ypoissant

Nice finelyd etailed shovel. It does

I clicked and watched the clip before I read your explanation and I must say that my reading of the clip matched perfectly your goal. This is very well executed. Bravo.

I Love it.

1) By adjusting biases and/or by applying the porcelain material. 2) Splines with only 2 CPs cannot have their bias adjusted. Those can only be straight lines. Because you extruded, the connecting splines are 2CPs splines. What you can do is extrude 2 times more and get rid of the internal patches that will be created because of that. Then you can tweak the biases to get a beveled effect. 3) This is a real-time glitch, The spline doesn't actually warp. Just press spacebar to refresh the display and they should get back to normal. 4) Indeed, those are to be removed and avoided. TinCan tutorial about internal patches will help you. 6) For valid 5pt patches, all the 5 points should be in a convex organization. 3 CPs aligned on a straight line as you have at several places will not work. Apart from that, sometimes, you may have to flip the normals of the 5pt patches to have them fill. If you find a 5PT patch will not allow you to click the 5PT-patch button, then hide everything else, reselect the 5 CPs using the grouping tool and then your 5pt patch button will be available.

Thanks

I'd say that concerning the lighting and texturing, you pushed it far enough. Now you need to think about your composition. In your current composition, there are too much space where it is not required and too little where it would be important. There are too much space at the top and left side of your composition and way too little space at the right side. The floor is acting as a frame while the boxes should take on this role much more. The boxes should frame your car. And your car should be the unmistakable center of interest. Right now, your car is crammed in the lower left side of your composition. This cramming is reinforced by the fact that it is turned toward the outher left-lower direction. But it doesn't even have space to move in this direction. In the attached quick composition, I widened the image so that the car have more space in front to move. I also moved the car so that it is better centered. Some overlap with the box is not a bad idea too. But now that I look at it, I would move the car a little more in the north-east direction. Then I moved the plastic base so that it participate in the framing of the car with the boxes. This is just a quick shuffling of some objects and of course, I couldn't rotate the objects around but that gives you an idea of some explorations in composition with your subject. A simple and effective rule to follow is the what is called the Golden Section. For starters in composition, this is an easy to apply rule but don't get too carzy about it. This is only a tool that can help you think about your composition. Another generally good idea to apply is to use objects in the scene to frame your main subject. That is framing.

Floor saturation is much better and believable. I would lower the floor contrast a little though. Here's another observation about reality: The boxes should be more reflective and with a smaller specularity size where there is black paint. And the interior should be almost completely diffuse (with no reflection and specularity). Also, the transparent plastic on the box should show at least some form of manipulation. I would put at least a bump map if not a displacement map there. As for light leaks, could you post a bird's eye view of the choreography along with your radiosity properties?

A 4M pixel camera should be enough. But if you intend to render your map this close, you would have to photograph it that close too. As for how far you go for realism? I fully agree with Martin view that realism is not needed for animation. In fact it is not at all required for stills done with an illustration purpose either. I personally prefer an artist who have developped a stylized recognizable style. And this is not generally based on realism. Of course, a realist artist can distinguish himself with subject matter and composition. But as far as color and light scheme are concerned, the choice of realism pretty much impose strong limits to the artistic freedom. I guess, one untold story of realism in 3D is that it requires a lot of knowledge that lies more on the scientific / technical / analytic side than truely intuitive / artistic expression. The underlying argument here is that it is not sufficient to just trow a ton of photons in a scene to get a realistic render. You must also get all your surface properties (textures) right and within the same color space, and all the surface details right. Here I only touched the bump maps for instance. But what about reflectivity maps, and specular size maps? In this scene, I would also use soft reflection. The hard stratas of the wood would need to have a smaller specular size than the soft stratas. And then, when that is done, you still need to take care of the next surface property. That sort of knowledge seems to never end when your goal is to do realism. And this is where the expression "Render Masturbation" takes all its meaning. It takes an immense amount of fiddling with the renderer before a truely satisfying realistic result comes out of it. Is it worth it? Which side of your brain do you prefer to use? Left side of the brain or right side? If your main goal is to get a message or a mood across, then photorealism is clearly not the road that will get you there faster.

Cool model. I like a lot. I can't wait to see her moving and singing. You should try the skin shader on her.

