Hey Guys,

This mainly directed towards students since the professionals already know everything about writing shaders for renderman. I have 3 different renderman books but I suggest for students wanting to learn that they first read Essential RenderMan Fast by Ian Stephenson. This book breaks down how to write a rib and then goes on to tell you how to generate ribs from C code, then how to write surface shaders, displacement shaders, light shaders, etc. This book is a much need foundation before trying to read Advanced Renderman or Texturing and Modeling a Procedural Approach. The Essentails book does not require you to know Calculus 3, Linear Algebra, and Calculus based Physics I & II; however, Advanced Renderman and Texturing & Modeling a Procedural Approach do. This is just for all those students that want to have a better understanding of Houdini and how to write and create shaders in VEX for Mantra.

Cheers,
Nate Nesler

I know that Renderman is cool and that it's becoming the standard de-facto of the shading language theory.

But, just curious, - is it really that necessary to study Renderman in order to understand how the shaders work? Isn't Mantra and VEX language not enough? I mean - well PRMan is tough, but - 1st - it's pretty expensive, 2nd - Mantra is not that bad at all and it's constantly developping, and it is also sometimes consiedred to be a production quallity renderer (remember “The Time Machine”), 3rd. - Mantra has its' advantages over PRMan - like very fast SSS and Motion Blur. So - unless u r planning to do another StarWars or smth. - why spending time and money (and this is really BIG oney) studing PRMan?

BTW - what is Physics I, II and Calculus 3?

First RenderMan set's the standard for feature film rendering right now. Actually for quite some time.

Second, it is necessary to study how shaders work and how RenderMan and Mantra call shaders in what order and when, if you want to do more than just layer shaders and use simple lighting models like phong, blinn, lambertian with lots and lots of texture maps.

The big issue with writing shaders for RenderMan and Mantra is managing anti-aliasing and the nyquist limit (what???) due to the very nature of how the two renderers evaluate shaders at the micro-polygon level. Advanced Renderman and the Procedural Texturing and Modelling books do go in detail on how to anti-alias your shaders and that is where the understanding of calculus comes in most.

You can apply what you learn to VOPs though and for the most part completely disregard aliasing issues as each VOP is written to be self-anti-aliasing. Just looking at a vop network won't get you there for advanced shaders. How do you write an illuminance loop let alone do one in VOPs? You have to read the Renderman Companion to get the basics then read the Advanced Renderman Guide to fill in the holes then attempt in VOPs. It isn't that hard, just takes time.

You ask “why spending time and money (and this is really BIG oney) studing PRMan?”
Answer 1: There are so very few people in this industry that can write even simple trivial shaders. This industry needs more visual effects artists that can re-assemble ribs and write their own shaders to get around production issues. All shops certainly could use strong shader writers along with good Houdini skills. Not another slim user please.
Answer 2: Learning how to assemble a shader, how to deal with attributes and the general understanding of programming constructs will make you an even stronger houdini user.


-jeff

Thanks Jeff,

that was quite informative.

But what I was talking about was not actually “Who needs PRMan?!” but - isn't learning Mantra and VEX Language (not VOPs) does the same? I mean - u learn VEX Langage - u write your own shaders - u understand how shaders work - no?

And Second - Is Mantra really that bad so that PRMan is the only rendering option?

Thanks.

PS. This is not just arguing for arguing - I'm interested in this, so maybe somebody else is interested too.

I think the question is a little misplaced. It's not a question of “is learning rman any different than learning mantra” - the answer to that is yes, of course it's different! They share many similarities but they have differences in some approaches and certainly syntactically. What I believe Jeff was trying to point out were the concepts behind the renderers…*those* are tremendously valuable and can indeed be shared between packages.

But “learning mantra” is more than knowing to convert a vector to three floats…it's that for sure, but learning techniques outside of the specific renderer then applying those concepts is what it's all about, once you get past tmaps and shadows.

Mantra isn't bad at all, it's not even “not bad” - it's really good. I've watched Mario scream, yell, and beat his head against the wall: but he gets some really nice pictures out of mantra - it's all in the little niggly details of implementation that the differences start to show.

And it's worth noting that the screaming and yelling used to happen with prman too. They all have their strengths and weaknesses.

Cheers,

J.C.

Thanks, JColdrick - u've been most helpful as usual.

1. I do not argue the statement that it's always usefull to know more then just basic tmaps and lights - i.e. understanding the “background” - how shaders work and such… But what I'm rrying to say that u necesserily learn it whenever u go behind the VEX Language. - So and if that is the goal (understanding *how the shaders actually work*) - why learning PRMan then?

2. That's just the problem of distributing your time - that's what I'm trying to clarify - Why learning 2 renderers that are both *production quallty renderers* ? Isn't it enough learning just one?

Again I mean - yes - u unavidably learn that *shaders background and math* whenever you learn PRMan - but don't u do the same, when u start learning Mantra and VEX Language?

Thanks.

I think the short answer is that you're not going to find books like the rman ones mentioned here, or online help for that matter, that goes into the same technical depth for mantra, that's all. There's few *pure* theory books on shading - they usually use an existing app as an example. Unless of course you want to study Siggie papers, which are *very* technical.

Cheers,

J.C.

Ok,

Hope, I see your point, JColdrick,

So learning PRMan is similar to learning some *Shaders' Bible* ? And even if I wished to find some *Mantra's Shaders Bible* - I simply won't be able to, coz it doesn't exist - is that right?

And,btw, does SESI plan to public more materials on the topic in the fufure?

Thanks.

I think, PRman have a lot of materials for learning already. But VEX is not. That's why… They are quite simillar shading language. I know VEX can handle sop, cop and more.
try like this:
Learning PRman -> Understand different context(Happy rendering with VEX is good) -> Try to convert RSL to VEX and build VOPs network)

Hey,

Yeah That is correct about there not being bibles of Mantra. You need to understand that these renderman books do not teach you the math you have to already know it and the physics. They just tell you how to implement the math and physics for your shaders. Physics comes into play more in regards to lighting. Besides Render Man is still kind of the industry standard. Last time I looked they held around 70% of the market but that was a while back I am sure Mental Ray has cut into their percentage by now. Mantra is more like render man than probably anything else so it seems the best choice for learning the more complicated ideas and structures for building sucessful shaders.

Cheers,
Nate Nesler

Thanks everyone.

Hope it clarifyied much for me.