[samsam]

Mihai - email attachment sent - let me know how you get on!

[Fernando Tella]

This is an important issue. Needs to be fixed as soon as possible.

[sconlogue]

Fixed?? So I went back to my big bump map and changed unchecked Lambertian and set my roughness to 50. Bump @ 20. Jaggies are gone! Still using turbosmooth in max so that wasn't the issue. When I get a minute I will post the offending material and the working material. Stange thing is that I had create other "fabric" Lambertian mats the same way and they worked fine. This was the only one with problems.

[samsam]

Fixed?? - Sadly as you can see the approach of unchecking Lambertian did not work in this particular instance. So yes it does need 'fixing'.

[3dtrialpractice]

this has been driving me crazy
all my geo seems to be doing it now..

Ive tried everything that came to mind.. i've lost 2 day s of my life with this now.. i imported the scene into studio in studio changes all sorts of setti9ngs.. and renderered renderered rerenderered... ahh.. Changed the smoothing angles in studio .. changed smoothing with "recalc normal" none of that worked.. changed scene scales.. changed object scales.. since i use BIG objects forlike cars my torus in my example was 8 meeters wide.. so i reduced it to 8 cm .. even to .01 cms !! ARRH no.. still had those jaggies.. Then i was like its my geo its this scene.. Closed everything.. REBOOT the computer.. created all new geometry from scratch.. STILL have jaggies.. subdividing still has jaggies just makes them smaller ... even in studio i import allnew geo created all new projectors.. tried it on geo without projectors.. no matter what i do i cant render a scene w/o the jaggies.... This Sucks.. im getting pissed .. either at myself or my computer.. frustrated with v1 which is just a beta program still..not to mention that all the plugins are still in beta... and i long wish for the days of alpha when we acctually had an update or fix every week or two..

why do some ppl experiance this jaggies and others not?

i would say they are sun/sky related since they show up in my sun scenes.. almost like distance sun ray clipping problem..

or its geometry related maybe render engine related.. but then why dont everyone get them jagiies.. and they stopped for sconlogue..

[Mihai]

subdividing still has jaggies just makes them smaller

Could you post the scene with the subdivide geometry?

It seems to be sun related and it's strange that it only happens for some people...since this smoothing is really a fake, it's interpolating the shading between those polys so it looks like it's actually highrez, there could be a problem using lowrez objects. In the scene by samsam, I tested and the smoothing problem appears on the lowpoly sphere, although it's pretty low poly and you can see it's outline so I wouldn't use it in a scene. For the other spheres these artifacts aren't visible. I hope this is something that can be improved but until then use at least objects of higher resolution if it's supposed to be curved, so you don't see the poly outline. Lower rez objects wouldn't be acceptable anyway, smoothing artifacts or not.

But it would be interesting to see what it looks like with your subdivided geometry....

[samsam]

In the scene by samsam, I tested and the smoothing problem appears on the lowpoly sphere, although it's pretty low poly and you can see it's outline so I wouldn't use it in a scene. For the other spheres these artifacts aren't visible.

Sorry Mihai - if you look closely it's there on the 40x40 and 60x60 spheres (jpeg compression has masked it a little).

Let's be realistic about this matter - there are plenty of examples of how awful these artifacts look. - Also its worth remembering that they get more evident as the SL's increase. I've frequently rendered test scenes only to find this issue cropping up when I do my final higher res version.

While I appreciate that occasionally it's possible to increase the subdivs on an offending object - this isn't always an option especially with complex geometry / mapping. i.e real production work...

None of this fills me with confidence...

[Mihai]

But I've only seen this on lower poly objects, and if it's not lowpoly there's something about the maps. For example it may be that a very heavy bump breaks the interpolation, even for a higher rez mesh. But again if you already need to make a pretty large image, you need to have at least the geometry so you won't see the edge polygons and in these cases this smoothing problem doesn't seem to apply. That's why I asked for a scene which uses a higher rez objects and still shows this problem.

