[Bf-cycles] Biased Rendering

Marco G marco.gzt at gmail.com
Wed Jul 2 14:25:10 CEST 2014


I've asked about this few times. Spoke with Brecht directly and he said he
wouldn't be against something like this.
I've got in touch with 2-3 devs to ask for an optionally biased integrator
but nothing came out of it, whether for free or paid.

According to what Brecht said, it is indeed a big project, not much the
first rough implementation  (few weeks for a moderately experienced guy)
but polishing and mantaining wouldn't be straightforward.

There is Mitsuba implementation (but GPL, not direct copy of it
allowed...), tested quite a bit and it has some major problems with
blotches, so it isn't complete i suppose (nothing compared to vray or
corona accuracy)

In Cycles Optimization page there is also a point saying "Use one of the
recent filtering techniques to denoise the image. These gives artifacts in
final renders but can give a better preview with few samples".. he said
these filtering techniques are closely related to irradiance caching, and
some of them can be done online (interactive) and perhaps they could be
"applied" just on secondary rays, limiting visible artifacts, don't know
really.

But all in all i think it's a matter of finding someone who can do such
stuff, there is no manpower for this atm i think.

------------------
As far as archviz/interiors/noise goes, working on it with Cycles daily,
the top features i think would boost speed/quality in this field would be:
- Sampling many lights more efficently with standard path tracing (not
branched);
- Light portals;
- Approximated indirect diffuse rays/GI (irradiance cache,
filtering...can't say)
- Reinhard like tone mapping (?), don't know details but with other engines
if i clamp to 1.00 to avoid burnt areas the image isn't that flat as
Cycles, it maintains good bright spots and good dark areas, need more
testing though;
- Wide gamut color space (?), heard someone might work on it eventually

MG


2014-07-02 12:22 GMT+02:00 Mafteiu Scai Vlad <mafteiuscaivlad at gmail.com>:

> Now that would be nice to have :)
>
> Vlad Mafteiu Scai
>
> Mobile: +45-5333-30-67
>
>
> *Mobile UI Designer, 3D Architectural Vizualizer & Product Visualization
> Artist*
> Email: mafteiuscaivlad at gmail.com
>
> Portfolio: http://www.behance.net/motionseadesigns
>
>
> This e-mail is confidential and is intended solely for the use of the
> recipient.  If you are not the intended recipient, be advised that you have
> received this e-mail in error and that any use, dissemination, forwarding,
> printing, or copying of this e-mail is strictly prohibited.
>
>
> 2014-07-02 13:20 GMT+03:00 boura david <def4d.fr at gmail.com>:
>
> +1000
>>
>> 2014-07-02 12:19 UTC+02:00, x3108 at chello.at <x3108 at chello.at>:
>> > Would it be a big problem to implement a „Redshift-like“ irradiance
>> cache/GI
>> > for biased rendering ?
>> > I recently saw how incredibly fast Redshift on GPU is, it is biased, but
>> > boy… you aren’t even noticing it. And so interactive!
>> > Because Interiors and nearly all renders are incredible fast with
>> irradiance
>> > caches & Co., not to mention pristine clean!
>> >
>> > Now Cycles hast almost every feature set needed, i guess only another
>> biased
>> > kernel need to be dispatched to the CPU/GPU for that ?
>> > Am i right, or would it be too much of a programming nightmare ? Can it
>> be
>> > only a modified cycles code, or does it need a complete reprogramming ?
>> >
>> > Then again, even if it would be some programming, wouldn’t biased be
>> worth
>> > it ? ;-)
>> > Meaning that small artists don’t have a render farm for animation, and
>> BI is
>> > not actual anymore by far all tuts are with cycles.
>> > I mean even Cycles, is not really a true spectral render (and MLT is
>> slow),
>> > basically being RGB biased, so… more biasing would not hurt at all. :-D
>> >
>> > What do you guys say ?
>> >
>> > Flavius
>> >
>> _______________________________________________
>> 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
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/bf-cycles/attachments/20140702/2f8b3278/attachment.htm 


More information about the Bf-cycles mailing list