Hi guys, i've been working some more on my retro, and i ran a high quality radiosity render. However, when i did that, the reflections off the car became far less reflective, and dull. What could be causeing this? Is it the color of the material or the radiosity? Well, i hope you guys can help. I've put two examples, one with the radiosity, and another without it.

Did everyone that used to frequent this forum die? i posted this some time ago and no responses have come out yet.

now i have another problem, but i'm not sure it will ever get answered. There is this line type thing that appears on the right most wall. The picture i first posted i edited with photoshop... (you can tell cause it's all blury and the grain isnt there.) It wont go away with upping the samples, or maximizing jitter. What could be causing this problem.

I cant do this alone, and i've used up all my ideas and experimental knowhow... wich is actually quite small in comparison to the programmers around here... actually it's more like non existance... i can barely use the property graph mapper thingy. Please help me.

Is that wall made of just one large patch? If so, try rebuilding it with a few more splines (not difficult if you use the grid wizard).

I had a similar problem on a recent project. I had a sparsely constructed floor (sparse compared to the character that was walking on it) which showed definite signs of a line. The glitch showed up quite badly as the camera moved over it. The problem was vastly reduced by increasing the floor's patch count. That said, the floor was vastly bigger than your wall, so I might be talking utter rubbish.
[attachmentid=10473]

The answer is complex but since we are in the lighting forum, here it is.

In a radiosity rendered scene, ambiance values on surfaces doesn't mean anything usefull. The ambiance surface property have no equivalent in reality. It was originally placed in the rendering equation to cheaply simulate what radiosity is now doing in a time where radiosity was only a concept in physics research. So it conceptually is an ultra cheap radiosity solution.

Following this logic, in A:M, ambiance surface property is treated the same way as radiosity. It is not added to radiosity. Because of that, an ambiance surface may not be brighter than the not-ambiance surface next to it in a radisoity scene. In some circumstances, it may even be darker because the ambiance is equivalent to a 7 f-stops in photography. In a radiosity scene, the total illumination may very well be way higher than that.

This is what you are seeing in your car reflections.



There is usually an explanation for that sort of artifact. Did you try to change your light positions and properties? I would need to take a look at the project file to find out. So I suggest you submit your project with a description of your issue to A:M report if you still get this artifact after changing your light settings.



In situations like that, where your texture is to be seen in very sharp perspective, the best approach for a solution is to use a higher resolution bitmap for your decal.

ok, here is the project file. I changed some light properties... but still no luck.

It was a low-res decal with lots of repeats - it never occurred to me to use a higher resolution bitmap. I'll try it - thanks.
And my apologies to Dan for hijacking his thread.

I'm trying to understand this (radiosity)... So, if it's okay, I'll make a statement based on my perception of the second render and what Yves has said... maybe someone (Yves?)can correct it if it's wrong.

Statement:

" In the radiosity render, I can see that the car paint is reflective and "a good paint job" (well buffed out), etc... I think the render is pretty accurate in that the "skylights" in the "garage" appear to be well above the camera, so they are high (this is a very high ceiling) and so, the light "weakens" over distance. If you want the reflectivity to show better in this picture your lights would have to be closer to the vehicle and brighter if possible. Otherwse, for a more shiny example of the car, the original (non radiosity render) is more (commercially) appealing because it makes the car look as we would hope to see it in the "bright" world."

End of Statement

I base my statement on a couple of observations of radiosity renders and the following text.

According to a book I recently got thru inter library loan:




This is from: Inspired 3D Short Film Production By Jeremy Cantor and Pepe Valencia

The bracketed [?s] are my own additional questions and the bold highlight is mine, too.

BTW, the book is a "college type" text book... a bit of a dry read, but full of lots of info, and lots of mentions of Alien Song by Victor N... and a CD with Raf Anzovin's Puppet. I recommend looking at the book, but I won't spend $60 of my own money for it since it barely mentions A:M and I won't be using any of the other animation/modelling programs (but this is not a book on technique, it is a text covering all of the aspects of short film production from concept thru distribution.)

'nuf said... sorry for the long post.

True. But this leads to a misconception in the case of reflected images.

The light attenuation is true because photons spread out from the light source to the illuminated object. So given a light source, as the object goes further from the light source, it receives less and less photons. So if you look at a surface that is being illuminated by a light, this surface will get darker as the light travels away.

For reflected images. This is also true for photons that reaches the eye (or the camera). But there is a catch. As the object goes further away from the camera, it also appears smaller. In fact it appears smaller in the same proportion as the photons spread out. So in the end, the object appears just as bright even though it looks smaller. Suppose that the light shining object is placed in a completely dark room. As it moves away, it covers less and less of your visual field. So your visual field receives less and less light but the light it receives is more and more concentrated. In more mathematical terms, the eye or the camera, receives the same quantity of photons per solid angle.