What are your render specs? Nice character BTW. I find his sides a little too flat though.

Very nice render Stian. Here are some comments: 1) Indeed, the floor looks way too saturated and contrasted and makes the rest of the objects appear to float or not to live in the same world. That is a tricky issue to get along when atempting photoreal. The main problem here is that the wood texture color space does not match the boxes color spaces. Currently, the wood saturation, brightness and contrast is way above that of the boxes graphics which should normally be much more eye catching. My guess is that the two texture maps do not come from the same capturing devices. You mention that you scanned the boxes. But I doubt you also scanned the wood floor. You probably photographed the floor. Now, you must match the wood texture saturation, lightness and contrast with that of the boxes (not the reverse). You can do that in photoshop where you would load the two maps and pick your eyedroper tool and note the saturation and brightness from the HSB display of the brightest yellow and orange colors on the box. Now, using a "Hue, Saturation..." and a "contrast-brightness" adjustment layers, you must decrease both the saturation, lightness and contrast until you get a match with those of the boxes. The attached modified render I did in Photoshop gives you an idea. 2) The shadows are missing from that render. Theyt are obviously there but they are too light and too soft to really support the objects in the scene. Make sure your shadow darkness is set at 100% and reduce the width of your light to maybe 50% or their current size. And of course, never, ever, use z-buuffered shadows in a radiosity render. 3) The light leaks under the boxes and the car. There are two issues to consider here. The first thing to consider, and the most important, is the thickness of your boxes cardboard. Any object you model to put in a radiosity scene must have some thickness to all their surfaces. When photons are deposited on surfaces, they will show through on both sides of the surface. If your surface have thickness (is double walled) then the second wall will stop their leekeage. I see that your boxes cardboard seem to have thickness. But it may not be thick enough. And your car may not have thickness too, especially under the car, which leaks light on the floor under the car. Note that the larger the "Sample area" the thicker must be the surfaces. Second, you can somewhat control the effect of the light leak by enlarging the photon mapping "Sampling area" and "Photon Samples". You will have to experiment to find a set of parameters that work well for your particular scene. Increasing those values have the effect of bluring the light leaks and they become less obvious. 4) Your floor is missing bumps to make it look more realistic. On wooden floor, the darker strates are harder and the lighter strates are softer. THe effect os that the darker strates forms mounds while the lighter strates forms valleys. So you take your bitmap, convert it to gray, invert it and basically apply it as a decal specifying bump type of decal. You then control the visual appearance of the bumps by adjusting the percent value. I also like to add darker lines to the bump bitmap to add groves that mark the separations between the planks.

By applying to the skin group, you mean that you have a group for the skin that excludes the eyes and that you set the diffuse shader for that group to Skin Shader? What would help would be you post a screen capture of the PWS tree of the model and the groups with that group properties expanded.

There is no clear cut rules. Deciding when to use radiosity and when not to, requires a lot of knowledge about light behavior. This knowledge cannot be communicated through a simple answer. I invite you to visit the Radiosity forum. Inparticular the Lighting tutorial and the Cornel box tutorial. With that knowledge, you should be in a better position to decide to use radiosity or not. Radiosity is not a magic wand that is sure to make your render look right. It is more like a tool that can do marvels in the hand of one who knows exactly what to expect from it. If you don't understand the technical implication of using radiosity, chances are you are not using it right, it will not improve the render quality and will considerably lengthen your render time for nothing.

I just wanted to mention that you really don't need radiosity for that sort of render. Cool render BTW.

