I’ve been trying to learn how to render photorealistic images, and this is as close as I can get:
How could I improve this image to make it more realistic?
I’ve been trying to learn how to render photorealistic images, and this is as close as I can get:
How could I improve this image to make it more realistic?
What is it? Ice?
Supposed to be, yeah.
There’s a lot that goes into photorealism. A good portion of it has to do with how the subject matter is presented, textures you use, lighting contitions, and material properties. With presentation if you want a chunk of ice to look photorealistic, there has to be something else in the scene cueing you in that this is real life and not just a neat material effect.
If it’s a chunk of ice on snow maybe you should put some twigs in the scene to add interest. Also lower the area of snow around the twigs to make the snow behave the way it would if a twig just fell there. Just adding little things makes a difference.
Also nature never has one light souce, and it’s never pure white. Shadows are not black. They are typically the compliment color of the lightsource. If this is an outdoor scene add a white/orange lightsource as you main one, and then a sky/blue hemi light so that the shadows are blue. Add a background texture so that the ice has more light to refract (and reflect if you have reflections on).
Thanks, that helps.
Which is better for realism, raytraced or buffer shadows?
Which ever gives you the most realistic results :). For me I would prefer Buffered shadows even though it might take a bit more time to set up (setting the bias can make a big difference, you’ll just have to play with it). Right now until that distributive raytracing comes out, if you want to get soft shadows you should use buffered shadows. Raytraced shadows give you very percise results but they are hard shadows, and also take a lot more time to render. And since soft shadows look more realistic, I would use those.
Once thing that buffered shadows don’t do is create softer shadows the farther away an object is from the lightsource and the shadow receiving surface. To fake it you’ll need to have more than one spotlight, mostly even a cluster of them to acheive the right affect.
I use alot of the tipc I’ve received and came up with this, I think it looks worse:
It has 3-point lighting as well…
Here’s another try, I think it’s the best one, 'cept for the no AA and crappy water:
If somebody could help me with good rendering, I’ll be on MSN, Yahoo Messenger! and AIM.
Yahoo: sychosisdragon
MSN: [email protected]
AIM: sychosisdragon
I think it has improved. Except where did the snow go? The edge of the water looks to distinct.
I have a question though. How do you use buffered shadows? I just figured out how to enable ray tracing by selecting “ray” on the rendering menu. I don’t see a buffered option…
I’d suggest that you buy or check-out some good books on studio photography. The kind of stuff that is (still) done using 4x5 view cameras and lighting setups.
The phrase “photo-realistic” is really an oxymoron. A good looking photo isn’t realistic; not really. Nor does something that is realistic look good. You can take a crappy photograph with a cheap camera and … by definition it is “realistic” (that’s reality in front of that lens, after all), but it doesn’t look good.
When you are doing an image, consider this. You are depicting something, either with CG modeling or with film, and that image is immediately going to be squashed from three dimensions to two. It’s going to be compressed from the gamut of colors our eyes can see, to the color-space of a CRT screen, or perhaps that of a digital printer, or perhaps that of color negative/print film. The light that radiated from the object into the camera lens has done its work and is gone; now all we have is an image that can be seen in reflected light from a printed page (subtractive, not additive, colorspace), or emitted from a CRT. The subtleties of contrast are gone. The ability to selectively focus with your eyes is gone. All of that original information is now forever lost. But, when the viewer looks at the image, all of that information is going to be reconstructed ‘at a glance.’ Since the actual information is gone, the reconstruction is going to occur based on what is left … specifically, on the cues that you put there expressly for the purpose.
And this is what a good photographer, or visual artist, must learn how to put in. The result is not, emphatically is not, “real.” But it is something that the eye will not only accept as ‘real,’ but will find pleasing and that will serve the dramatic purposes of the picture.
Even though CG sounds worlds-apart from photography, videography, or film-cinematography … it plays under the same rules of physics.
Here’s an update, I haven’t added and room reflections yet, but I will. The reason the edge of the water is so defined is because I don’t have the OSA on:
Once again, if someone could help me with this, I’ll be on MSN, Yahoo Messenger! and AIM.
Yahoo: sychosisdragon
MSN: [email protected]
AIM: sychosisdragon
Okay, new image, but I think it looks more realistic than the ice cube:
Any way to improve on this?
Is there a way to make lights non-reflective, so the light source doesn’t show up on a reflective object?
Yeah, just turn the specular off. It’s under the material options. You can also do that with the light source as well.
>>Okay, new image, but I think it looks more realistic than the ice cube:<<
Personally, I thought the last ice cube render was more realistic. Your ball looks like it’s trying to be a chrome ball with glass attributes, and not really doing either one too well.
Do you render that last ice cube using scanline or raytracing?
Uh, I don’t know what scanline is, but I used buffer shadows, if that’s what you’re asking me?