[Sidefx-houdini-list] Shading artifact when really close to surface

guillaume laforge guillaume at alamaison.fr
Thu Oct 22 09:50:02 EDT 2009


"vm_binarysplit" doesn't help solving patch cracks in my case :-/.

Tweaking "vm_shadingquality" and/or "vm_measurescale" give me a closer look
to the one using "Non-raster Measuring" but not the same quality.
If I increase shading to a very big number, the closest area are better but
the horizon is very very long to render (and still not the same quality on
very near parts...).

Cheers

Guillaume


On Thu, Oct 22, 2009 at 3:15 PM, Francois Duchesneau <sidefx at trinix.ca>wrote:

> I wonder if you wouldn't have to increase the shading quality a lot more
> than what you usually do with "non uniform" because now it's not dicing
> based on the pixel size. Or maybe there's an other property to add to
> specify that level of dicing.
>
> You can also try "vm_binarysplit" if you haven't already done it.
>
> François
>
> > For info, the scanline measuring set to "uniform" give a perfectly
> > smoothed
> > surface but it is not usable if you need to change N and/or P in your
> > material...
> > You lost all the details of a displacement shader for example.
> >
> > I also tried to use a sop transformation instead of objects
> transformation
> > but there is still some shading artifact.
> >
> > Finally, as my surface (it is an ocean) is perturbed by a lot of noise,
> > the
> > shading artifacts are not really distinguishable.
> >
> > But this dicing problem is still a mystery to me...
> >
> > Cheers
> >
> > Guillaume
> >
> > On Wed, Oct 21, 2009 at 9:11 PM, guillaume laforge
> > <guillaume at alamaison.fr>wrote:
> >
> >> François :so the info was not so hidden in the doc :). Thanks again for
> >> the links.
> >> Ivan :
> >> I tried to subdivide the sphere but it doesn't help much. It was not a
> >> primitive sphere as UVs were needed.
> >>
> >> If you download my sample scene, you will see that I moved the camera to
> >> the center of the world and that the surface artifacts are close to the
> >> center of world too.
> >> I'm not sure it is a floating point error.
> >>
> >> Andrew :
> >> I'm not 100% sure but in one of my test to remove those artifacts(before
> >> François help me with the vm_measure parameter),
> >> I tried to scale down my big sphere by parenting all objects under a
> >> null
> >> scaled down to 0.01. As It doesn't help, I created a new object and get
> >> the
> >> scalled down sphere using an object merge. From this new object the
> >> shading
> >> was better. I guess some transformation are more accurate at sop level ?
> >>
> >> Cheers
> >>
> >> Guillaume
> >>
> >>
> >> On Wed, Oct 21, 2009 at 7:50 PM, Francois Duchesneau
> >> <sidefx at trinix.ca>wrote:
> >>
> >>> http://www.sidefx.com/docs/houdini10.0/rendering/dicing
> >>> http://www.sidefx.com/docs/houdini10.0/props/mantra10_0
> >>>
> >>> > It works :)
> >>> >
> >>> > I tried to add and tweak randomly different combination but didn't
> >>> get
> >>> the
> >>> > good one.
> >>> >
> >>> > Is there any doc about all those dicing parameters ?
> >>> >
> >>> > Thanks a lot Francois !!!
> >>> >
> >>> > Cheers
> >>> >
> >>> > Guillaume
> >>> >
> >>> > On Wed, Oct 21, 2009 at 6:48 PM, Francois Duchesneau
> >>> > <sidefx at trinix.ca>wrote:
> >>> >
> >>> >> If you add the vm_measure to your object and set it to "Uniform
> >>> >> Measuring", the problem goes away. It's probably less efficient
> >>> because
> >>> >> it
> >>> >> will have to dice too much far away.
> >>> >>
> >>> >> Let's hope it won't introduce artifacts elsewhere.
> >>> >>
> >>> >> François
> >>> >>
> >>> >> > My scene is just a sphere :).
> >>> >> >
> >>> >> > For info, here is a sample file showing the problem :
> >>> >> > http://www.vol2nuit.fr/repro_shading_pb.hip
> >>> >> >
> >>> >> > Cheers
> >>> >> >
> >>> >> > Guillaume
> >>> >> >
> >>> >> > On Wed, Oct 21, 2009 at 4:01 PM, Francois Duchesneau
> >>> >> > <sidefx at trinix.ca>wrote:
> >>> >> >
> >>> >> >> In that case what I don't understand is if I'm creating a sphere
> >>> with
> >>> >> a
> >>> >> >> radius of 1 with a close up camera aiming at the top of the
> >>> sphere,
> >>> I
> >>> >> >> won't get artifacts.
> >>> >> >>
> >>> >> >> This means there's must be something in particular that causes
> >>> this.
> >>> >> In
> >>> >> >> my
> >>> >> >> case all the fractured piece of geometries were separated. Is
> >>> your
> >>> >> >> sphere
> >>> >> >> made up of different polygons not fused together?
> >>> >> >>
> >>> >> >> François
> >>> >> >>
> >>> >> >>
> >>> >> >> > Hi François,
> >>> >> >> >
> >>> >> >> > Thanks for the info on your similar experience !
> >>> >> >> >
> >>> >> >> > I tried to scale down all except the camera but it doesn't
> >>> solve
> >>> >> the
> >>> >> >> > problem
> >>> >> >> > :-/.
> >>> >> >> >
> >>> >> >> > Btw, it is using 9.5.310. Didn't try in 10 yet.
> >>> >> >> >
> >>> >> >> > Cheers
> >>> >> >> >
> >>> >> >> > Guillaume
> >>> >> >> >
> >>> >> >> > 2009/10/21 François Duchesneau <sidefx at trinix.ca>
> >>> >> >> >
> >>> >> >> >> I ran into similar issues too.
> >>> >> >> >>
