POV-Ray : Newsgroups : povray.unofficial.patches : Is anyone working on a distributed / stochastic / Monte Carlo ray-tracing patch for 3.7? Server Time
15 Oct 2024 15:33:33 EDT (-0400)
  Is anyone working on a distributed / stochastic / Monte Carlo ray-tracing patch for 3.7? (Message 8 to 17 of 17)  
<<< Previous 7 Messages Goto Initial 10 Messages
From: Nekar Xenos
Subject: Re: Is anyone working on a distributed / stochastic / MonteCarloray-tracingpatchfor 3.7?
Date: 23 Jul 2014 14:05:38
Message: <op.xjgpvdseufxv4h@xena.home>
On Tue, 22 Jul 2014 14:26:28 +0200, clipka <ano### [at] anonymousorg> wrote:

> Two other things that had always bothered me about MCPov is that it  
> doesn't allow the use of classic light source (which would be far less  
> noisy and hence much faster than using bright spheres), and that it has  
> a factor-2 error in diffuse computations that make it necessary to use  
> different finish settings. Both make it excessively difficult to create  
> scenes that render essentially identical (except for artifacts or noise)  
> in both POV-Ray and MCPov. Needless to say that UberPOV is intended to  
> do a better job in that respect.
>

And there was also a problem with media. I hope UberPov will be able to do  
what MCPov did and still be able to render different types of media  
correctly.

-- 
-Nekar Xenos-


Post a reply to this message

From: Nekar Xenos
Subject: Re: Is anyone working on a distributed / stochastic / MonteCarloray-tracing patch for 3.7?
Date: 23 Jul 2014 14:12:54
Message: <op.xjgp7gp3ufxv4h@xena.home>
On Tue, 22 Jul 2014 00:53:47 +0200, clipka <ano### [at] anonymousorg> wrote:

> Am 21.07.2014 21:12, schrieb Nekar Xenos:
>> On Sun, 20 Jul 2014 20:45:27 +0200, clipka <ano### [at] anonymousorg>  
>> wrote:
>>
>>> Am 20.07.2014 19:44, schrieb Nekar Xenos:
>>>>
>>>> Is anyone working on a distributed / stochastic / Monte
>>>> Carloray-tracing patch for 3.7?
>>>
>>> Distributed raytracing patch: Planned for official POV-Ray; don't hold
>>> your breath though.
>>>
>>> Stochastic / Monte Carlo raytracing patch: Maybe UberPOV
>>> (https://github.com/UberPOV/UberPOV) already offers what you're
>>> looking for; it already uses stochastic rendering for various
>>> features. If stochastic global illumination is what you're after, just
>>> drop me a note and I'll make that happen as well.
>>>
>>
>> I'm not 100% sure what the correct term is, the Wikipedia entry was a
>> bit confusing. I can describe it as following the ray from the light
>> source to the screen. Just like MCPov, only with all of the 3.7  
>> features.
>
> Hmm, there's definitely some misunderstanding going on here.
>
> "Following the ray from the light source to the screen" is /forward  
> raytracing/, as opposed to following the ray from the screen to the  
> object surface and ultimately to the light source, aka /backward  
> raytracing/. The latter is the classic approach, and the one used by  
> POV-Ray.
>
> MCPov also uses /backward/ raytracing like POV-Ray.
>
I don't know why I thought MCPov used forward ray-tracing. I think it was  
probably because of the slow rendering and good results after waiting a  
long time ;)

> /Stochastic/ raytracing can be done with both approaches (I guess it is  
> the only way to do forward raytracing, but it is also a way to do  
> backward raytracing), or even with combinations of the two, such as  
> bidirectional raytracing, Metropolis Light Transport or some such.
>
>
> So let's forget about the technical terms, and start at the other end:
>
> What is it about MCPov that you actually want?
>
> Blurred reflections? Definitely go for UberPOV, as it has them. Yay!
>
> Blurred refractions? Not supported by UberPOV yet, but shouldn't be too  
> difficult to add once I decide on a nice syntax.

That would be useful.
>
> A less artifact-prone replacement for the effect that official POV-Ray  
> uses radiosity for (so-called Global Illumination)? Raise your hand, and  
> I'll make it happen in UberPOV.
>

*raises hand*

> Render over and over again while you sit and watch, until you're happy  
> with the result? Not supported in UberPOV yet; it's on the agenda, but  
> may require quite some intrusion into the POV-Ray architecture. What  
> UberPOV does already support, however, is an "anti-aliasing" mode (or,  
> more to the point, an oversampling mode) that allows you to specify what  
> quality you'd be happy with, and UberPOV will do the  
> rendering-over-and-over-again in the background until the quality  
> criteria are met.
>

Yes, please :)
>
> As a rule of thumb, if MCPov has it, it is also planned (if not already  
> implemented) for UberPOV.
>
Whoo-hoo!