[LarsSon]

But it's still visible in "cylinder 40". That's really bad i think.

I tried same scene with small distant emitter and same thing happend.
So it's not problem with sun.






-LarsSon

[Hybaj]

I've seen this problem with other renderers and they later fixed with a some kind of a trick. Hope NL programmers can implement this trick into maxwell too... if it's possible ...

[samsam]

But again if you already need to make a pretty large image, you need to have at least the geometry so you won't see the edge polygons and in these cases this smoothing problem doesn't seem to apply. That's why I asked for a scene which uses a higher rez objects and still shows this problem

Mihai, I take it you're saying that if the profile of the poly object looks stepped then you dismiss any suggestion that shadow termination artifacts can be concidered a limitation of the renderer.

I would suggest most modern renderers overcome this shadow termination problem even over fairly large i.e. 18 degree polynormal deltas. I've commited my machine to another overnight render to show how maxwell still exhibits this shadow termination artifact on an 80x80 polysphere (with a polynormal delta of 4.5 degrees) - as you can see even though the profile of the sphere looks perfect - the maxwell rendering is not.

I would surmise from this that anyone working with a curved surface with a interpolygon normal delta of more than say 3 degrees might expect to see unpleasant artifacts on their rendering when using the maxwell sun light system.

Do you think this is a satisfactory situation for a rendering system that proports to produce some of the most realistic looking renders in the world?

[3dtrialpractice]

Mihai, first thanks for your patience..

here are my mxs scenes.. including the original, and subdivided,AND poly smoothed..

http://lukeroman3d.cbeckleydesign.net/M ... 9tests.zip

also I included OBJ format of the scene so that you can load it in any 3d app..

I even tried Triangulating.. still had jaggies...

I did find a "fakefix" in maya to apply "polysmooth" with Expinential smothing NOT linear... this seems to make the jaggies either disapear or they are really small in some areas.. BUT the BIGGEST problems with using polysmooth expinential on your object.. is that it WILL messup you UV's and It may in the case of the Cone it will DEFORM you geometry... Ohh and it still doesnt help my objects that had textures assigned.. essp NORMAL maped shaders..

I dont know how ppl get their normal maps shaders to work .. When ever i apply a normal map to my shader it jut Destroys the shader into Darkness.. not just in my scene but in MXED as well the shader completly goes way dark.. But this is another problem

example of poly smoothing with expinential option set on..

expinential with 2x subdivision of faces.. (acctually roundsoff the polycone at base) but the jaggies seem to almost all be gone.. still a bit left on cone tip,torus inside,and look real close still on rightside of sphere too(very faint)...



here is the EXACT same goe scene.. etc.. exept applid a texture to it..




ohh and ya i tested this with sun AND with emmiters.. emmiters also show jaggies so i guessits not a sun problem...






maybe you mihai or someone else will brighten the situation up... and help me see how this is something that I am doing or someway of fixing.. i wish that this was user error not maxwellerror..

thanks mihai for stickin with us and exploring this problem..

[pBarrelas]

I guess we'll have to wait 'til the next version! I'm having these problems too, so i guess that this isn't a user error!!:(

[mrcharles]

As a reference to this thread... similar discussion of the subject can be found at my threads on this subject...

http://www.maxwellrender.com/forum/view ... highlight=

and this...

http://www.maxwellrender.com/forum/view ... highlight=

I was not encouraged by the results of either thread.

I would love to see this fixed. I like using physical sky + sun for illumination.

Does anyone out there have a test using a 3rd party physical sky + sun HDRI? Do you have the same problem with bump mapped textures?

NL... is there any update on this subject?

[pBarrelas]

mrcharles - thanks for those links on the subject!! I'm having problems with bump maps either! This is trully frustrating, we have a very powerful renderer and then an image can be ruined because of this! I'm spending too much time trying to find a way to go around this.
I'm sure this is a bug NL is taking care!!