Of course, on that first line, I meant to write "those in imagebufalgo.h".
On Jul 13, 2012, at 11:39 PM, Larry Gritz wrote:
> The public comments (i.e., those in imagebufalgo) don't need to explain the
> algorithms, except as possibly a one-line formula involving only the
> pre-multiplied imag
The public comments (i.e., those in imagebufalgo) don't need to explain the
algorithms, except as possibly a one-line formula involving only the
pre-multiplied images. E.g., for "over", you might say
// The "over" operation is: R = A + (1-A[alpha]) * B
and leave it at that.
But the
I confirmed from other sources that it only makes
sense to use over with blend modes, and other Porter-Duff
operations should be separate from blend modes.
That makes things so much easier now.
Larry, should I write the derivations in the
comments for the functions or is it enough to write just th
Oh yah, I forgot about that.
+1 for matching nuke's screen math (which boils down to max(a,b) if
either a or b is outside of 0,1, and screen otherwise. It's less
surprising to artists than clamping.
-- Jeremy
On Fri, Jul 13, 2012 at 10:52 AM, Jonathan Egstad wrote:
> fwiw,
> Some of the issues
fwiw,
Some of the issues with merging HDR pixels have been hashed out in Nuke for a
while, so a peek at its Merge node may be helpful:
For example 'screen' in Nuke is implemented to avoid the 1-A negative problem:
if (A >0 && A<1 && B>0 && B<1) C = A+B - A*B
else if (A > B) C = A
else
Yup.
screen = A+B-AB, where AB term represents the if intersection of the
subpixel geometry.
But in the case where you have two mattes where you know the
correlation == -1, (which denotes no overlap in the subpixel
geometry), adding mattes reduces to a simple A+B. An example of no
overlap in th
Is that really necessary? The way the doc you sent describes things, the blend
modes were always used with "over". Do PhotoShop and others give independent
control of blend and composite? Does anybody really need to do "screen + in"?
I was thinking that you'd want two separate functions:
1.
On second look I think the simplification Larry suggested won't work.
There are X number of Porter-Duff operations and Y number of blend
modes. If I am doing such simplifications, I will have to do it for
each pair (Porter-Duff operation, blend mode), and that would be
X = 12, Y = 20, X*Y=240 func