-- 
-Nekar Xenos-


Post a reply to this message

From: Nekar Xenos
Subject: Re: Is anyone working on a distributed / stochastic / Monte Carloray-tracing patch for 3.7?
Date: 23 Jul 2014 14:18:44
Message: <op.xjgqg7fsufxv4h@xena.home>
On Sun, 20 Jul 2014 20:45:27 +0200, clipka <ano### [at] anonymousorg> wrote:

> Am 20.07.2014 19:44, schrieb Nekar Xenos:
>>
>> Is anyone working on a distributed / stochastic / Monte  
>> Carloray-tracing patch for 3.7?
>
> Distributed raytracing patch: Planned for official POV-Ray; don't hold  
> your breath though.
>
> Stochastic / Monte Carlo raytracing patch: Maybe UberPOV  
> (https://github.com/UberPOV/UberPOV) already offers what you're looking  
> for; it already uses stochastic rendering for various features. If  
> stochastic global illumination is what you're after, just drop me a note  
> and I'll make that happen as well.
>

What I would like to see with radiosity is a system that distinguishes  
between different materials. Smooth shiny materials should have a smaller  
cone of scattering than cement which should have a very wide angle of  
scattering.

-- 
-Nekar Xenos-


Post a reply to this message

From: clipka
Subject: Re: Is anyone working on a distributed / stochastic / MonteCarloray-tracing patch for 3.7?
Date: 23 Jul 2014 15:52:39
Message: <53d01287$1@news.povray.org>
Am 23.07.2014 20:18, schrieb Nekar Xenos:

> What I would like to see with radiosity is a system that distinguishes
> between different materials. Smooth shiny materials should have a
> smaller cone of scattering than cement which should have a very wide
> angle of scattering.

Well... no, not really. In smooth shiny materials, the primary direction 
of the cone of scattering virtually always depends on the angle of the 
incoming light, approaching classic specular reflection as the object 
gets even smoother.

To model such cases, POV-Ray's radiosity approach is inherently 
unsuited, because the sample information needs to be very 
memory-efficient and thus quite limited - POV-Ray only stores a single 
weighted average of all the light coming in, so any information about 
the direction of the incoming light is lost entirely.

What really kicks ass for smooth shiny (but not perfectly polished) 
materials is blurred reflections; with a strong blur it can even give a 
subtle edge to comparatively dull materials


That doesn't mean your idea would be all nonsense though - see the stuff 
I happen to have just shown off in the "Brilliant Spheres" thread on 
povray.binary.images ;-)

(Actually it's not /that/ much of a coincidence: As a first approach at 
"replacing" radiosity I plan to actually just keep using it but not 
cache the results (plus make sure that the secondary rays shot are 
different each time); and as I want UberPOV to fully support non-default 
brilliance settings, I figured that I might just as well add that to 
POV-Ray's radiosity first... And then one thing led to another, and I 
found myself adding all those fancy finish improvements that had been on 
my agenda for so long...)


Post a reply to this message

From: Alain
Subject: Re: Is anyone working on a distributed / stochastic / MonteCarloray-tracingpatch for 3.7?
Date: 23 Jul 2014 17:21:57
Message: <53d02775@news.povray.org>


> (Actually it's not /that/ much of a coincidence: As a first approach at
> "replacing" radiosity I plan to actually just keep using it but not
> cache the results (plus make sure that the secondary rays shot are
> different each time); and as I want UberPOV to fully support non-default
> brilliance settings, I figured that I might just as well add that to
> POV-Ray's radiosity first... And then one thing led to another, and I
> found myself adding all those fancy finish improvements that had been on
> my agenda for so long...)
>

And you did a marvelous job at that. Tank you very much :)


Post a reply to this message

From: Nekar Xenos
Subject: Re: Is anyone working on a distributed / stochastic / MonteCarloray-tracing patch for 3.7?
Date: 24 Jul 2014 11:58:37
Message: <op.xjienol6ufxv4h@xena.home>
On Wed, 23 Jul 2014 21:52:32 +0200, clipka <ano### [at] anonymousorg> wrote:


> (Actually it's not /that/ much of a coincidence: As a first approach at  
> "replacing" radiosity I plan to actually just keep using it but not  
> cache the results (plus make sure that the secondary rays shot are  
> different each time); and as I want UberPOV to fully support non-default  
> brilliance settings, I figured that I might just as well add that to  
> POV-Ray's radiosity first... And then one thing led to another, and I  
> found myself adding all those fancy finish improvements that had been on  
> my agenda for so long...)
>

You're our hero :)

