Thread-topic: More interlacing woes in FCP 6.0.2
Hey FCP guys (most notably Paul :-)),
So we're doing some more serious development and have hit another brick
wall. This bug has been there since 5.1 it seems and continues to be there
in 6.0.2
We often need to apply our plugins to sequences instead of directly to clips
For example: because FxPlugs cannot extend durations of a clip, people using
Twixtor need to pad their clip to the desired duration of the extended clip
in one sequence, and then apply Twixtor in another.
Another example: let's say we have two different plugins that both deal with
multiple images in time. In order to descrease the calculation time we
often suggest to customers to apply one plugin in one sequence, render that
sequence, then drop that prerendered sequence into another and apply the
second plugin. IN this way you don't needlessly recalculate frames (because
both plugins process multiple input images to create a final output image,
if you don't do this then many frames get recalculated as the second plugin
evaluates multiple images of the first-applied plugin).
========================================================================
Anyhow, that's just motivation for you to understand that we need the
following bug fixed:
If you apply an FxPlug to a sequence within another sequence, then the input
bitmap is always reported as progressive (FxFieldOrder==0) and is
indistinguishable from a clip that is progressive that should be processed.
This is very problematic.
To repeat: if the input is a CLIP (and not a sequence), then the input field
order is always correct (the field order on the input image given to
-renderOutput) and so is the output field order (the field order on the
outputImage given to -renderOutput. (of course getting to all the fields of
an interlaced clip within a progressive sequence is problematic, but that's
another bug report ;-))
If my FxPlug is applied to a sequence (instead of clip) within another
sequence, then the input field order is always reported as progressive,
regardless of whether the input sequence is set to have fields or not (and
the input field order should be set to whatever field setting the INPUT
sequence has as its setting).
Another problem happens with sequences dropped into clipwells. If I drop an
interlaced sequence into a clipwell on a clip within an interalced sequence,
then the field order for the input is reported as lower. For the first
field in time I get both the input and output field is set to 2. For the
second field in time I get that the input field is 2 (wrong) and the output
field is 1 (correct).
I'll file a proper bug report.
Pete
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Pro-apps-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden