Not all video has 8 bits per pixel. if I use HD-SDI input from a an
AJA Kona card, or have a DPX QT wrapper or TIFF image sequence wrapped
in a reference movie it should be seen as 32 bit 'HDR' or even FCP has
10bit uncompressed 4:2:2 components included for QT, id love that
extra 2 bits to be seen, especially if outputting on a nice display
device.
This would be similar to working at 32bit in After Effects, better
color accuracy, for blurs and highlight, level control, etc.
Does the video input/movie player/image input source patches determine
the bit depth on a per codec basis via QT, or are all images reduced
to 8 bit within a normal QC context?
Interesting about Render in Image, I do not see a GUI method of
setting 32bit. How is this accomplished?
At this point this is more of an academic curiosity. For reference,
the GLSL shader that is having issues with video versus image is the
one I posted in reply to 'Vertex Displacement Mapping'. I will admit I
am a bit out of my expertise, so I assume I have an issue on my end.
Ive made sure to force 2D textures with the "Image Texturing
Properties" patch and resize the image to a power of two, to no avail.
Oddly any size jpeg, tiff, etc seem to work.
Thank you again Pierre and Alessandro for your help - QC 3.0 is a
really huge update and very exciting.
On Nov 5, 2007, at 1:30 PM, Pierre-Olivier Latour wrote:
I notice in QC 3.0 release notes there is support for floating
point textures. When are these used, and how does one enable
floating point output from source patches?
If you use a QCRenderer to a floating-point context, then QC should
use floating point textures internally
Otherwise, the only way to "force" a texture to be floating point is
to use a Render in Image and set its internal format to 32bits /
pixel. But this is pointless and is just gonna use a lot more VRAM
unless you know exactly what you are doing.
Are images vs video treated differently?
No, they're all "images"
I have a glsl shader that seems to work with a image input patch,
but not a movie player / video input patch? Ive tried some
workarounds like using render to image in hopes it changes the
destination texture type, and putting in a cikernel, with no luck.
This sounds like a bug: please report to bugreport.apple.com
So in short, is there a way to force float textures for video ?
Why would you wanna do that? Video input has 8 bits resolution
anyway, so you wouldn't gain anything.
________________________________
Pierre-Olivier Latour - [EMAIL PROTECTED]
Technologist - Apple, Inc.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Quartzcomposer-dev mailing list ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/quartzcomposer-dev/archive%40mail-archive.com
This email sent to [EMAIL PROTECTED]