-- 
-Nekar Xenos-


Post a reply to this message

From: clipka
Subject: Re: Is anyone working on a distributed / stochastic /MonteCarloray-tracing patch for 3.7?
Date: 24 Jul 2014 13:53:24
Message: <53d14814$1@news.povray.org>
Am 24.07.2014 17:58, schrieb Nekar Xenos:
> On Wed, 23 Jul 2014 21:52:32 +0200, clipka <ano### [at] anonymousorg> wrote:
>
>
>> (Actually it's not /that/ much of a coincidence: As a first approach
>> at "replacing" radiosity I plan to actually just keep using it but not
>> cache the results (plus make sure that the secondary rays shot are
>> different each time); and as I want UberPOV to fully support
>> non-default brilliance settings, I figured that I might just as well
>> add that to POV-Ray's radiosity first... And then one thing led to
>> another, and I found myself adding all those fancy finish improvements
>> that had been on my agenda for so long...)
>>
>
> You're our hero :)

Thanks. Sometimes I am. ;-)


Post a reply to this message

From: jhu
Subject: Re: Is anyone working on a distributed / stochastic /MonteCarloray-tracing =
Date: 1 Sep 2014 12:20:00
Message: <web.54049bc4b743e78d19b0ec40@news.povray.org>
clipka <ano### [at] anonymousorg> wrote:
> Am 24.07.2014 17:58, schrieb Nekar Xenos:
> > On Wed, 23 Jul 2014 21:52:32 +0200, clipka <ano### [at] anonymousorg> wrote:
> >
> >
> >> (Actually it's not /that/ much of a coincidence: As a first approach
> >> at "replacing" radiosity I plan to actually just keep using it but not
> >> cache the results (plus make sure that the secondary rays shot are
> >> different each time); and as I want UberPOV to fully support
> >> non-default brilliance settings, I figured that I might just as well
> >> add that to POV-Ray's radiosity first... And then one thing led to
> >> another, and I found myself adding all those fancy finish improvements
> >> that had been on my agenda for so long...)
> >>
> >
> > You're our hero :)
>
> Thanks. Sometimes I am. ;-)

How difficult would it be to adapt this to GPU rendering via OpenCL or CUDA?


Post a reply to this message

From: clipka
Subject: Re: Is anyone working on a distributed / stochastic /MonteCarloray-tracing =
Date: 1 Sep 2014 15:57:39
Message: <5404cfb3@news.povray.org>
Am 01.09.2014 18:16, schrieb jhu:
> clipka <ano### [at] anonymousorg> wrote:
>> Am 24.07.2014 17:58, schrieb Nekar Xenos:
>>> On Wed, 23 Jul 2014 21:52:32 +0200, clipka <ano### [at] anonymousorg> wrote:
>>>
>>>
>>>> (Actually it's not /that/ much of a coincidence: As a first approach
>>>> at "replacing" radiosity I plan to actually just keep using it but not
>>>> cache the results (plus make sure that the secondary rays shot are
>>>> different each time); and as I want UberPOV to fully support
>>>> non-default brilliance settings, I figured that I might just as well
>>>> add that to POV-Ray's radiosity first... And then one thing led to
>>>> another, and I found myself adding all those fancy finish improvements
>>>> that had been on my agenda for so long...)
>>>>
>>>
>>> You're our hero :)
>>
>> Thanks. Sometimes I am. ;-)
>
> How difficult would it be to adapt this to GPU rendering via OpenCL or CUDA?

Very. At least that's what I still suspect.


Post a reply to this message

From: jhu
Subject: Re: Is anyone working on a distributed / stochastic /MonteCarloray-tracing =
Date: 2 Sep 2014 09:55:01
Message: <web.5405cb1c65080cc0d19b0ec40@news.povray.org>
clipka <ano### [at] anonymousorg> wrote:
> Am 01.09.2014 18:16, schrieb jhu:
> > clipka <ano### [at] anonymousorg> wrote:
> >> Am 24.07.2014 17:58, schrieb Nekar Xenos:
> >>> On Wed, 23 Jul 2014 21:52:32 +0200, clipka <ano### [at] anonymousorg> wrote:
> >>>
> >>>
> >>>> (Actually it's not /that/ much of a coincidence: As a first approach
> >>>> at "replacing" radiosity I plan to actually just keep using it but not
> >>>> cache the results (plus make sure that the secondary rays shot are
> >>>> different each time); and as I want UberPOV to fully support
> >>>> non-default brilliance settings, I figured that I might just as well
> >>>> add that to POV-Ray's radiosity first... And then one thing led to
> >>>> another, and I found myself adding all those fancy finish improvements
> >>>> that had been on my agenda for so long...)
> >>>>
> >>>
> >>> You're our hero :)
> >>
> >> Thanks. Sometimes I am. ;-)
> >
> > How difficult would it be to adapt this to GPU rendering via OpenCL or CUDA?
>
> Very. At least that's what I still suspect.

Oh well. Knight's Landing is just around the corner so OpenCL/CUDA might be
moot, but hopefully it's not too expensive.


Post a reply to this message

<<< Previous 7 Messages Goto Initial 10 Messages

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.