Re: FCP FxPlug (software) - rendering in 10-bit YUV?
Re: FCP FxPlug (software) - rendering in 10-bit YUV?
- Subject: Re: FCP FxPlug (software) - rendering in 10-bit YUV?
- From: Paul Schneider <email@hidden>
- Date: Tue, 24 Jul 2007 15:56:35 -0700
Also, it's been my experience that even with 10 bit images you can be
presented with 8 bit RGB. The difference seems to come into play
between
rendering (floating point) and clicking around the timeline (8 bit),
Yes, that's correct.
In the RT display path, display to the canvas and video out is always
8-bit. Rendering (Cmd-R) happens in float for high-precision sequences.
In the legacy, NRT display path, display to the canvas is always 8-
bit. Video out will be rendered in float for high-precision
sequences, if you have a 10-bit video out device. Rendering (Cmd-R)
happens in float for high-precision sequences.
This should be covered in the "FCP Rendering Paths" document that
comes with the FxPlug SDK; if the information is missing or unclear,
please file a bug and I'll fix it!
- Paul
On Jul 24, 2007, at 12:29 PM, Peter Litwinowicz wrote:
Paul Miller wrote:
What does FCP 6 do while rendering in 10-bit YUV? Is it promoting
everything to float for the FxPlug software pipeline, or
something else?
Nevermind, I found an old message from this list that mentions
that in
fact it does.
What do other people do in this case? I have some filters that
don't yet
handle full float. Is it better to not support float at all (ie.
10 bit
YUV) or clip?
That's a marketing decision, but if you are only supporting 8 bit,
then
technically you'll need to at least convert the floating point to
8 bit
and then process and convert back (which I presume you were going
to do).
I don't think returning a black frame when presented with a
floating point
(10 bit YUV) image will make your users happy. Of course, when
they get
different bit depths out than what they put in, that will make them
unhappy
too (and users DO notice, trust me... We get lots of emails about
our AE
plugins which can only support 8 bit within FCP).
Also, it's been my experience that even with 10 bit images you can be
presented with 8 bit RGB. The difference seems to come into play
between
rendering (floating point) and clicking around the timeline (8
bit), but
perhaps I am wrong on this. Paul Schneider, can you elaborate? Am
I insane
on this point?
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:
40apple.com
This email sent to email@hidden
_______________________________________________
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