"St." <dot### [at] dotcom> wrote:
> >     I particularily like this one.
>
>
>      Sorry, I meant 'sky_7.png'. 'Sky_9.png' looks good too, but I see this
> more as a sunset or sunrise, so, a lot more red in there maybe? Casting red
> over the land too?
>
>
>        ~Steve~

Thanks!

The sphere and the ground are plain white and radiosity is on. If you use a
color picker, you can notice they take a tint according to the colors of the
clouds and the that of the sky. In daylight, they get a yellowish tint where
exposed to the sun, and blueish in shadows, which is the natural behaviour.

Render times for a 640*480 with an Athlon 3000 is 1 to 24 hours, depending on
the number of layers, the sun's elevation, the number of intervals/samples I
need and adaptive method's behaviour. Lines near the horizon take the longest
of course. Normally, in RL there is always some haze or fog hiding this part.
But I want to keep it clear for those tests. However haze or fog can be
obtained with a layer and an appropriate density function.

Here is another try that took about 1:40 to render. I highlight that 70% of
these scene are occupied by the skyscape. In normal outdoor scenes, it is less
because you have something else to show. But I want for my own pleasure to make
scene where the sky is the subject (TC-RTC topic ;)?).

    Bruno

Post a reply to this message

Attachments:
Download 'sky_16.png' (254 KB)

Preview of image 'sky_16.png'

"Thomas de Groot" <t.d### [at] internlDOTnet> wrote:
> I was thinking about these:
>
http://news.povray.org/povray.binaries.images/thread/%3Cweb.44f8fbd1d1fee11a506816740@news.povray.org%3E/
>
http://news.povray.org/povray.binaries.images/thread/%3Cweb.4446b51a716c65e7b3cf832f0@news.povray.org%3E/
>
http://news.povray.org/povray.binaries.scene-files/thread/%3Cweb.4447da16ade663f78069ddd30%40news.povray.org%3E/?ttop
=250513&toff=100
>
> Thomas

Oohh. I missed this discussion! The results are VERY good!

Having a look at the source code, I confirm that I have a similar approach, but
maybe a little more advanced and closer to nature's laws. Some colors were
obtained by applying a non very natural values to the scattering media of the
atmosphere, like "scattering{1 rgb .01*<1,.8, .6> extinction 0}".

In RL, atmosphere's colors (blue at daytime, yellow/orange/red-ish at sunset)
are due to Rayleigh scattering (type 4), plus eventually some other media
(scattering/absorption of specific colors) for non-pure air particles or
pollution. In addition, as explained in the documentation, extinction is always
1. Clouds are type 2 (or 3 in some cases) scattering media. They usually have no
coloring behaviour themselves and are only white-scattering. Clouds are tinted
by the scattered light coming from the sun and neighbourhood. That's what I try
to reproduce.

In order to make sure of the realisticness of some of my skyies, I try the same
scene with different elevations of the sun, only changing the sun's power to
compensate illumination decrease. I'll render one with only the atmosphere at
different elevations to illustrate.

    Bruno

Post a reply to this message

message de news: web.4732e7267579baf9e8ba46670@news.povray.org...
>
> In order to make sure of the realisticness of some of my skyies, I try the 
> same
> scene with different elevations of the sun, only changing the sun's power 
> to
> compensate illumination decrease. I'll render one with only the atmosphere 
> at
> different elevations to illustrate.
>
>    Bruno


I am very interested by your approach.
I had a try at it but my lack of scientific background drew me back.

Maybe you could write a page about the subject?

Marc

Post a reply to this message

"Bruno Cabasson" <bru### [at] alcatelaleniaspacefr> wrote in message 
news:web.4732de627579baf9e8ba46670@news.povray.org...

> Here is another try that took about 1:40 to render. I highlight that 70% 
> of
> these scene are occupied by the skyscape. In normal outdoor scenes, it is 
> less
> because you have something else to show. But I want for my own pleasure to 
> make
> scene where the sky is the subject (TC-RTC topic ;)?).

     Heh, well, the present topic of 'Aeroplanes' could be a good candidate 
for your skies! I like sky_16 too - I swear that I saw something *very* 
similar only two evenings ago. I think you're cracking this.  :)

       ~Steve~



>
>    Bruno
>

Post a reply to this message

"Bruno Cabasson" <bru### [at] alcatelaleniaspacefr> schreef in bericht 
news:web.4732e7267579baf9e8ba46670@news.povray.org...
>
> Oohh. I missed this discussion! The results are VERY good!
>
> Having a look at the source code, I confirm that I have a similar 
> approach, but
> maybe a little more advanced and closer to nature's laws. Some colors were
> obtained by applying a non very natural values to the scattering media of 
> the
> atmosphere, like "scattering{1 rgb .01*<1,.8, .6> extinction 0}".
>
> In RL, atmosphere's colors (blue at daytime, yellow/orange/red-ish at 
> sunset)
> are due to Rayleigh scattering (type 4), plus eventually some other media
> (scattering/absorption of specific colors) for non-pure air particles or
> pollution. In addition, as explained in the documentation, extinction is 
> always
> 1. Clouds are type 2 (or 3 in some cases) scattering media. They usually 
> have no
> coloring behaviour themselves and are only white-scattering. Clouds are 
> tinted
> by the scattered light coming from the sun and neighbourhood. That's what 
> I try
> to reproduce.
>
> In order to make sure of the realisticness of some of my skyies, I try the 
> same
> scene with different elevations of the sun, only changing the sun's power 
> to
> compensate illumination decrease. I'll render one with only the atmosphere 
> at
> different elevations to illustrate.
>

I am really impressed. I wait impatiently for further developments along 
these lines. I sure am interested!

