 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 10-1-2014 9:59, Ive wrote:
> Am 10.01.2014 09:21, schrieb Thomas de Groot:
>> The bright dots are still a problem indeed. They are also present in
>> Christoph's image. I wonder if this is not light leaking along mesh
>> edges....
>>
> It is light leaking and it gets even worse when using higher quality
> (i.e. lower error_bound) settings. I guess the only solution is to
> remodel the room and give the walls and the ceiling some real thickness,
> thats what I did. As with CSG where it is a good idea to let walls and
> ceiling overlap.
You confirm my own guess. I am going to process the walls and pillars
objects in Silo.
>
>
>> presently, radiosity settings are:
>>
>> radiosity {
>> pretrace_start 0.08
>> pretrace_end 0.004
>> count 500, 1000
>> nearest_count 10, 5
>> error_bound 1
>> recursion_limit 4
>> low_error_factor 0.3
>> gray_threshold 0.0
>> minimum_reuse 0.015
>> maximum_reuse 0.1
>> brightness 1
>> adc_bailout 1/100000
>> normal on
>> media off
>> always_sample off
>> //max_sample 1.0
>> }
>>
>
> When you add no_radiosity to all window/door glasses you'll get away
> with 2 recursions and faster render speed - without any visual impact.
Good idea. I didn't think of that possibility.
> And I'm not sure about this low adc_bailout within the radiosity block,
> if such low value is needed shouldn't it go into the global settings
> block itself - to keep the photons visible?
> Maybe Christoph can shed some light on this.
I put in that low adc_bailout in preparation to the photons/caustics
generation, as Christoph suggested earlier in this thread. I am a newbie
as far as this is concerned.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 10.01.2014 10:11, schrieb Thomas de Groot:
>> And I'm not sure about this low adc_bailout within the radiosity block,
>> if such low value is needed shouldn't it go into the global settings
>> block itself - to keep the photons visible?
>
> I put in that low adc_bailout in preparation to the photons/caustics
> generation, as Christoph suggested earlier in this thread.
>
Yes, thats what I'm saying too, but it seems you missed the "global
settings block itself" - and not the radiosity block within the global
settings.
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 10.01.2014 09:21, schrieb Thomas de Groot:
>> Still, to me the overall light levels of the shadowed parts seems
>> somewhat too low (but this might be a question of my system's gamma -
>> currently 2.0)... and there are still some bright artifact dots along
>> the upper edges of the walls. What radiosity settings did you use?
>
> I agree about the light levels but I don't know how to compensate that
> other than by an additional light (I don't like) or by a stronger light
> maybe (I shall try). However, I personally like this and my impression
> is that this is a /natural/ situation. Comments welcome.
The proper approach to compensate would be to increase the walls'
brightness. /If/ you want to compensate at all. Looks natural enough to
me, too.
> The bright dots are still a problem indeed. They are also present in
> Christoph's image. I wonder if this is not light leaking along mesh
> edges.... Comments welcome.
It's not specifically leaking along mesh edges, but through /any/ edges.
Surprisingly, similar artifacts can be observed with CSG as well. Up to
this day I haven't really figured out what's causing this, but the
number of radiosity bounces (i.e. recursion_limit) does seem to play a
role (I think I've never seen these artifacts with recursion_limit 2 or
lower).
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 10.01.2014 10:18, schrieb Ive:
> Am 10.01.2014 10:11, schrieb Thomas de Groot:
>>> And I'm not sure about this low adc_bailout within the radiosity block,
>>> if such low value is needed shouldn't it go into the global settings
>>> block itself - to keep the photons visible?
>>
>> I put in that low adc_bailout in preparation to the photons/caustics
>> generation, as Christoph suggested earlier in this thread.
>>
> Yes, thats what I'm saying too, but it seems you missed the "global
> settings block itself" - and not the radiosity block within the global
> settings.
Actually I proposed to use such a low value in the global settings
/photons/ block.
