[bodebodebode]

During some experimentation I did notice that the diaphram setting had a lot to do with the DOF.

In fact, it wouldn't have such a big influence unless there's the following bug with 1.6.1
http://www.maxwellrender.com/forum/view ... hp?t=26925
(Btw, this is already fixed in the next version.)

So the tip for 1.6.x is to use circular diaphram ?

[tom]

So the tip for 1.6.x is to use circular diaphram ?

That's not necessary unless you have excessive bokeh which would look incorrect with polygonal.

[kami]

thank you. I must have missed that thread. So I guess, that will fix the whole problem.
Looking forward to the next update then

[kami]

I just wanted to do another quick test... This time with a circular diaphragm:
On the left still the DSLR, in the middle: 6 Blades and on the right: circular diaphragm


It's better than polygonal, but still not very close to reality ... And I thought the circular diaphragm wasn't buggy? So I suppose, the result should have been closer to the DSLR image.
Any thoughts on that?

[KurtS]

FWIW, I get the same result as Mihai, using 1.6.1

f-Stop: 2.8

(added hdri lighting and smoothing angle 91 to the rings to get a little closer to your reference photo)

[kami]

this is f-stop 2.8? my last images were 1.4 (forgot to add that).
Could you post your modified .mxs?

[KurtS]

ok, the file can be downloaded here: http://www.mediafire.com/?ugi9tbeju7z

[KurtS]

I don't know whats going on here. When I reopened the file and tried to render with the SAME SETTINGS as for the image above, it looks like this:

[KurtS]

This is weird.

to the left, kami's file downloaded, opened in Studio, click render. (adjusted shutter and ISO in MXCL because it was so bright)

to the right: saved mxs with new name, click render.

SAME f-stop (1.4) on both renderings! Only thing I've done is to make a save-as copy of the MXS-file.

[kami]

Thanks for your tests KurtS. Very weird behavior indeed.
I made two images with your mxs:
Opened it in studio, changed f-stop to 1.4, diaphragm to circular (top) or polygonal/6 (bottom) and hit render.



Much sharper than your 2.8-version.
A bit weird is also the fact, that the axis of the view has changed a little, which could be caused by the blade-bug.
Somebody of the ML team should analyze the issue ...

[Siegfried Kircheis]

Did you consider the factor of Crop Ratio of DSLR ?
If I'm making it right, that Crop Raio of canon DSLR should be somewhere around x 1.6.

If I set the lens length of camera to 35mm in Rhino, that means it's equilvalent to Canon DSLR in 35x1.6=56mm, and I think it affects the bokeh a little bit.

I've been thinking about this issue for long time, as in real world, that it's rather difficult to achieve blur bokeh from lenses in wide angle ( say 12mm ) , I think the situation in Maxwell is a little bit different.

[kami]

Did you consider the factor of Crop Ratio of DSLR ?
If I'm making it right, that Crop Raio of canon DSLR should be somewhere around x 1.6.

I did consider it by changing the film size in rhino to the equivalent value.

It's been a year and the problem isn't fixed yet ...

[Siegfried Kircheis]

THere are lots of optical design on lenses, and I've read from magazine, that bokeh varies from lenses to lenses, say compare Nikon 50/4 to Canon 50/4, the bokeh effect could be slightly different.

I didn't download your file, so I cannot check out the distance between your camera(Maxwell) and object being focused. But for DSLR, there should be a small symbol on the body to reresent the location of CCD/CMOS for distance calculation, the symbol is similar to "Ø". I guess it helps.