> "And" <49341109@ntnu.edu.tw> wrote:

>> Maybe I can express my opinion.
>> Once I rendered a scene with large area surfaces that switched the
>> specular/phong on. The multiple light sources can make it slow truely.
>> I only have the experience that time. Because I almost use the reflection
>> instead of specular.
>
> This only gives good results of you use photons, and I've never had good results
> with photons.
>
> Regards,
> A.D.B.
>
>
>

With photons, you need to set the first active object lighted with a 
target directive.
You follow that with reflection on and/or refraction on.
If you have several stacked active objects, only the first in line 
should be set as a target. If you set an object as a target and that 
object is beheind another, all photons shoot in it's direction will be 
intersepted by the first object. If you use count, it will steal photons 
from the other objects...

If you have one problem object, you can use target float. That value 
will multiplied with the spacing. Usefull for an object needing a high 
density of photons while most others only need a moderate coverage.

I find it beter to use spacing rather than count. That's even more 
importent when you have several light_source and target objects.
You need to adjust spacing according to the dimention of your objects 
relative to the scale of the scene. If, on average, 1 unit covers only a 
few pixels, use a large spacing, but if 1 unit could visualy cover a 
large area, use a small spacing.

Normaly, I don't set any other parameter other than spacing and autostop 
0 and, maybe, jitter.

Remember that the camera NEVER directly see the photons. Photons can 
cause caustics, but never ever cause highlights.

If you use reflection to get your highlights, the light_source must be 
visible by using looks_like or been unioned with some object.


Alain

Post a reply to this message

"Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> So I've been working on this for several months now (off and on) and it's
> nowhere near finished.
>
> The problem I'm having is that even with the crummy radiosity settings I used on
> this render, it took several days.  The reason for this, I think, is the
> combination of the lighting and the water (I know the water needs work).
>
> I've been experimenting with pure radiosity for lighting, but I'm having trouble
> getting good illumination from flames. I've tried pumping the emission (media)
> as high as 1200, but as soon as I try to add color to the flames the light all
> but vanishes.
>
> I hadn't really wanted to show this off yet, but I need help with the lighting.
>
> This render uses conventional lighting but, as I said, it was slow.
>
> Regards,
> A.D.B.

The hole debate here is only about theory. Can you post the source code? Most
likely I'm not the one to spot the difficulties but others may be able to do so.

Best regards,
Michael

Post a reply to this message

"MichaelJF" <mi-### [at] t-onlinede> wrote:
> "Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> > So I've been working on this for several months now (off and on) and it's
> > nowhere near finished.
> >
> > The problem I'm having is that even with the crummy radiosity settings I used on
> > this render, it took several days.  The reason for this, I think, is the
> > combination of the lighting and the water (I know the water needs work).
> >
> > I've been experimenting with pure radiosity for lighting, but I'm having trouble
> > getting good illumination from flames. I've tried pumping the emission (media)
> > as high as 1200, but as soon as I try to add color to the flames the light all
> > but vanishes.
> >
> > I hadn't really wanted to show this off yet, but I need help with the lighting.
> >
> > This render uses conventional lighting but, as I said, it was slow.
> >
> > Regards,
> > A.D.B.
>
> The hole debate here is only about theory. Can you post the source code? Most
> likely I'm not the one to spot the difficulties but others may be able to do so.
>
> Best regards,
> Michael

First of all, the original question was about theory.  Secondly, I'm not asking
for help fixing code: I'm using a water material (slightly modified) that was
provided in a previous thread, if that doesn't work right nothing will.

The problem is most likely due to six overlapping light boxes, plus water which
has been said to be slow under ideal conditions.  The only viable option I can
think of is to get rid of the point lights entirely and switch to some sort of
radiant lighting, which is what the question was actually about.

so, to reiterate, does anybody have a good way to make luminous flames without
point lights?

Regards,
A.D.B.

