[Bf-cycles] Rendering with Radio waves
storm
kartochka22 at yandex.ru
Wed Apr 9 15:20:10 CEST 2014
Why not Cycles? It is best integrated with Blender, have very well
tested for years, have two features not related to physical property of
light or anything (think general quasi monte carlo integrator
framework), optimal triangle/ray intersection test and QMC generator.
And it already work on some GPU as well as most CPU on planet.
1. Try to get a bit familiar with general build process, ckeckout
blender git tree, compile it, enjoy awesome real time Cycles preview
wingow.
2. Look at /blender/intern/cycles/kernel/kernel_path.h that contain most
inner integrator loop, function kernel_path_integrate() (there are many
similar named in that file, ignore them for now, they are only for very
special cases)
3. Look at line "bool hit = scene_intersect(kg, &ray, visibility,
&isect);" it is most time consumer, test arbitraty ray with scene and
return hit related data. Think how to reuse it in your case (i think in
can be used as is w/o any modification)
4. A bit more complex is main Monte carlo intergator guts close to the
and of loop. Basicaly, qmc generator for dimension used
"path_state_rng_2D(kg, rng, &state, PRNG_BSDF_U, &bsdf_u, &bsdf_v);"
You can stick with path_rng() instead, it return float random sample
between [0,1] for given dimension. You can reuse it for your case, to
sample direction, or polarisation vector, or spectre, only need to add
new dimension in kernel_types.h file.
>From my Point of view, you just need to add few dimensions to qmc
(wavelength, and for polarisation ), and make special BSDF node to
respect them. Few changes, and you get good interactive playground for
your task, with endless possibilities as saving scene, converting
geometry data from many formats, ability to render it as usual light for
debug reason.
В Ср, 09/04/2014 в 19:39 +1000, Trevor Anderson пишет:
> Thanks heaps everyone, I'll have a look into Lux and pbrt.
> Audio rendering is an interesting idea. Definitely something worth
> considering.
>
> I was hoping that at least initially I could limit myself to
> considering large/smooth enough targets that the optical assumptions
> of light would be sufficient.
> I plan on taking some test images with a mmW radar system to compare
> to, so I will pay close attention to the interference patterns and may
> need to take another approach, thanks for the input.
>
> Cheers,
> Trev.
>
>
>
>
> On Wed, Apr 9, 2014 at 6:56 PM, Ton Roosendaal <ton at blender.org>
> wrote:
> Hi,
>
> For physical accurate wavelength mechanics (millimeter waves
> interference and interaction with matter) you might have to
> build a complete independent system - I'm afraid neither
> cycles or lux would help much here. You probably could
> investigate audio rendering (auralization).
>
> Light is in the nanometres spectrum, which allows a lot of
> assumptions and models to simulate an environment where light
> goes around and how to render it. These assumptions I wouldn't
> make for mm waves.
>
> -Ton-
>
> --------------------------------------------------------
> Ton Roosendaal - ton at blender.org - www.blender.org
> Chairman Blender Foundation - Producer Blender Institute
> Entrepotdok 57A - 1018AD Amsterdam - The Netherlands
>
>
>
> On 9 Apr, 2014, at 8:34, Michael Fox wrote:
>
> > I would recommend looking into Luxrender as it has full
> spectral rendering, and is free and opensource and have very
> nice blender intergration, i would ask the luxrender people to
> help you, they are very helpful
> >
> > On 09/04/14 16:24, Trevor Anderson wrote:
> >> Thanks Lukas.
> >>
> >> I will take a more serious look at that then. :)
> >>
> >> Regards,
> >> Trev.
> >>
> >>
> >> On Wed, Apr 9, 2014 at 2:41 PM, Lukas Stockner
> <lukas.stockner at freenet.de> wrote:
> >> Hi,
> >> for a project like this, I wouldn't go for Cycles since it
> has no spectral rendering support and a quite complicated
> integrator code.
> >> My choice would be PBRT, since it has spectral support,
> clean and simple code and great documentation in form of a
> book :)
> >>
> >> Lukas Stockner
> >>
> >> Am 09.04.2014 01:38 schrieb Trevor Anderson
> <trevor.g.anderson at gmail.com>:
> >> >
> >> > Hi all,
> >> >
> >> > I am fairly new to blender and to development for that
> matter. So if anyone can point me in the right direction that
> would be much appreciated. Specifically if there is perhaps
> another open source ray-tracing tool that might be better for
> what I am looking to do.
> >> >
> >> > I am looking to use cycles to simulate millimetre
> wavelength radio wave imaging. As part of that I suppose I
> would be looking at trying to track polarization of the rays,
> but primarily changing the wavelengths captured by the camera.
> Ideally also exporting range information too.
> >> >
> >> > I was wondering if any of this had been done already,
> even for something like infra-red or hyper-spectral imaging,
> something that would give me a starting point.
> >> >
> >> > Thanks,
> >> > Trev.
> >> _______________________________________________
> >> Bf-cycles mailing list
> >> Bf-cycles at blender.org
> >> http://lists.blender.org/mailman/listinfo/bf-cycles
> >>
> >>
> >>
> >> _______________________________________________
> >> Bf-cycles mailing list
> >>
> >> Bf-cycles at blender.org
> >> http://lists.blender.org/mailman/listinfo/bf-cycles
> >
> > _______________________________________________
> > Bf-cycles mailing list
> > Bf-cycles at blender.org
> > http://lists.blender.org/mailman/listinfo/bf-cycles
>
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org
> http://lists.blender.org/mailman/listinfo/bf-cycles
>
>
>
> _______________________________________________
> Bf-cycles mailing list
> Bf-cycles at blender.org
> http://lists.blender.org/mailman/listinfo/bf-cycles
More information about the Bf-cycles
mailing list