EMAIL:                  simon_smith@zen.co.uk
NAME:                   Simon Smith
TOPIC:                  Mist and Fog
COPYRIGHT: I SUBMIT TO THE STANDARD RAYTRACING COMPETITION COPYRIGHT.
TITLE:                  Planetary Nebula
COUNTRY:                UK
WEBPAGE:                http://www.simon-smith.org/raytracing/index.html
RENDERER USED:          PoVRay 3.00a.riscos.cc5.06 (unofficial version)
TOOLS USED:             Trial and error, a text editor and 2D drawing tool
RENDER TIME:            1360x1024 took about 28 hours
HARDWARE USED:          600MHz Iyonix (30321 ARM processor) No FPU :-(

IMAGE DESCRIPTION:      This is a mock-up astronomical image intended to be
                        compared with views of the Cat's Eye Nebula, the
Hourglass Nebula, the Ant Nebula and others. I have provided two views - a
greyscale version and a 'false colour' version. The 'false colours' used
were based on those in this picture of the Cat's Eye Nebula;
http://www.astro.washington.edu/balick/WFPC2/catseye.jpeg, although I did
decide to use them in different order in my scene. I actually think I like
the greyscale version slightly better. :-)


DESCRIPTION OF HOW THIS IMAGE WAS CREATED:

I anticipated that most of the submissions for this round's topic would be
earthbound scenes - early morning mists and the like. And given how
expensive fog is to render, I never seriously intended to enter this month's
competition, simply because my home PC lacks the computing power. But with
18 days to go to the deadline, inspiration struck; a planetary nebula would
be a legitimate subject - it just happens to be an atmospheric effect for a
stellar atmosphere rather than a planetary atmosphere. And I knew that these
nebulae tend to throw off their atmospheres in concentric 'shells' which
could be modelled reasonably well using standard PoVray textures.

Of course, 18 days, a slow computer, and other commitments made it
impossible to do as much work as I would have liked to have done. I lost a
couple of Saturdays and a couple of weekday evenings which I really could
have used. A complaint with which every other raytracing hobbyist will be
thoroughly familiar. :-)

In fact, most of my early test renders just used some overlapping ellipses
and they rendered quickly and looked quite reasonable. It was always the
textures that were going to be the more important part of this scene. It was
only quite late on that I switched to a more complex, organic nebula shape
based on blobs. By that time, the layered textures were slowing things down
beautifully, and using complex blobs instead of simple ellipses made little
difference.

The colours in the nebula are based on the Cat's Eye nebula picture
mentioned above. Its shape was mostly my own creation, but I was
certainly inspired by NGC7009 (see the Wikipedia link below), the Hourglass
Nebula, and M2-9, the Wings of a Butterfly Nebula (APOD 2005 June 02
http://antwrp.gsfc.nasa.gov/apod/image/0506/m2-9b_hst_big.jpg) (hm, looks
more like a squid to me) and others besides.

When producing this scene I did about 150 test renders. Of those, about a
dozen were focused on getting the shapes right - all the rest were tests and
refinements of the textures.

This is one of the most 'faked' scenes I have ever produced. I would have
liked to have made the star ultra-bright so that it really did burn through
the nebula shell in the direction of the camera. The current setup uses a
wood texture with a transparent core in which the star sits. This produces
much the same effect, but would require the wood texture to be rotated if
the camera was moved.

The blobby nebula shape started out perfectly symmetrical - there is a
sample render in the zip file. I then randomly added or subtracted a small
amount to or from each parameter, and this distorted it by about the right
amount. I also made sure that it was scaled and rotated by varying amounts
each time it was used in the scene, and the textures were given similar
treatment. One does not need big displacements when doing this sort of thing
- a 2-5% linear distortion or 1 or 2 degree rotation in each direction seems
plenty.

A similar method was used for the nebula ring. In early test renders it was
just a torus. I have no idea whether it would be possible for a real
planetary nebula to produce a ring structure like this.

When testing the various wood textures, I draw a couple of ellipses in a 2D
graphics package and interpolated them with ten colour bands. I could then
overlay the nebula shape and see roughly where the colour transitions would
occur. (My colour maps were set using using 0.1 point 'steps'.) This was
very helpful and saved a lot of tedious test renders. But it does make it
doubly annoying that I still managed to lose the central star at the last
minute. Oh well. There's a test render in the zip file showing how it was
supposed to appear. (ring.png)

The final image, a greyscale and colour montage, is the same image file
saved twice, once in full colour and once in greyscale. AFAICS saving in
greyscale seems to be a standard feature of the JPEG format, even if not
widely used. For the purists, my attached zip file contains nebulabw.pov
which simply averages the RGB components of each colour. This may not look
exactly the same as the greyscale image in my submission, but I can honestly
say I have not been able to percieve any significant difference.


OMISSIONS:

You may note that there are no background stars. I did have some originally
(and what's more, I had made a point of ensuring that they were different
sizes and colours rather than all shades of grey) but the main sample
nebula images I was working from were often so closely focused on their
target that they did not have any background stars visible. So in the end I
decided to take them out of my own scene too.

Lens flare: I had decided not to use any 'lens flare' on the central star,
even though this is quite common among nebula images. I'm assuming I have an
idealised camera that is not subject to such distortions. This was decision
was made moot by the fact that a final texture change made the central star
completely invisible anyway. See below.

One annoying and unintentional omission is the central star from my own
scene. The smaller test renders tended not to pick it up anyway, because it
is quite small, and it was only too late that I found a texturing oversight
had erased it (again, darnit!) from my final high-quality render. That's why
there's an ugly black blob in the foreground 'ring' structure. The central
star is supposed to have 'burned through' the thin shell of gas there.
Unfortunately it's been mugged by an error in the inner nebula texture, I
suspect, so it isn't visible after all. I will be correcting this when I
get a chance to re-render and re-post the image on my own web site. Probably
in a week or two.


LINKS:

For how a false-colour image of a real planetary nebula might be built up,
see here:
http://hubblesite.org/sci.d.tech/behind_the_pictures/meaning_of_color/catseye.shtml

For a wide variety of planetary nebulae see:
http://www.astro.washington.edu/balick/WFPC2/

Image montage in pdf format at:
http://www.astro.washington.edu/balick/WFPC2/plneb.pdf

Photo gallery of images at:
http://oposite.stsci.edu/pubinfo/PR/97/pn/photo-gallery.html

See also the Astronomy picture of the day at:
http://antwrp.gsfc.nasa.gov/apod/planetary_nebulae.html

And planetary nebulae have reasonable coverage in the Wikipedia at:
http://en.wikipedia.org/wiki/Planetary_nebula