Post a reply to this message

> "MichaelJF" <mi-### [at] t-onlinede> wrote:
>> "Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
>>> So I've been working on this for several months now (off and on) and it's
>>> nowhere near finished.
>>>
>>> The problem I'm having is that even with the crummy radiosity settings I used on
>>> this render, it took several days.  The reason for this, I think, is the
>>> combination of the lighting and the water (I know the water needs work).
>>>
>>> I've been experimenting with pure radiosity for lighting, but I'm having trouble
>>> getting good illumination from flames. I've tried pumping the emission (media)
>>> as high as 1200, but as soon as I try to add color to the flames the light all
>>> but vanishes.
>>>
>>> I hadn't really wanted to show this off yet, but I need help with the lighting.
>>>
>>> This render uses conventional lighting but, as I said, it was slow.
>>>
>>> Regards,
>>> A.D.B.
>>
>> The hole debate here is only about theory. Can you post the source code? Most
>> likely I'm not the one to spot the difficulties but others may be able to do so.
>>
>> Best regards,
>> Michael
>
> First of all, the original question was about theory.  Secondly, I'm not asking
> for help fixing code: I'm using a water material (slightly modified) that was
> provided in a previous thread, if that doesn't work right nothing will.
>
> The problem is most likely due to six overlapping light boxes, plus water which
> has been said to be slow under ideal conditions.  The only viable option I can
> think of is to get rid of the point lights entirely and switch to some sort of
> radiant lighting, which is what the question was actually about.
>
> so, to reiterate, does anybody have a good way to make luminous flames without
> point lights?
>
> Regards,
> A.D.B.
>

Object with a high emission component in the finish.
Object filled with some emissive media. Need to add hollow to that object.

Using media, you need to use some pattern to modulate the emission. Some 
that can be usefull:
spherical (usualy with some turbulence)
bumps (normaly with uneven scalling, also with some turbulence)
Other patterns can be used creatively.

The media container will normaly be fully transparent, but that's not 
obligatory, it just need to be mostly transparent.



Alain

Post a reply to this message

Alain <kua### [at] videotronca> wrote:

> > "MichaelJF" <mi-### [at] t-onlinede> wrote:
> >> "Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> >>> So I've been working on this for several months now (off and on) and it's
> >>> nowhere near finished.
> >>>
> >>> The problem I'm having is that even with the crummy radiosity settings I used on
> >>> this render, it took several days.  The reason for this, I think, is the
> >>> combination of the lighting and the water (I know the water needs work).
> >>>
> >>> I've been experimenting with pure radiosity for lighting, but I'm having trouble
> >>> getting good illumination from flames. I've tried pumping the emission (media)
> >>> as high as 1200, but as soon as I try to add color to the flames the light all
> >>> but vanishes.
> >>>
> >>> I hadn't really wanted to show this off yet, but I need help with the lighting.
> >>>
> >>> This render uses conventional lighting but, as I said, it was slow.
> >>>
> >>> Regards,
> >>> A.D.B.
> >>
> >> The hole debate here is only about theory. Can you post the source code? Most
> >> likely I'm not the one to spot the difficulties but others may be able to do so.
> >>
> >> Best regards,
> >> Michael
> >
> > First of all, the original question was about theory.  Secondly, I'm not asking
> > for help fixing code: I'm using a water material (slightly modified) that was
> > provided in a previous thread, if that doesn't work right nothing will.
> >
> > The problem is most likely due to six overlapping light boxes, plus water which
> > has been said to be slow under ideal conditions.  The only viable option I can
> > think of is to get rid of the point lights entirely and switch to some sort of
> > radiant lighting, which is what the question was actually about.
> >
> > so, to reiterate, does anybody have a good way to make luminous flames without
> > point lights?
> >
> > Regards,
> > A.D.B.
> >
>
> Object with a high emission component in the finish.
> Object filled with some emissive media. Need to add hollow to that object.
>
> Using media, you need to use some pattern to modulate the emission. Some
> that can be usefull:
> spherical (usualy with some turbulence)
> bumps (normaly with uneven scalling, also with some turbulence)
> Other patterns can be used creatively.
>
> The media container will normaly be fully transparent, but that's not
> obligatory, it just need to be mostly transparent.
>
>
>
> Alain

I have tried all of these things.  The media gives good results as long as I
don't care if it's pure white... The problem of course is that the parts that
affect the radiosity will be well beyond the white-point and any attempt to add
color reduces their brightness to such a degree that they no longer have the
desired effect.

Regards,
A.D.B.

Post a reply to this message

"Anthony D. Baye" <Sha### [at] spamnomorehotmailcom> wrote:
> "And" <49341109@ntnu.edu.tw> wrote:
> >
> > Maybe I can express my opinion.
> > Once I rendered a scene with large area surfaces that switched the
> > specular/phong on. The multiple light sources can make it slow truely.
> > I only have the experience that time. Because I almost use the reflection
> > instead of specular.
>
> This only gives good results of you use photons, and I've never had good results
> with photons.
>
> Regards,
> A.D.B.


This is not true, I seldom use photons. And in the scene you show it is no need
of using photons. The effect on the water is only the highlight.

I don't have any computer now.
Maybe I can teach you making the highlight with reflection this weekend.