> >>> >> >> >> I wanted to render a pre-fractured ground with camera
> >>> projection
> >>> >> but
> >>> >> >> >> many artifacts were visible around the seams. After many hours
> >>> of
> >>> >> >> >> research I tested the same in Maya and it worked so I decided
> >>> to
> >>> >> test
> >>> >> >> it
> >>> >> >> >> with Ray Tracing in Mantra and like you the problem disappear.
> >>> My
> >>> >> >> >> geometry had a big scale too and was also tangent to the
> >>> camera.
> >>> >> >> Maybe
> >>> >> >> >> there are some tesselation settings out there in the
> >>> properties
> >>> to
> >>> >> >> try
> >>> >> >> >> out.
> >>> >> >> >>
> >>> >> >> >> I thought it might be an issue with micro-polygon but honestly
> >>> I'd
> >>> >> be
> >>> >> >> >> surprised RenderMan had the same issue. Unfortunately I don't
> >>> have
> >>> >> it
> >>> >> >> to
> >>> >> >> >> test it.
> >>> >> >> >>
> >>> >> >> >> Out of curiosity, I guess if you recreate that kind of render
> >>> in
> >>> a
> >>> >> >> >> smaller scale with a non-scaled camera you won't find any
> >>> >> artifact.
> >>> >> >> I'm
> >>> >> >> >> asking because one day I had to scale a whole scene including
> >>> the
> >>> >> >> camera
> >>> >> >> >> and a volume and I had to increase the volume sample a lot to
> >>> make
> >>> >> it
> >>> >> >> >> render at a reasonable time.
> >>> >> >> >>
> >>> >> >> >> Good luck and let me know if you find anything.
> >>> >> >> >>
> >>> >> >> >> François
> >>> >> >> >>
> >>> >> >> >> guillaume laforge wrote:
> >>> >> >> >> > Hi,
> >>> >> >> >> >
> >>> >> >> >> > I've got some shading artifacts when using micropolygon
> >>> >> rendering
> >>> >> >> with
> >>> >> >> >> a
> >>> >> >> >> > camera really close to a geometry.
> >>> >> >> >> > The camera is tangent to the surface of a sphere. The sphere
> >>> >> radius
> >>> >> >> is
> >>> >> >> >> 6371
> >>> >> >> >> > unit with a white lambert shader.
> >>> >> >> >> > When the camera distance to the surface is around four
> >>> units,
> >>> >> some
> >>> >> >> >> patterns
> >>> >> >> >> > appeared in the shading of the sphere.
> >>> >> >> >> > At a distance of one unit, it is really "bumpy" !
> >>> >> >> >> > If I'm rendering polygons as subdivision surfaces it is
> >>> worst
> >>> >> (and
> >>> >> >> >> when
> >>> >> >> >> need
> >>> >> >> >> > to render this way).
> >>> >> >> >> > If I increase the shading quality, it is worst too.
> >>> >> >> >> > If I switch to Raytracing renderer, the problem disappear.
> >>> >> >> >> Unfortunately
> >>> >> >> >> we
> >>> >> >> >> > need to use Micorpolygon Rendering for this project.
> >>> >> >> >> >
> >>> >> >> >> > I tried to scale down all the scene but it doesn't work. I
> >>> guess
> >>> >> >> the
> >>> >> >> >> problem
> >>> >> >> >> > came form the ratio between sphere radius and camera/surface
> >>> >> >> distance
> >>> >> >> >> being
> >>> >> >> >> > too big...
> >>> >> >> >> >
> >>> >> >> >> > Is there any magic settings to remove this kind of artifact,
> >>> or
> >>> >> is
> >>> >> >> it
> >>> >> >> >> a
> >>> >> >> >> > precision limitation of the micropolygon system.
> >>> >> >> >> >
> >>> >> >> >> > Thanks for any advice,
> >>> >> >> >> >
> >>> >> >> >> > Cheers.
> >>> >> >> >> >
> >>> >> >> >> >
> >>> >> >> >> _______________________________________________
> >>> >> >> >> Sidefx-houdini-list mailing list
> >>> >> >> >> Sidefx-houdini-list at sidefx.com
> >>> >> >> >>
> >>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >> >> >>
> >>> >> >> > _______________________________________________
> >>> >> >> > Sidefx-houdini-list mailing list
> >>> >> >> > Sidefx-houdini-list at sidefx.com
> >>> >> >> >
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >> >> >
> >>> >> >>
> >>> >> >>
> >>> >> >> _______________________________________________
> >>> >> >> Sidefx-houdini-list mailing list
> >>> >> >> Sidefx-houdini-list at sidefx.com
> >>> >> >>
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >> >>
> >>> >> > _______________________________________________
> >>> >> > Sidefx-houdini-list mailing list
> >>> >> > Sidefx-houdini-list at sidefx.com
> >>> >> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >> >
> >>> >>
> >>> >>
> >>> >> _______________________________________________
> >>> >> Sidefx-houdini-list mailing list
> >>> >> Sidefx-houdini-list at sidefx.com
> >>> >> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >>
> >>> > _______________________________________________
> >>> > Sidefx-houdini-list mailing list
> >>> > Sidefx-houdini-list at sidefx.com
> >>> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>> >
> >>>
> >>>
> >>> _______________________________________________
> >>> Sidefx-houdini-list mailing list
> >>> Sidefx-houdini-list at sidefx.com
> >>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >>>
> >>
> >>
> >>
> >>
> >>
> > _______________________________________________
> > Sidefx-houdini-list mailing list
> > Sidefx-houdini-list at sidefx.com
> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >
>
>
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>



More information about the Sidefx-houdini-list mailing list