Maxwell Render | Official public forum

I understand that RPC is fake but this is the fastest way I have ever seen to make nice scenes. Is there a chance of using RPC with Maxwell?

What's RPC? Never heard of the term in relation to CG.

This has been asked for many times before; I've even talked to Arch Vision who has said they are willing to work with NL on developing a plug-in for Maxwell. As far as I know this is a very low priority for NL and you shouldn’t expect to see it any time soon. I've been using AXYZ Design people since there actually 3D models, they work pretty well as long as you don't get to close to them.

I officially hope they DON'T put RPC support in Maxwell. RPCs drive me absolutely crazy. They are PHOTOGRAPHS of people/cars/trees/etc. So unless your rendering looks like a photo (and how many of us can truly achieve that) they stand out like a sore thumb.

I find them very usefull.
I try to use my own RPC material, for diferent reasons, bur The tool is great.
I vote for RPC support in Maxwell

Ernesto

misterasset wrote:I officially hope they DON'T put RPC support in Maxwell. RPCs drive me absolutely crazy. They are PHOTOGRAPHS of people/cars/trees/etc. So unless your rendering looks like a photo (and how many of us can truly achieve that) they stand out like a sore thumb.

I agree, but I need them because the clients like them and they are "lighter" than 3d models of people.
Also, it can be a way for us to improve quality because with maxwell, our renderings SHOULD look like photos.

JTB & Misterasset - RPCs although are real photos they don't fit well with a photo realistic render because they look like they were put on in photoshop. They don't have realistic lighting on them and shadows etc are problematic. I use them only for NPR stuff, but they are quick and dirty. So I like them for some stuff. AXYZ people or lowpoly seem to be the best solution so far, as long as they are not too close. I think they are starting high resolution models though so that may change.

Regards Peter.

The problem with RPC's is they are basically ignored in most render engines except for Max scanline and radiosity. They need to change the nature of the RPC so that they interact with light because the concept behind them is no different from the way we use maps in our materials.

Peter Shupe wrote:JTB & Misterasset - RPCs although are real photos they don't fit well with a photo realistic render because they look like they were put on in photoshop.

So what???? they are quick, faster when rendering, many to choose from.
The clients (at least most of them) don't need 100%quality, but they all need 1000% speed. (Talking about architectural renderings)

AXYZ people or lowpoly seem to be the best solution so far, as long as they are not too close. I think they are starting high resolution models though so that may change.

Regards Peter.

Why are you all talking about people models only? What about car and trees???? Imagine 50 trees from ONYX or SPEEDTREE and compare them to 50 RPCs.
Then you will understand why you don't have a clean picture after 50 hours and why you can't render more than 1600X1200.
There are many advantages of using RPC technology.
I agree that the best results are when we use 100% models but...
1. I don't need the best results for my work.
2. I don't get paid for the best results, I get paid for the results that make my client happy (Surely you don't expect from every client to even know what RPC or AXYZ is).

First off, I'm not being mean, but, if you're extremely worried about time and you want to use RPCs then why are you trying to use Maxwell.

Not to go back to the car analogy, but if you want to carry a family around, you buy a mini-van. If you want to have a midlife crisis, you buy a sports car. You don't tell the people who make the sports car that you want to put your whole family in it.

misterasset wrote:First off, I'm not being mean, ...

Are you sure??? You have to try more

1. I am not worried, I am not trying to use Maxwell, I have already used Maxwell (and loved it) for more than 8 buildings.
2. About the car analogie....

You don't tell the people who make the sports car that you want to put your whole family in it.

I know, but the people who make the sports car, know that most of the customers have families so they take care of them too...

Anyway, it was just a simple question... I think we would all benefit from RPC and M~W together, just my opinion...

But what would be the difference between putting in RPC trees in the render, and pasting trees later in PS, after the render is done?

Many people want semi transparent vegetation, so in PS I can tune that transparency to the exact value they want. If I was using RPC, I would have to re-render the entire scene wouldn't I? Also in PS I can more easily tweak color, clean them up etc, because everything is separate.

Trees and vegetation rendering together with the scene makes sense to me if they will be influencing the lighting on the building a lot, in which case I'd use real tree geometry (tweaking it so each tree isn't 300 000 polys).

It seems to me with RPC trees I get neither the benefit of realistic light interaction with the environment, neither the benefit of easy tweaks in PS.

What about animations, you can't realistically put 300,000 Onyx trees into a scene without running into extremely long render times or running out of memory. That's where RPC trees, cars and people really do well since they have a very small memory foot print.

In my opinion, RPC is a view dependant object.
If that is the definition; I find it very usefull, and much better than using flat planes with maps projected on them.
The main advantage is that you do not have to rotate the flat objects each time you move the camera.
I do not care RPC (brand) people, they have a very bad alpha channel and their maps comes with a fake shadow that I do not like.
But RPC as a tool, is great. I mean RPC as View dependant objects.
I create my own RPC objects with custom photographs, and also using Imagecels people library. I find them very usefull, and very realistic.
Which is the sense of pasting them in photoshop as posprocessing?
If you place them in the scene you will get real reflections, real shadows, and realistic lighting on every people you place.
It is the ONLY way to get a perfect solution. Otherwise you have to be aware to place people at the right scale, pasting shadows, adjusting brightness etc. With the disadvantage of loosing all that working time in case the viewpoint is changed.
The best way is placing RPC or View Dependant Objects People, and render a batch. Then you will get all the images ready to print.

Why should I wait forever to get the image done in maxwell, if I am going to fake it pasting stuff in photoshop!!!!!
No way! this is my strong opinion!

Ernesto

PD if you do not like RPC company at least create a view dependant object that can be used into maxwell Studio!


Up to now there is no other solution for people.
The 3D low polygon people is not good enough. You will find that they work from certain angles only. They look like monsters in some cases.
the 3d people is not good enough nowadays, and they are too heavy.
If you have a good detailed model, it will have wrong position or wrong proportions, if you find a good proportioned model, it will not have enough detail.
Regarding the Flat tree maps or RPC trees, they are veruy usefull.
I personally do not like RPC brand trees, they are bad 3d trees rendered into the RPC file! I prefer the 3d trees as Mihai 1000 times than an RPC tree.
But sometimes it is usefull for backgrounds, or special details which would be very difficult to model.

Please separate the RPC brand name from the View Dependant Object Concept.
I do not care if the first or the second solution is selected, but we need the View Dependant Tool, no matter the brand name!

Ernesto