 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
I have been experimenting with media-based clouds using pigment functions.
After lots of playing, I think I've finally begin to understand the balance
between emitting, absorption, and scattering medias. I've also found that
method 3 with 1 interval and a decent amount of samples seems to work the
best (in terms of speed/quality).
Images number 1 and 3 were the slowest, about (estimated) 17h and 24h on a
P4 2.66GHz machine running at 60% duty cycle.
Image 2 is much faster, about 4 hours on a P4 3.0 @ 3.5 running 100% duty
cycle
Originals were 800 x 600, except for 3 which was 1024x768
Still working on making thunderheads, I think I've found a way using the
cylindrical pattern and a few transforms...
Reactor
Post a reply to this message
Attachments:
Download 'cloud_test1.jpg' (32 KB)
Download 'cloud_test2.jpg' (22 KB)
Download 'cloud_test3.jpg' (18 KB)
Preview of image 'cloud_test1.jpg'
Preview of image 'cloud_test2.jpg'
Preview of image 'cloud_test3.jpg'
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Reactor" <rea### [at] hotmail com> wrote in message
news:410716af@news.povray.org...
> I have been experimenting with media-based clouds using pigment functions.
> After lots of playing, I think I've finally begin to understand the
balance
> between emitting, absorption, and scattering medias.
Very impressive! I tried this a couple months ago, made some progress, but
not enough.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Reactor wrote:
> I have been experimenting with media-based clouds using pigment functions.
> After lots of playing, I think I've finally begin to understand the balance
> between emitting, absorption, and scattering medias. I've also found that
> method 3 with 1 interval and a decent amount of samples seems to work the
> best (in terms of speed/quality).
>
> Images number 1 and 3 were the slowest, about (estimated) 17h and 24h on a
> P4 2.66GHz machine running at 60% duty cycle.
> Image 2 is much faster, about 4 hours on a P4 3.0 @ 3.5 running 100% duty
> cycle
> Originals were 800 x 600, except for 3 which was 1024x768
>
> Still working on making thunderheads, I think I've found a way using the
> cylindrical pattern and a few transforms...
>
>
> Reactor
>
WOW! Those look amazing! I especially love the visible light shining
through the coulds in the third picture.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
If you published the code, you might get some offers of marriage... But I
wouldn't take any of them up on it.
Blummin nice work
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Curious what scale you used. I think that's one of the trickier parts
of choosing media settings.
Was 1 POV unit 1 meter? 1 mile? 1 cloud? arbitrary?
Was the sun <rgb 1> <rgb 100> etc.
Do the media settings have any correlation to any turbidity/opacity
units (miles of visibility?)
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ron M" <rm_### [at] cheapcomplexdevicescom> seems to have written in message
news:410### [at] cheapcomplexdevicescom...
>
> Curious what scale you used. I think that's one of the trickier parts
> of choosing media settings.
>
> Was 1 POV unit 1 meter? 1 mile? 1 cloud? arbitrary?
>
> Was the sun <rgb 1> <rgb 100> etc.
>
> Do the media settings have any correlation to any turbidity/opacity
> units (miles of visibility?)
First of all, thanks all for the kind responses. I am posting a source
snipet that should answer a few questions. A macro is in the works, but it
might take a while...
I used 1 pov unit = approximately 1 foot.
I have placed comments in the attached source snipet. It isn't a full
scene, just the media block and explanation. Please feel free to play with
it, and, if someone figures out how to make good thunderheads, please post
your method... I am experimenting with a few things, and I have a
quarter-baked idea for the "Desert" topic (isn't fully formed enough to be
half baked).
Be aware that media clouds have very long render times.
Reactor
Post a reply to this message
Attachments:
Download 'clouds_description.pov.txt' (3 KB)
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Reactor wrote:
>> "Ron M" <rm_### [at] cheapcomplexdevicescom> seems to have written in
>> message news:410### [at] cheapcomplexdevicescom...
>>>
>>> Curious what scale you used. I think that's one of the trickier
>>> parts of choosing media settings.
>>>
>>> Was 1 POV unit 1 meter? 1 mile? 1 cloud? arbitrary?
>>>
>>> Was the sun <rgb 1> <rgb 100> etc.
>>>
>>> Do the media settings have any correlation to any turbidity/opacity
>>> units (miles of visibility?)
>>
>>
>>
>> First of all, thanks all for the kind responses. I am posting a
>> source snipet that should answer a few questions. A macro is in the
>> works, but it might take a while...
>> I used 1 pov unit = approximately 1 foot.
>>
>> I have placed comments in the attached source snipet. It isn't a
>> full scene, just the media block and explanation. Please feel free
>> to play with it, and, if someone figures out how to make good
>> thunderheads, please post your method... I am experimenting with a
>> few things, and I have a quarter-baked idea for the "Desert" topic
>> (isn't fully formed enough to be half baked).
>>
>> Be aware that media clouds have very long render times.
>>
>>
>> Reactor
It looks to me that the shape of the cloud is governed by the shape of the
container. In your case this is a flat box. Looking at the cloud from a large
distance (or when it is scaled to a smaller size), the box shape becomes
visible. IMHO, if you adjust the size of the container then you can make any
shape of cloud. So, making a hammerhead shaped container and filling it with
"cloudy media" shoud get you where you want to be.
--
Maurice
Quote of the minute:
As a goat herd learns his trade by goat, so a writer learns his trade by wrote.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> It looks to me that the shape of the cloud is governed by the shape of the
> container. In your case this is a flat box. Looking at the cloud from a
large
> distance (or when it is scaled to a smaller size), the box shape becomes
> visible. IMHO, if you adjust the size of the container then you can make
any
> shape of cloud. So, making a hammerhead shaped container and filling it
with
> "cloudy media" shoud get you where you want to be.
>
> --
> Maurice
>
> Quote of the minute:
> As a goat herd learns his trade by goat, so a writer learns his trade by
wrote.
>
I tried something somewhat similar, but it requires the "cloudy media" to
have a uniform density, which may look ok from great distances, but wouldn't
work well up close. The media in my inplementation isn't uniformly dense,
which allows the thickening of a cloud toward the center and a few other
different effects (visible shafts of light, realistic "puffy fringe"up close
for flyby animations, etc). I originally tried isosurfaces and blob
objects, both of which I felt looked less realistic.
My current attempt at thunderheads centers around using MegaPov's blob
pattern, which allows me to model the basic shape of the thunderhead using a
blob... seems to work so far, but after playing with media in Pov 3.5, you
quickly noticed difference between 3.1. Still working, produced some nice
images (but very slow), will post a decent one in the next few days.
I would also like to point out that if you are looking at the cloud from a
large distance/unique angle/wide camera angle, you should adjust the max and
min distances appropriately. You can also achieve faster render times if
you don't have a needlesly large cloud object (beware of shadows that should
be present from portions of the clouds that are off screen, if you make the
cloud box too small your scene might not look right).
Reactor
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Reactor wrote:
> I have been experimenting with media-based clouds using pigment functions.
> After lots of playing, I think I've finally begin to understand the balance
> between emitting, absorption, and scattering medias. I've also found that
> method 3 with 1 interval and a decent amount of samples seems to work the
> best (in terms of speed/quality).
Awesome... especially the first image! But the rendering durations are
also awesome - better not to think about how long such a scene would
take on my pre-war rattletrap (AMD K6-2/400 MHz)!
See you in Khyberspace!
Yadgar
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