I agree with most of Peter's comments. His expressive use of light, in particular, I agree completely. Indeed, it is a godd advise to start exploring with as minimal number of lights as possible. Unless one have experience with lighting, it is better to start with one main light to get the most expressive basic lighting setup. Then add secondary lights like a fill light and then a rim light. After that you can add bounce lights to help get the mood and enhance relationship between the objects and characters. But beware that too many lights can truely kill a scene by making all objects appear flat. Secondary lights should be used parcimoniously and subtly. I agree only partially with the comments concerning skylights however. I also find it odd that several beginners starts by putting a skylight rig in their scene without even understanding how to manage even one single light. But that is not a reason to dismiss skylight rigs altogether. Skylight rigs can make the objects in a scene appear much less flat because of the ambiant occlusion effect. But that is if the skylight rig is used correctly and with the proper settings. It is true that skylight rigs are not easy to use. Finding the proper babance between a skylight rig light intensity and color vs a sun light intensity and color to get the right mood can require a lot of experimentation. Skylight rigs are not magic wand that one can use to automagically get a cool clay look. So skylight rigs are not for beginners. Most of the beginners I saw trying to use a skylight rig did it the wrong way and had big difficulties in setting them right. The bottom line is it is better to start with the basic 3 light setup. Learn what is this beast that is a light first. How it works, how it is adjusted, what it does. etc. first. Then move on to more complex light setups. BTW, this was not a comment addressed specifically to you Doug. I know you didn't use a skylight rig in your scene. And bravo for setting yourself such a learning goal. I invite to move your effort in lighting in the Radiosity forum so that we can follow your progression and advise as you go. And have everybody inerested in lighting learn at the same time. About your setup, Doug. 1 ray cast lights will produce way too harsh shadows. You could use z-buffer shadows but that have the potential to make your objects look floaty. Instead of using only one ray on your sun light, you would get better results by using like 5 rays but reducing the size of your sun light. The shadow smoothness is controlled by the light size. From the softness of your first shadows, I would say that your sun light is very very large. You could easily reduce its size to at least 10% its current size.

You might want to start with the Cornell Box tutorial that you will find in the Radiosity forum. Also check those tutorials about Photon Mapping.

Try a 25 light rig and render at 9 passes. That should give you about the same render time but better distributed light rays and thus better shadows. Nice model. I like the smoothness of the body.

Brian's window thread.

Nice render indeed. The robot is super cool of course but I particularly like the attention to little details in the scene, like the cat in this instance, that adds to the story and to the composition. That you didn't use jittering in final gathering produces this painterly daubed look. In some scenes, this can look not so nice but you managed to pull it really well here and this gives your render an added artistic touch. Bravo. BTW You busted my record render time which was 147h two weeks ago. And I was not even using Photon Mapping, Just a skylight but with a lot of hair.

Doug, I hope you realize that your display is set to some odd color settings. I checked your posted images and they all appear to use only 4 bits (16 levels) for red and blue and 5 bits (32 levels) for green. That is even less than 16bits. That explains the heavy banding in all your images. I would suggest you set your display property to 32 bits or 24 bits to get optimal view of your preview renders.

I wouldn't advise doing your storyboard with your models and scenes and with the same application you will do your animation. At least if you can draw, let your imagination drive your drawing. Explore facial expression, body language and camera expressiivity. Things you can more easily do with a pencil than with the 3D applicaiton. To get the same finesse in expressivity that you can get with a pencil will require much more work with a 3D appplication no matter how easy it is to use. And it is better to put on paper the images you have in your mind rather that whatever you can accomplish in a given time with he application.

The joints between the metal sheets are curved inside so they look like they are beveled. But some of your sheet joints have a too large radius which makes them look balooned. See attached pointers.

Very nice render. The DOF help a lot with the horizon and the sky gradient help getting the car form out. Two things I would suggest. 1) Use soft reflection and a very large scale but tiny height roughness on the ground to make it look more real. 2) make the sheet joints curvature tighter. Edit: I mistypes "soft shadow" while I was meaning "soft reflection".

Good texturing job. Great attention to details here. There are just some little details that are killing the render. Those black tubes in the janitor trunk which are completely lacking the level of details that the rest of the objects have. And they immediately look like our typical garden hoses and because we are seeking for reference objects, they make the janitor look like a 1 or 2 feet tall robot. For a mechanical object this size, sense of scale would be better if there were more small-scale mechanical nurnies. Lowering the camera position would also help with the sense of scaling issue.

Really nice model. I like your other models, like your cow too. That bicycle is full of mechanical details which adds to its believability. For your chrome shot, I woull add a skydome with gradient from deep blue at zenith to white at horizon, to act as reflective cue. This should give your model, or at least the whashed white reflection part, a much more 3D feel. Even if you add color to some part, the nice gradient of such a skydome will add nice visual depth and curvature cues. Also, switch your main light from z-buffer to multi-ray raytrace shadows. The regularly blured shadow doesn't give enough support to your model.