Photo realistic rendering using TV3D?

[harikumar001]

Hey all,

Have a great day. I wanted to clarify something on rendering with TV3D. I understand that with TV3D what we get is a real-time rendered output.

My question is, is it possible to have a higher quality output of the scene using Tv3D rendering of a single frame. Something like 3DS Max rendering. Don't want it to be real time, can take longer. Just like getting a high quality snapshot of what is on the screen. Pardon me if my question sounds dumb, but pls share your views.

Thanks & Regards

[uncasid]

yeah, you can..  you just have to make it   my guess would be this:

1. make an engine that runs in the background of the application
2. have engine render to texture (image)
3. conquer the world!

[harikumar001]

Thank you uncasid 
Still unclear where to start and how to proceed with non real time rendering
Has multi-pass rendering got to do anything with what I aim to achieve?

[uncasid]

well, the way it works with TV.. you do it all yourself.. if you want to do multi-pass.. go for it.  What I would suggest doing is getting used to how to load objects and things of the sort.  Do you plan on making it so you can edit meshes (like in 3d studio)?  If that is the case, you will have a bit of work to do.

So, tell us your goals first, and I will tell you if you are better off with a 3d modeler that is already built

[jviper]

I'd say if you really want TV3D to do the heavy lifting of a non-real time application, it would be no different than how you usually render with TV3D. One difference is you now don't have to worry about the time constraints you have in real-time. That might actually make it easier. Though you might want to watch out for things like automatic refresh where your output disappears when you re-size or move the screen, but that could probably be solved easily keeping an extra buffer/texture/surface or whatever to store your final output image so it wont just disappear.

Remember, TV3D is a rasterizer. It could still be used for non-real time, but if you don't have time constraints, a ray-tracer may be better, or at least more intuitive when it comes to drawing photo-realistic scene. Depends on how you have it planned I suppose.

[harikumar001]

Thank you uncasid and jviper

My application allows users to place furniture in a room in 3D. It also allows them to walk inside the room. That is all real time rendered.

What I am trying to achieve is, I want the user to be able to fix a camera position and be able to render a photo realistic image of what is on the screen to be saved to disk. I hope you get what I mean. Fix the camera view, freeze it and render it in high quality.

While rendering in high quality time is not a constraint. So as jviper said, if I had to proceed with a ray-tracer how would I have to start. Will TV3D be of any help with that or do I have to do it all by myself?

Thanks again guys for your views  .

[Lenn]

Ray-tracing is the least of your problems. A serious lighting engine such as that contained in some popular rendering plugins (example: V-Ray) is a very serious undertaking only for the experts. If you are asking these questions, that means you are certainly not ready to do it yourself. However, there are some open-source light mapping applications, which you may have a look at. The most important characteristic in making images look photorealistic is the second-bounce lighting (and 3rd,4th, while some very detailed renders have a lot more than that). I haven't seen any open-source solutions of high quality that you could have a look at. Still, even if you did find it, it would probably take you months to implement it. So, I would recommend trying out some already existing lighting shaders that make it appear as if there is ambient, direct and indirect lighting in them, such as SSAO, and PCF shadowing, etc. This would require you to know a lot about HLSL though, so good luck.

[harikumar001]

Thank you Lenn 

That post was really helpful.

I have been reading about rendering engines that are Renderman compatible. Correct me if I am wrong. Say if I generate instruction from my program according to the Renderman interface specifications for a rendering engine, wouldn't it be able to render a high quality image based on the input provided.

There are many open source rendering engines available that are Renderman compatible. Pixie is one rendering engine that an application can interact directly with.

So is that not what I am looking for? Am I on the correct track?
Your views are appreciated.

Thanks and Regards

[uncasid]

Just so I understand you..

You are looking to make a program that users can modify layouts of a house.  You have a toolset already built for them, all they do is design the house / put the furniture?  Sounds like what you want to do is make a simple program that does all this stuff, then saves a file format that can be openend in a photo realistic renderer.  I haven't worked with renderman, but I am pretty sure it will have a command driven interface.  You could do something like this   Problem is, you will have to know renderman inside out / the files it works with / lighting / etc..

[Mithrandir]

I briefly looked at Renderman API interface and it seems to be pretty straight forward.
I believe any API using this interface would work the TV. You would have to implement all those methods but I don't see anything that should stop you from doing that.

[harikumar001]

Thank you uncasid and mithrandir 

You are correct uncasid, thats exactly wat i am trying to do.

I have already started reading the Renderman Specification document. There are many rendering engines which are Renderman compliant.

Anyway, thanks so much again for your views. They were really helpful. Keep up the good work.