Thomas

Post a reply to this message

"Thomas de Groot" <t.d### [at] internlDOTnet> wrote:
> "Bruno Cabasson" <bru### [at] alcatelaleniaspacefr> schreef in bericht
 I'll render one with only the atmosphere
> > at
> > different elevations to illustrate.
> >
>
> I am really impressed. I wait impatiently for further developments along
> these lines. I sure am interested!
>
> Thomas

Thanks for encouragements.

Here is an illustration for atmosphere, considering pure atmosphere. The only
changing parameter is the elevation of the sun (simple looks_like white sphere
at the moment).

You can obtain different colors (including reddening/decay) by tweaking some
parameters like the color of the scattered blue (rgb <0.25, 0.65, 1> here), the
power of the sun, and parameters controlling the density function (basically:
exp(-k*h/T), k=mult.factor, h=height at the given point, T=constant thickness
of atmo). You can also add absorption, but I prefer leave that to cloud layers.

The camera is at 2m above the ground. When you put it higher, you get a deeper
blue automatically because there is less gaz above you: the blue is less
scattered and the blackness of deep space contibutes more.

    Bruno.

Post a reply to this message

Attachments:
Download 'atmo_demo.jpg' (39 KB)

Preview of image 'atmo_demo.jpg'

"Bruno Cabasson" <bru### [at] alcatelaleniaspacefr> schreef in bericht 
news:web.473446d07579baf9e8ba46670@news.povray.org...
> Here is an illustration for atmosphere, considering pure atmosphere. The 
> only
> changing parameter is the elevation of the sun (simple looks_like white 
> sphere
> at the moment).
>
> You can obtain different colors (including reddening/decay) by tweaking 
> some
> parameters like the color of the scattered blue (rgb <0.25, 0.65, 1> 
> here), the
> power of the sun, and parameters controlling the density function 
> (basically:
> exp(-k*h/T), k=mult.factor, h=height at the given point, T=constant 
> thickness
> of atmo). You can also add absorption, but I prefer leave that to cloud 
> layers.
>
> The camera is at 2m above the ground. When you put it higher, you get a 
> deeper
> blue automatically because there is less gaz above you: the blue is less
> scattered and the blackness of deep space contibutes more.
>

Excellent! That looks very intuitive to use. For skies I usually like to use 
LightSysIV, but there the problem is that I do not really understand the 
background and the correct way to change things. Your method seems really 
simple for the end-user. I suppose that, given the correct parameters, you 
could simulate a Martian sundown for instance as easily as a terrestrial 
one.

Thomas

Post a reply to this message

"Thomas de Groot" <t.d### [at] internlDOTnet> wrote:

> Excellent! That looks very intuitive to use. For skies I usually like to use
> LightSysIV, but there the problem is that I do not really understand the
> background and the correct way to change things. Your method seems really
> simple for the end-user. I suppose that, given the correct parameters, you
> could simulate a Martian sundown for instance as easily as a terrestrial
> one.
>
> Thomas

I intend to make it only declarative. You specify values for parameters, scalar
or functions/color_maps or so, and they are processed by an 'engine'. In this
case, the engine is very simple.

Sky_17 is currently rendering and here is the state now. It is a 3 cloud layers
sky (plus the atmosphere), and it already took about 12-14 hours to render
(Athlon 3000). I think it'll average 6-8 pps. I needed 2 hours to tweak this
one before the final render. Seeing it now, I find it a bit too bright, but
presentable however.

    Bruno

PS: The fact is ... I AM a martian!

Post a reply to this message

Attachments:
Download 'sky_17_rendering.jpg' (58 KB)

Preview of image 'sky_17_rendering.jpg'

"Bruno Cabasson" <bru### [at] alcatelaleniaspacefr> schreef in bericht 
news:web.473485927579baf9e8ba46670@news.povray.org...

> I intend to make it only declarative. You specify values for parameters, 
> scalar
> or functions/color_maps or so, and they are processed by an 'engine'. In 
> this
> case, the engine is very simple.
>
> Sky_17 is currently rendering and here is the state now. It is a 3 cloud 
> layers
> sky (plus the atmosphere), and it already took about 12-14 hours to render
> (Athlon 3000). I think it'll average 6-8 pps. I needed 2 hours to tweak 
> this
> one before the final render. Seeing it now, I find it a bit too bright, 
> but
> presentable however.
>

Yes, that is the price to pay. Abe Madey's skies take about the same time 
IIRC. However, the final result is worth the wait.

> PS: The fact is ... I AM a martian!

I am not surprised....   :-)

Thomas

Post a reply to this message

Bruno Cabasson wrote:
> "Thomas de Groot" <t.d### [at] internlDOTnet> wrote:
> 
>> Excellent! That looks very intuitive to use. For skies I usually like to use
>> LightSysIV, but there the problem is that I do not really understand the
>> background and the correct way to change things. Your method seems really
>> simple for the end-user. I suppose that, given the correct parameters, you
>> could simulate a Martian sundown for instance as easily as a terrestrial
>> one.
>>
>> Thomas
> 
> I intend to make it only declarative. You specify values for parameters, scalar
> or functions/color_maps or so, and they are processed by an 'engine'. In this
> case, the engine is very simple.
> 
> Sky_17 is currently rendering and here is the state now. It is a 3 cloud layers
> sky (plus the atmosphere), and it already took about 12-14 hours to render
> (Athlon 3000). I think it'll average 6-8 pps. I needed 2 hours to tweak this
> one before the final render. Seeing it now, I find it a bit too bright, but
> presentable however.
> 
>     Bruno
> 
> PS: The fact is ... I AM a martian!
> 
> 
> ------------------------------------------------------------------------
> 

Wow!

-- Arttu Voutilainen

Post a reply to this message