The bottom line, a light should not appear darker in a reflected image.

But this is not the issue in the render. The patch is no light. It is a white colored patch with ambiance setting.


No. Soft reflection is a different phenomenon not related to radiosity. It only depends on the surface roughness of the reflective material and not on the distribution of light in a scene.


In radiosity parlance, specularity and reflectivity means exactly the same thing. The specularity property that is used in traditional ray-trace render is a cheap (in the sense that it is very easy to compute and can render really fast) trick to simulate soft reflection of lights. This works fine with raytracers because nothing contributes light in a scene except lights. But for radiosity scene, everything would contribute lights in a scene and thus, specular surfaces needs to sample the whole environment to for each pixel to simulate specularity. This is very costly to compute. This is why it is said to be not good at handling specular surfaces.

Thanks, Yves...

I learned quite a bit more. Your examples and expertise are always helpful and informative.

Sincerely,

So then what exactly is causing this line yeves? I read that post you made several times, but i dont think you said what exactly to do in order to fix it. What should i do to get ride of that line or the reflective problem? I'm still pretty new and you scared me a bit with all that tech talk, i understood all of it but... man

sorry guys... i just now looked at the project i archived, and apparently when you transfer over the project file it doenst transfer the objects or.. anything at all with it. sorry for not picking that up sooner. So here is the full project file. Please let me know if it's a problem lies with the lights or the scale of the model itself, or something other than that.

Thanks for all your support guys. You've realy helped me out a lot. I apprieciate it.

Well, i guess it's not doing the line any more... i may encounter it later, but the latest render i did had no visual problems. I'm still not sure how to fix the reflection problem though. How do i get the reflections of the luminairs to be white but still retain the deep red of the car? is that even possible? I should probably create an environment to show this car in.

Ok. I've created an environment for the car to reflect now. everything except the brightest part, the sun, is reflecting normaly. Is there any reason it's yellow? the decal is pure white up there. Is there any trick or gimmic that will make that area super bright? Other than that... everything is looking good, and only the sun has reflective dulling now. yay

Tricks is to play with Reflective Blend ans Reflective Filter.


For light reflection, you will control that through Specular Size and Specular Intensity and you can also change the Specular Color if you need but my guess is that you can get what you want by increasing the Specular Intensity and reducing the Specular Size..

Wow! that worked out great yeves. I had to increase the spectral intensity to 500% to get it to be white. But it's looking much better now. I redid all the lighting in the shot and now there's a much more pronounced sun as well as 9 sky colored lights. Thanks yeves for all your help so far.

New problem tough now there's this very strange black box appearing in the reflection of the sky on the roof of the car. I've moved the camera to look at the area and that black box isnt there on the dome. Any idea on what's causing this problem?

Never mind. I found that the problem actualy came from an anchored spline. I've since removed the anchored spline and created a full spline to wrap around the roof.

Looking good. I find the front hood flatness does not quite fit with the rest of the car though.

BTW, My name is written Yves. not yeves.

ya... i was thinkin that too. i have a whole bunch of things i want to do with it... i'm going to add a separate part, to create a crease, so the hood looks like it would be openable in real life, as well as add the logo in the front and bring that trim accross it. Hopefully that will be enough... and not too much.

sorry yves... i still havent figured out how to say your name well enough... so sadly that spelling mistake is how i'm sayin it. but... spelling mistakes aside, thanks for all your help again on the tech issues.

Pronounced "eev". But that will not help with the spelling I'm afraid

wow dan, that looks really good...!

But you shouldn't have told us that the black spot was splinage, you could have said...

"It was a bald eagle that was flying over at the moment I took the photo."

I woulda bought it... really!

And ain't Eeev great? He's so knoledygable!

Snicker snicker...

With a middle name like Chassaignac... I really shoudn't pick on anybody elses name though.... huh?

A square bald eagle aye? Regardless... that is not a reflection of the sky rather it is a defect on the surface of the model surface itself. When i move around in animation it looks like there's a black piece of electrical tape on that area. It seems every time i fix one of these anchor points, a new surface "black box" appears near another anchor point somewhere else on the cars geometry. It's starting to get on my nerves. A movie i just rendered of the car spinning showed that a black box sprang up near the back on the roof of the car.

by the way... how do you post videos on the forum?

if you can't post them here, try work in progress (WIP) There is a limit (i think of 1 MB) though so it needs to be compressed.

Square eagle... nah just very high up... no paint job is that good... (and besides, I didn't click to enlarge, so it was just a spot)

And BTW, What Yves sed about the flat hood... very hard to make a flat panel in steel IRL... needs to have some arc/crown to it... usually in both directions.

Not to detract from your Retro... it's just a plain fact... and slight crowning would make it reflect more interestingly, too. Just MHO, your mileage may vary.