MLT and aperture shape experiment

[arc en ciel]

hello,

studying MLT and aperture diffraction's settings, I noticed something intrigating.
If I just put an image in

Code: Select all

<aperture_shape>

without changing any other parameters, Indigo is running faster (samples/sec), but quality (denoise) does not follow.
- -
I join 2 images, both are 10 minutes render,
one running at 395 ksps without "aperture_shape"
second running at 763 ksps with D69 image in "aperture_shape"
- -
if someone have an explanation ?
- -
thank's and have a superday

[fused]

I think it has something to do with how Indigo samples the aperture shape image. Rejected samples are being counted, iirc.

[arc en ciel]

probably Fused, I read again the announcement's thread of aperture in Indigo (v 1.0.4)
but I don't understand what's up....
actually I find an other image making Indigo running 10x faster, but depending maxchange value, large mutation probe, noise could be really strong in highlights areas
anyway, I exported each light'layers individually, traited differently with photoshop and voilà (ce que ça donne)

[arc en ciel]

an other try using aperture painting concept
Clouds by CTzN !
no HDR, just a sun at 88°, two scatters

[CTZn]

Thanks for the credit ! Your textures(?) have too much of something, gamma or saturation rather, that conflict somewhat with the sunset tones. I hope you don't mind the critic

[arc en ciel]

you're right ! I updated the image with a less saturated one, especially in green.
this is 3 layers recomposed in photoshop, so there can be many variations.

...that conflict somewhat...

on the other hand, this creates a contrast that I do not deny
thank you for your comment