Post a reply to this message

>> The hole debate here is only about theory. Can you post the source code? Most
>> likely I'm not the one to spot the difficulties but others may be able to do so.
>>
>> Best regards,
>> Michael
>
> First of all, the original question was about theory.  Secondly, I'm not asking
> for help fixing code: I'm using a water material (slightly modified) that was
> provided in a previous thread, if that doesn't work right nothing will.
>
> The problem is most likely due to six overlapping light boxes, plus water which
> has been said to be slow under ideal conditions.  The only viable option I can
> think of is to get rid of the point lights entirely and switch to some sort of
> radiant lighting, which is what the question was actually about.
>
> so, to reiterate, does anybody have a good way to make luminous flames without
> point lights?
>
> Regards,
> A.D.B.
>
I'm still not quiet sure that you can solve this riddle only by theory. 
I assume there are additional conditions at work here to slow down the 
rendering of your water that dramatically.

Since I have a somewhat simliar scene in my mind I played around a bit. 
I stole a fire from the IRTC (emissive media), modeled a small scene 
which is somewhat similiar to yours (at least the position of the 
"fires" and the water with respect to the camera should be comparable). 
I added a dimmed yellow light_source to the media to get the water 
transparent. The "fires" are repeated 6 times. The water material is 
yours from your "looking for water" thread but I incorporated clipkas 
comments on the demo scene water - and scaled the normal down. The 
rendering was slowed down by command line options (+bs8 +wt6 +r4). 
Radiosity settings were Radiosity_IndoorHQ from rad_def.inc. Parsing and 
rendering was done in 14m 52s. Without the water: 2m 5s. Machine: Core 
i7@3.4 GHz.

So I must have left out a constraint you have in your scene.

My own scene will be different from yours, but the lighting situation 
and the position of the pool relative to the fires are similiar. The 
media of the fires must be improved of course.

Best regards,
Michael

Post a reply to this message

Attachments:
Download 'waterradtest02.png' (447 KB)

Preview of image 'waterradtest02.png'

> Alain <kua### [at] videotronca> wrote:

>> Object with a high emission component in the finish.
>> Object filled with some emissive media. Need to add hollow to that object.
>>
>> Using media, you need to use some pattern to modulate the emission. Some
>> that can be usefull:
>> spherical (usualy with some turbulence)
>> bumps (normaly with uneven scalling, also with some turbulence)
>> Other patterns can be used creatively.
>>
>> The media container will normaly be fully transparent, but that's not
>> obligatory, it just need to be mostly transparent.
>>
>>
>>
>> Alain
>
> I have tried all of these things.  The media gives good results as long as I
> don't care if it's pure white... The problem of course is that the parts that
> affect the radiosity will be well beyond the white-point and any attempt to add
> color reduces their brightness to such a degree that they no longer have the
> desired effect.
>
> Regards,
> A.D.B.
>
>

In your media, you are not limited to the 0..1 range.
The emission value can be as high as you want or need. emission<100, 95, 
40> is perfectly fine and gives an orangish tint.

Also, your density can use a colour_map with values that are also not 
limited to the 0..1 range. Negative values are supported as well as 
values much larger than 1.

Depending on the actual dimentions of your containing object, the needed 
values will need to adjusted.

You can add some absorbing media to your flames. That media don't need 
to have the same pattern as the emissive media. In fact, in most 
relatively small flames, the emissive part barely absorb light at all, 
and the absorbing parts are often the cooler parts that don't emit much.

A very bright emissive media can look white in it's brightest parts, but 
have reflection that are definitively not white, and cause coloured 
radiosity illumination.
If you use media with radiosity, you need to add "media on" in the 
radiosity block if you want your media do be able to actualy illuminate 
it's surrounding.



