Jump to content

Help! Large image comes out really bright!


3dpatient2
 Share

Recommended Posts

I'm still somewhat new to vray and I'm having the following problem: I'm rendering a nighttime scene with about 6 lights in it and a few vray material wrappers with high outputs to light my scene. My problem is that at low resolutions my image comes out fine. When I push the image size above about 2000 x 1000 pixels however, everything starts to get really bright.

 

I'm using GI with an irradiance map for the first bounce and light cache for the second. I set the irradiance map settings to "medium" (this has worked for large images for me in the past). Also using linear color mapping.

 

Has anyone had this problem before? Please help!

Link to comment
Share on other sites

Right. Thanks for the tip.

 

But even if I use exponential mapping would the result differ according to the size of the render? That's the crux of the issue here: I render two images with exactly the same settings except for the image size and I get two wildly different results.

Link to comment
Share on other sites

The resolution influences the amount of rays emitted in each pixel of the scene - of course, if you make the scene larger, you ought to largen the amount of subdivs in irmap settings.

Also, try practicing. To be honest, never rendered such a large-res scene, but with 1280 everything works just fine

Link to comment
Share on other sites

Okay, but even if you didn't increase the subdivs why would the image get brighter?

 

Maybe my confusion is coming from the fact that I have practiced this before: with daylight renderings where I'm using just one direct light and an environment map image. When I render those at low res and then again at high res, I have no problem.

 

The only real x factors I've got here are the number of lights and the vray material wrappers. Those are the only things that are different from what I normally do.

 

Also, there isn't a sliding scale here: I get the same result if I render something 300x100 as I do if I render something 2000 x 700. It's only after that 2000 pixel mark that I'm getting problems.

Link to comment
Share on other sites

If i use linear - that's only when i need unrealistic results in animations.

When u increase the resolution, u increase the number of pixels. The more pixels there are with the same amount of subdivs, the less subdivs there are per each pixel of the image. Therefore, bugs occur.

Link to comment
Share on other sites

My friend, nobody uses linear color mapping! Try exponential! Linear is unreal, does not bring photorealistic results, which are crucial in archrendering.

 

Linear mapping is critical and the most useful, as long as you use an sRGB LUT on your frame buffer. Using exponential will burn in your LUT into the image making it much more challenging to comp.

 

I use Linear exclusively.... with the proper EXR output and comping, it is the only way.

Link to comment
Share on other sites

whoo hoo.... is any one else glad chris adds his knowledge to this site???

I was starting to second guess my entire workflow, till I saw his post.

 

So call me nobody ;) I also use linear exclusively..... with gamma correction to and exr of course.

 

 

anyhoo..... back to the first question..... what objects have the material wrappers for lighting the scene? I've never taken that approach, I rely solely on vrayLights, ies, hdri, and now the vraysky for lighting scenes.... I use the wrapper more as a control object.

 

if you remove those objects from the scene do you still get the blow outs? I would consider a different lighting method if I ran into your problem.

 

good luck to ya.

Link to comment
Share on other sites

I was reading about how Cinema4D handles rendered gamma. It renders at 1. I tried rendering a test without colormapping to 32bit, and the result looked washed out in Photoshop. But leveling it out gave the picture a very nice exposure. I then tried colormapping HSV/exp to 8 and 16 bit, and the result matched the 32bit image after correction and conversion to 8 or 16bit. What I took from that was that its good to have the engine process the colormapping in its internal process, and then a 16bit or even 8bit output will be in good shape.

 

Chris--when you talk about burning in the LUT (lookup table?) what do you mean? Wouldn't that be like biasing a file to a certain profile? Or is that the point--to force the display curve onto your image?

Link to comment
Share on other sites

Hey, everyone! I was the one who originally misfamed linear color mapping. I really did not know all that stuff about correcting the colours then, after the render, using the LUT - what's that exactly, if u do not mind? Couple of links would be useful please, since i am absolutely out of that source of knowledge, unfortunately...

Why i wrote that linear brings unrealistic results is because all visualiserz, whom i know, use the exponential, that is why. Sorry, for misguiding!..

Link to comment
Share on other sites

Chris--when you talk about burning in the LUT (lookup table?) what do you mean? Wouldn't that be like biasing a file to a certain profile? Or is that the point--to force the display curve onto your image?

 

LUT mean look-up table. Basically it is the profile of your monitor Ideally what you want to do is calibrate your monitor to a standard profile so that everyone can see the image correctly. The easiest one to do is sRGB... that way if everyone is calibrated to sRGB then everyone will see the image exactly the same (ideally). Anyway. The data of the image (in terms of light information in floating point) does not correspond to what we actually see. Which is why you need a look-up table so that the linear data can be seen correctly... what you want is the data to stay linear but the monitor to display it correctly. That is why Max has that gamma setting in it, and why Vray now has that "sRGB" button in it. By using the exponential color mapping, what is being done is to force all the data to have an exponential curve in it. Your image will "look" correct but will not BE correct (as far as light data is concerned). Sorta like taking an image and doing a levels on it of 2.2... then saving your image... you will never get that back correctly. That is what I mean by burning in the LUT. Now if you never plan on editing your image ever again, and plan on saving an 8 bit image... well I guess... who cares? But I suggest the rendering out of 3D to only be the first step.

 

Now there is a lot more information about this, and it has be discussed at nausium... so if you are interested in reading more... Look up linear workflow (or LWF) in Vray on google. I am sure you will get plenty of results.

Link to comment
Share on other sites

Thanx, guys, very much! I'm almost a year in vray already, but never heard of what i have heard here! =) Thank you very much! I've googled "linear workflow" and took the first article there was. It's very detailed and nice. I've tried it out for myself - that's awesome! Thanx once again!

Once again i beg your pardom for disinformation!...

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...