Please post here anything else (not relating to Maxwell technical matters)
#366909
Hi Mihai,
The dof in the leaves sample really looks quite good! Still the unsharp portions for my taste still lack a certain luminous sparkle and burn quality when compared to a real image.
It's colder and also quite a tad noisier. The gears example is too simple to expose subtle effects such as what's going on on the hairs of the tomato crop.
While the overall image is a great achievement I'm not sure whether the DOF in the Superbaka image isn't edited in post, I see several (nicely done) post effects in his renders.

All I can honestly say that I'm often completely blown away with what one can do with a wide open photographic lense and very little effort.
In comparison I'm always a bit disappointed about the sober appearance of rendered dof (not just limited to Maxwell) There's also many physical lenses which create boring dof - here the choice of
measures is easy - I would not buy it :)

As said before I could well live with post effect on dof such as our current scattering (for as long as we could preview it while the rendering is running).
Maybe something along the lines of what Nox offers could be looked into?

I have no time currently to render sample scenes but I think replicating something along the lines of the tomato crop was a fantastic challenge for anyone who' d like to try this.
Last edited by Polyxo on Mon Apr 08, 2013 12:54 pm, edited 2 times in total.
#366915
I think Maxwell can come very close to what you you have posted but it is missing photometric burnout, and the aperture is lacking in character.
Image
Image
Image
(Images from DOFPro)

This is something I have been figuring out how to fake for ages, though at the moment the only way to achieve this is to f*** around with the bokeh balls in post.
It would be nice to see custom apertures and spherical aberration employed in Maxwell but I cant image it being easy to make this a reality. Photometric burnout on the other hand I am eagerly waiting for as an addition to simulens
#366919
neither, and i'd need to turn off the realtime scanner to view it, which i don't really want to do.

probably it is a false positive, but it tells me it is listed as site with potentially harmful content, so i was wondering if anybody with a different scanners gets any hits too?
#366931
bograt wrote:I think Maxwell can come very close to what you you have posted but it is missing photometric burnout, and the aperture is lacking in character.
(I'm getting that warning from ESET also)

The Maxwell camera model does get this "photometric burnout" or the DOF wouldn't work correctly, but I think from one of the images in that site, after searching through Google (http://dc111.4shared.com/doc/pCkveBiW/preview009.png), of the chess pieces also shows the localized effect of lens flare in my opinion. Depending on the lens of course you get more or less of it but it's something you usually try to avoid in a lens design, although in this case it makes things look more interesting and "analog".

So there are always improvements to be made, I just wouldn't be so harsh as saying it's totally not like the real thing. Two more examples I quickly made now:

Image

Image
#366936
Hi Mihai,
thanks a lot for these new images! They nicely demonstrate the burnout and especially the first one shows very good looking and
unusual clean DOF indeed! I wonder how Rustebergs plants would perform with this setup.

Please understand that I never wanted to sound harsh or even bashing and that I did not limit my observation to just Maxwell.
I am very impressed what this product can do but when one also photographs it's probably normal to compare results of both approaches.

Holger
#366939
Mihai, I think your images look very accurate.
As you said I think the effect I was referring too is down to localized lens flare rather than the burnout you get on out of focus objects, it is something that bothers me in scenes with very deep depth of field but maybe those effects are negligible in this discussion.
For the record I have always been amazed by the accuracy of Maxwells camera model but there is no harm in analyzing its features against real, imperfect cameras/lenses.
#366950
Well this is one of my favorite topics :) The "lens" system must be part of the rendering solution to produce photographic results and really this is what we want - we want it to look like a photo, this is what our acceptance of "real" is based on since we have never seen anything captured without a lens system, including our own eyes, and all are imperfect. So far I think Maxwell does one of the best representations out there but its certainly not the end of the story.

Related to this, I just want to mention that you should look at the render in Maxwell as a cameras raw file. You should first of all change the curves a bit to bring some contrast and it can change the perceptions a lot. For instance those orange chess pieces looked a little bland straight from render but just making a slight S curve it totally changed the look in my opinion, along with the "smoothness" impression of the bokeh. The good thing is if you work with the MXI or another 32b or 16b file you have a lot of headroom for tone mapping the output.
#367021
its certainly not the end of the story...
Yeah, while I think the DOF on my rendered hairy plants doesn't look shabby one can see that there's quite a bit more going on in the tomato detail photo on the previous page.
What I can say though that I could not improve the result by outputting a Zbuffer and postprocessing the image (Focal Point in this case but I'm sure Dof Pro would fail as well).
you should look at the render in Maxwell as a cameras raw file
Sure. These plant details were the first images in a long time I left as "straight from camera".

ok thanks for explaining. actually I do copy the T[…]

Sketchup 2026 Released

Fernando wrote: " Now that Maxwell for Cinema[…]

Hello Gaspare, I could test the plugin on Rhino 8[…]

Hello Blanchett, I could reproduce the problem he[…]