POV-Ray : Newsgroups : povray.general : df3 clouds Server Time
28 Mar 2024 13:10:03 EDT (-0400)
  df3 clouds (Message 1 to 5 of 5)  
From: Bald Eagle
Subject: df3 clouds
Date: 7 Jun 2017 10:20:00
Message: <web.59380b19fb407d23c437ac910@news.povray.org>
Pretty interesting renderings of storm data using a modified version of POV-Ray.

https://nnc3.com/mags/LJ_1994-2014/LJ/127/7486.html

https://nnc3.com/mags/LJ_1994-2014/LJ/127/7754.html


Now those are some clouds!


Post a reply to this message

From: Thomas de Groot
Subject: Re: df3 clouds
Date: 8 Jun 2017 02:46:20
Message: <5938f2bc$1@news.povray.org>
On 7-6-2017 16:18, Bald Eagle wrote:
> Pretty interesting renderings of storm data using a modified version of POV-Ray.
>
> https://nnc3.com/mags/LJ_1994-2014/LJ/127/7486.html
>
> https://nnc3.com/mags/LJ_1994-2014/LJ/127/7754.html
>
>
> Now those are some clouds!
>
>

Interesting.

-- 
Thomas


Post a reply to this message

From: Kenneth
Subject: Re: df3 clouds
Date: 8 Jun 2017 13:55:06
Message: <web.59398f2d644a5b5d883fb31c0@news.povray.org>
Thomas de Groot <tho### [at] degrootorg> wrote:

>
> Interesting.
>

Indeed. One of those sites gives a very good description of how POV-Ray uses df3
files with isosurfaces. I would assume that some of the source-code patches
mentioned/created by the author (or the other fellow named Suzuki) have not been
ported to official POV-Ray(?)  Being able to model thunderstorms would be quite
interesting!

From the article...
"My thunderstorm numerical model data cannot be represented as a mathematical
function [the method POV-Ray uses to create isosurfaces]; instead, it is
represented as three-dimensional floating-point arrays..."

"Suzuki's code... contains routines that expand the functionality of POV-Ray
density files, including the option of rendering floating-point density file
data."


Post a reply to this message

From: Bald Eagle
Subject: Re: df3 clouds
Date: 8 Jun 2017 15:20:01
Message: <web.5939a276644a5b5dc437ac910@news.povray.org>
"Kenneth" <kdw### [at] gmailcom> wrote:
> I would assume that some of the source-code patches
> mentioned/created by the author (or the other fellow named Suzuki) have not been
> ported to official POV-Ray(?)  Being able to model thunderstorms would be quite
> interesting!

I think from my brief inspection of the source code modifications, that the
changes were quite minor, although having them available as a separate special
type of .df3 - "df32" would be a great addition to the current default source
code.

I believe the current isosurface code was written by Mr. Suzuki, so we have very
little to modify in order to make such a thing possible.

I know from trying to "scan-in" rgb data from an image file via eval_pigment(),
that it's very SLOW.  I don't know how trying to process a df3 in the same way
would be.


Definitely, the interesting thing about the code is how the hdf data is imported
as a pattern function, which is then used to generate the isosurface.  That must
have taken a LONG time to process, way back then, with the "slow" processors,
limited memory, and the huge number of data points.  And it was also probably
before the isosurface approximation macros were written.

I'm guessing that perhaps for simple clouds or low-resolution storms, that
perhaps the hdf file might be able to be converted to a df3, but I haven't
looked closely at the data format (though he does suggest that it's possible)

He's obviously been quite busy and successful since then:   http://orf.media/


Post a reply to this message

From: clipka
Subject: Re: df3 clouds
Date: 9 Jun 2017 11:58:06
Message: <593ac58e$1@news.povray.org>
Am 07.06.2017 um 16:18 schrieb Bald Eagle:
> Pretty interesting renderings of storm data using a modified version of POV-Ray.
> 
> https://nnc3.com/mags/LJ_1994-2014/LJ/127/7486.html
> 
> https://nnc3.com/mags/LJ_1994-2014/LJ/127/7754.html
> 
> 
> Now those are some clouds!

Anyone volunteering to port Suzuki's floating-point density files patch
to POV-Ray 3.7 would be welcome.

I currently don't have the time to implement it myself, but would do my
best to find the time to merge such a patch into official POV-Ray.


Caveat: I haven't looked at the file format yet, but if it's binary like
DF3, then care must be taken to make the code portable, so that it works
even on machines that don't use the standard IEEE single-precision
floating-point format, and/or use a different byte ordering for such
data. (IIRC there already exists an official mechanisms to convert
to/from IEEE format in the POVMS module, which could be co-opted for the
floating-point density files.)


Post a reply to this message

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