My idea was to bleed only light that exceeded the RGB dynamic range;
that is, color in excess of rgb 1.  Colors within the dynamic range
would remain un-smeared.  My first thought was to see if I could adapt
Sam's luminous bloom rig for this purpose, but I haven't been able to
figure out how he did it.

I settled on mathematical manipulation of a pigment function.  The
source image was rendered in the EXR format.  It was slow, but it got
the job done.  This POC tries to get the look of iris diffraction.

bleed_test_image2.jpg is a JPEG render of the source image.  As noted
above, the actual source image was an EXR.

bleed_poc1-excess2.jpg shows the colors in excess of rgb 1.

bleed_poc3-power4.jpg shows the excess color smeared out into 4
diffraction spikes.

bleed_poc4-final4.jpg is the final combined image using 10 diffraction
spikes.

Post a reply to this message

Attachments:
Download 'bleed_test_image2.jpg' (51 KB) Download 'bleed_poc1-excess2.jpg' (19 KB) Download 'bleed_poc3-power4.jpg' (28 KB) Download 'bleed_poc4-final4.jpg' (51 KB)

Preview of image 'bleed_test_image2.jpg'

Preview of image 'bleed_poc1-excess2.jpg'

Preview of image 'bleed_poc3-power4.jpg'

Preview of image 'bleed_poc4-final4.jpg'

That is pretty smart! Very well done indeed.

-- 
Thomas

Post a reply to this message

Cousin Ricky <ric### [at] yahoocom> wrote:
> My idea was to bleed only light that exceeded the RGB dynamic range;
> that is, color in excess of rgb 1.  Colors within the dynamic range
> would remain un-smeared.  My first thought was to see if I could adapt
> Sam's luminous bloom rig for this purpose, but I haven't been able to
> figure out how he did it.
>
> I settled on mathematical manipulation of a pigment function.  The
> source image was rendered in the EXR format.  It was slow, but it got
> the job done.  This POC tries to get the look of iris diffraction.
>
> bleed_test_image2.jpg is a JPEG render of the source image.  As noted
> above, the actual source image was an EXR.
>
> bleed_poc1-excess2.jpg shows the colors in excess of rgb 1.
>
> bleed_poc3-power4.jpg shows the excess color smeared out into 4
> diffraction spikes.
>
> bleed_poc4-final4.jpg is the final combined image using 10 diffraction
> spikes.

It looks good, Ricky!

My original method used either averaged pigments or averaged function
pigments... I can't remember which.

My latest released version (lb7b, I think) used averaged reflective textures
with planar normals. Each texture entry used the input image to influence the
metallic reflection of the finish. It also allowed the reflection_exponent to be
specified, which in turn allowed you to control how the glare came from bright
vs. dark areas.

If I make another version I will probably use averaged pigment functions again,
since there is less computation involved.

Sam

Post a reply to this message

Cousin Ricky <ric### [at] yahoocom> wrote:
> My idea was to bleed only light that exceeded the RGB dynamic range;
> that is, color in excess of rgb 1.  Colors within the dynamic range
> would remain un-smeared.  My first thought was to see if I could adapt
> Sam's luminous bloom rig for this purpose, but I haven't been able to
> figure out how he did it.
>
> I settled on mathematical manipulation of a pigment function.  The
> source image was rendered in the EXR format.  It was slow, but it got
> the job done.  This POC tries to get the look of iris diffraction.
>
> bleed_test_image2.jpg is a JPEG render of the source image.  As noted
> above, the actual source image was an EXR.
>
> bleed_poc1-excess2.jpg shows the colors in excess of rgb 1.
>
> bleed_poc3-power4.jpg shows the excess color smeared out into 4
> diffraction spikes.
>
> bleed_poc4-final4.jpg is the final combined image using 10 diffraction
> spikes.

I love it!

Post a reply to this message