The general global settings adc_bailout, as well as radiosity
adc_bailout, should both be ok with more conservative values.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 10-1-2014 12:47, clipka wrote:
> Am 10.01.2014 10:18, schrieb Ive:
>> Am 10.01.2014 10:11, schrieb Thomas de Groot:
>>>> And I'm not sure about this low adc_bailout within the radiosity block,
>>>> if such low value is needed shouldn't it go into the global settings
>>>> block itself - to keep the photons visible?
>>>
>>> I put in that low adc_bailout in preparation to the photons/caustics
>>> generation, as Christoph suggested earlier in this thread.
>>>
>> Yes, thats what I'm saying too, but it seems you missed the "global
>> settings block itself" - and not the radiosity block within the global
>> settings.
>
> Actually I proposed to use such a low value in the global settings
> /photons/ block.
>
> The general global settings adc_bailout, as well as radiosity
> adc_bailout, should both be ok with more conservative values.
>
Right. I missed that, as I am now entering unknown territory ;-)
I have some read-up on the matter to do, and some testing.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 10-1-2014 12:44, clipka wrote:
> The proper approach to compensate would be to increase the walls'
> brightness. /If/ you want to compensate at all. Looks natural enough to
> me, too.
I agree and, as far as I am concerned, I do not want to compensate.
Increasing luminosity helps too while keeping the balance between light
and dark within acceptable limits.
>
>> The bright dots are still a problem indeed. They are also present in
>> Christoph's image. I wonder if this is not light leaking along mesh
>> edges.... Comments welcome.
>
> It's not specifically leaking along mesh edges, but through /any/ edges.
> Surprisingly, similar artifacts can be observed with CSG as well. Up to
> this day I haven't really figured out what's causing this, but the
> number of radiosity bounces (i.e. recursion_limit) does seem to play a
> role (I think I've never seen these artifacts with recursion_limit 2 or
> lower).
>
Well, from my latest render (not included here) I must say that leaking
also appears with recursion_limit 2.
Iirc, Christoph Hormann treated this problem in Megapov with a radiosity
patch.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
From: Thomas de Groot
Subject: Re: LanuHum Global Test image - wip 3
Date: 11 Jan 2014 07:11:37
Message: <52d134f9@news.povray.org>
|
|
|
| |
| |
|
|
Does not seem much but I got rid of the light leaking. Following was done:
- meshes (walls, archs, windows) inc files through Poseray export as obj
files (weld vertices and recalculate normals);
- in Silo: giving thickness to the walls with the Shell utility (this
probably exists also in Blender?);
- in Silo: closing of the arch object and extending it a little bit
above the ceiling;
- in Silo: lowering the sky-windows which were /floating/ just above the
building letting light leak in-between;
- saving the different objects as obj files and through Poseray export
to POV as mesh2 files (recalculate normals);
- in the POV scene: replace the mesh{} objects by the mesh2{} ones;
Everything seems now to be in place to tackle the photons and caustics.
Thomas
Post a reply to this message
Attachments:
Download 'test.png' (558 KB)
Preview of image 'test.png'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Thomas de Groot <tho### [at] degroot org> wrote:
>
> - in the POV scene: replace the mesh{} objects by the mesh2{} ones;
>
Povray slows down the miscalculation if matrix is used
I didn't find the solution to calculate normal*matrix for mesh2
Maurice Raybaud it is possible will make
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 11.01.2014 13:11, schrieb Thomas de Groot:
> Does not seem much but I got rid of the light leaking. Following was done:
>
Seems much to me.
BTW before you start a final render: your aspect ratio for the image is
wrong!
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 11-1-2014 15:41, Ive wrote:
> Am 11.01.2014 13:11, schrieb Thomas de Groot:
>
>> Does not seem much but I got rid of the light leaking. Following was
>> done:
>>
> Seems much to me.
I know ;-) it is a fair lot of work for only little specks of light, but
the result is essential of course.
>
> BTW before you start a final render: your aspect ratio for the image is
> wrong!
hmm... 1000x800 and using all camera parameters as given by the original
scene file. I shall change the right to -x*image_width/image_height and
see if that makes any difference.
Presently, I have difficulties with photons/caustics. THey do not show
on the ceiling...
Ultimately, I shall add a media to the scene to show some dust in the
atmosphere, but not too much.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
