Page Last Updated: 03/05/99
	'Water'
	Time to trace: 22 hours 23 minutes 36.0 seconds (80616 seconds) Image Description The inspiration for this scene came from my summer holidays of many years ago. Most evenings, my father and I would go fishing off rocks around Anglesey, North Wales, watching the moon and stars come out and highlight the bay. Most nights, the local fishermen would also lay crab pots and mark them with buoys. This was my first entry in the IRTC -- I now appreciate just how much effort people put in to their entries! I know it's not a very imaginative scene, but it has come out almost as I wanted. If I had time, there are still a couple of things that I would change, mainly the quality of the land and the water normal. Image Notes The water is a large box with a dark, highly reflective surface and modified by the wave normal. I tried to get this to work on a *large* sphere to try and get a better horizon, but just couldn't get the thing to work. Oh, well, never mind... The small boat is straight forward CSG. I think it still needs work to add more details. The anchor chain is also built by CSG... just like it is described in the help :) The buoys are also simple CSG, including the addition of a hollow glass ball at the top of the buoy containing a light. There are actually 3 small area lights in each buoy, one for each of the xz, xy and yz planes, of 1/2 strength white light. I found this gives a better illumination of the area. The buoys are randomly positioned in a #while loop and are also rotated a random amount through their x and z axis to give the feeling of 'bobbing' in the water. The land is a height field. I used one of the samples packaged with POV-Ray. The sky is a simple sky_sphere with multiple grades of MidnightBlue. The moon is a large area light and 'looks_like' a sphere that has an image of the moon's surface wrapped around it. The actual image was downloaded from 'http://maps.jpl.nasa.gov/earth.html'. At the site there is also a 12Meg JPEG version of the image (which works out at a 70Meg gif image) and works very well when doing detailed renders of the moon. The stars are 8000 fixed distance point lights disguised as spheres, randomly distributed in a hemisphere around <0,0,0> via a #while loop. Each sphere has a random size and colour. The colours are all 'bright' (i.e. usually more than rgb <1,1,1>) with more emphasis on the blue. I think the effect works quite well (and would work correctly in an animation) but it doesn't half slow down rendering time. There were some very useful comments submitted for the image: From agage@csee.usf.edu: For other large-moon-over-water images, look up Ken Musgrave. Ah, but its the first one I've tried... From lachie@zip.com.au: A little dark, perhaps. That was a real problem with this image. I wanted it to be dark, but I didn't realise how dark it actually was. Next image will have a correctly adjusted gamma - I'll check it on several monitors. From xeo@home.com: A little on the dark side, but very good look to the models Thanks about the models, but there aren't very many of them. From witoslaw@kki.net.pl: the moon is just TOO BIG For a realistic rendering of the moon, yes it is, but it is the size I wanted it for this image. From jerry@hoboes.com: That is a good moon image, isn't it? You ran into the same problem I did, which is that the moon looks really tiny if you make it the right size and distance :*) I think what could use the most work is not the water, but those pyramid things in the water. I don't even know what they are! The stars are cool, but you might try distributing them differently; they look too evenly distributed. Yup. This was my first use of height fields and I didn't have enough time to make a nice terrain. I've since got Leveller and could probably do a lot better with the rocks now. From Sean_Hamilton@amrcorp.com: Good first image, and I like your starry effect. the rest of the image was too dark to see correctly on both of my computers, even with the brightness turned up all the way. Guilty... From gregj56590@aol.com: Poor contrast and/or lighting. ... as charged... From tlyons@gnn.com: too dark to see details, maybe change your assumed gamma? your Honour. From bsieker@techfak.uni-bielefeld.de: I really wonder what took the image so long to render on such a fast machine. The atmosphere of the image is quite good and credible, but it lacks a bit visible details. The image took a long time 'cos each of the stars is a light source, each with its own light buffer. Its the same as when you create an area light and the render time goes up - you are adding a lot more lights to the scene and giving the renderer more work to do. From johnson@pharmacy.arizona.edu: Ok, but the rocks are too regular and geometric. Yup. From Varyk@aol.com: I think a smaller moon, while less dramatic, would have looked more realistic. Perhaps a black disk centered on the other side of the moon, fading from opaque to transparent, would simulate the brightness of the moon drowning out nearby stars, also a similar effect for the horizon (or even some part of you star placement code making placement less likely the closer you are to the horizon.)(or maybe reducing the size of the star-spheres so ones near horizon appear fainter) These astronomer's suggestions being made, nice first entry! Mine too! Thanks for the ideas... From gmccarter@hotmail.com: Powerful lighting. Good concept. But the moon is at least 4x bigger than it should be, even allowing for artistic license. Did you know that the real moon actually appears to be only 1/2-degree wide? Yup, but the moon is the size I wanted. From Martial@Biosys.net: Well nice work for the stars but I don't think that is better than light. Too long time rendering for me (P100). For all Good work and idea! More light next time!