Alain

Post a reply to this message

Am 27.08.2015 um 23:08 schrieb MichaelJF:
>>> The hole debate here is only about theory. Can you post the source
>>> code? Most
>>> likely I'm not the one to spot the difficulties but others may be
>>> able to do so.
>>>
>>> Best regards,
>>> Michael
>>
>> First of all, the original question was about theory.  Secondly, I'm
>> not asking
>> for help fixing code: I'm using a water material (slightly modified)
>> that was
>> provided in a previous thread, if that doesn't work right nothing will.
>>
>> The problem is most likely due to six overlapping light boxes, plus
>> water which
>> has been said to be slow under ideal conditions.  The only viable
>> option I can
>> think of is to get rid of the point lights entirely and switch to some
>> sort of
>> radiant lighting, which is what the question was actually about.
>>
>> so, to reiterate, does anybody have a good way to make luminous flames
>> without
>> point lights?
>>
>> Regards,
>> A.D.B.
>>
> I'm still not quiet sure that you can solve this riddle only by theory.
> I assume there are additional conditions at work here to slow down the
> rendering of your water that dramatically.
>
> Since I have a somewhat simliar scene in my mind I played around a bit.
> I stole a fire from the IRTC (emissive media), modeled a small scene
> which is somewhat similiar to yours (at least the position of the
> "fires" and the water with respect to the camera should be comparable).
> I added a dimmed yellow light_source to the media to get the water
> transparent. The "fires" are repeated 6 times. The water material is
> yours from your "looking for water" thread but I incorporated clipkas
> comments on the demo scene water - and scaled the normal down. The
> rendering was slowed down by command line options (+bs8 +wt6 +r4).
> Radiosity settings were Radiosity_IndoorHQ from rad_def.inc. Parsing and
> rendering was done in 14m 52s. Without the water: 2m 5s. Machine: Core
> i7@3.4 GHz.
>
> So I must have left out a constraint you have in your scene.
>
> My own scene will be different from yours, but the lighting situation
> and the position of the pool relative to the fires are similiar. The
> media of the fires must be improved of course.
>
> Best regards,
> Michael
>
>
I have forgotten to mention that I turned on media and normal within the 
radiosity block (@Alain: thanks for this reminder) and I used aa 0.3 
from the Quickres.ini.

Best regards,
Michael

Post a reply to this message

> I'm still not quiet sure that you can solve this riddle only by theory.
> I assume there are additional conditions at work here to slow down the
> rendering of your water that dramatically.
>
> Since I have a somewhat simliar scene in my mind I played around a bit.
> I stole a fire from the IRTC (emissive media), modeled a small scene
> which is somewhat similiar to yours (at least the position of the
> "fires" and the water with respect to the camera should be comparable).
> I added a dimmed yellow light_source to the media to get the water
> transparent. The "fires" are repeated 6 times. The water material is
> yours from your "looking for water" thread but I incorporated clipkas
> comments on the demo scene water - and scaled the normal down. The
> rendering was slowed down by command line options (+bs8 +wt6 +r4).
> Radiosity settings were Radiosity_IndoorHQ from rad_def.inc. Parsing and
> rendering was done in 14m 52s. Without the water: 2m 5s. Machine: Core
> i7@3.4 GHz.
>
> So I must have left out a constraint you have in your scene.
>
> My own scene will be different from yours, but the lighting situation
> and the position of the pool relative to the fires are similiar. The
> media of the fires must be improved of course.
>
> Best regards,
> Michael


using those radiosity settings, my 3.01 Ghz Phenom II X6 machine sits on its
thumbs and shows a black window.

Even when I turn off the light source fading, the focal blur and the
anti-aliasing.  With those radiosity settings my computer just sits there.

I have no idea what's going on with this scene, but even when I remove the point
lights and replace them with emissive objects (simple spheres no ambient or
diffuse just emission 60 and no_image) the IndoorHQ settings from rad_def.inc
cause my scene to bog down at 0%.  It doesn't even get to the water.

here is the actual water material I'm using:

#declare M_Watx4 =
material {
    texture {
            pigment { color rgbft <0.235, 0.318, 0.618, 1.0, 0.96> }
            finish {
                diffuse 0
                ambient 0
//                reflection {0.1, 0.9 fresnel on exponent 0.8}
                reflection { 1.0 fresnel on }
                conserve_energy
                specular albedo 1
                roughness 0.007
                }
            normal {
                function {f_ridged_mf(x,y,z, 0.1, 3.0, 7.0, 0.7, 0.7, 2)} 0.3
                scale <1.0, 0.09375, 1.0>
//                scale <4.68, 3.2, 4.68>/12
                }
        }
    interior {
        ior 1.31
        fade_distance 10
        fade_power 1001
        fade_color <0.235, 0.318, 0.618>
        }
    }

my command line: +AM2 +a0.03 +H945 +W1680 +BM2 -RVP +WT6 +P serraglio.pov

I get some slowdown from the floor because the tiles are all geometry (prisms),
and I expect that effects the water some, but I don't think it explains
everything.

my lights are point lights with fade_distance 7 and fade_power 2

other than that, my radiosity settings are dumbed down from the tutorial for
conventional lighting in the docs.

radiosity {
    pretrace_start 0.04
    pretrace_end   0.005
    count 35

    nearest_count 10
    error_bound 0.5
    recursion_limit 2

    low_error_factor 0.1
    gray_threshold 0.2
    minimum_reuse 0.015
    maximum_reuse 0.2
    brightness 1.0
    normal off

    adc_bailout 1/256
   }

Regards,
A.D.B.

Post a reply to this message