Reinhard Burn Value Confusion

[jervistuazon]

Hi guys, hope you could help me. Im having a problem understanding how burn value works and what is the best method to fix this.

 

I have here two images, the first one is with a default burn value of 1. It looks nice overall but the white part in the center is overexposed.

 

On the 2nd image ive used a burn value of .35, the result is undexposed and desaturated.

 

So im quite confused what would be the proper burn value to used.

 

Do i have to increase exposure in the 2nd image to have a brighter image?

 

I use a vray universal settings with camera f8, 200 shutter speed, 200 iso.

 

[ATTACH=CONFIG]54893[/ATTACH]

Hope you could assist me thankyou very much.

[Francisco Penaloza]

Reinhards is a very flexible way to tonemap an image within V Ray but is not perfect. one of the drawback is it desaturate the image and loss of contrast while diminishing burn areas.

This can be easily corrected in Photoshop adding a curve contrast to your image. Otherwise you could do the same in V Ray render buffer.

If you want full control of your image, you need to render linear, save it a full float (32bits exr or HDRI) and tonemap in post.

[Corey Beaulieu]

Reinhard is a blend between Exponential and Linear Multiply. At a burn value of 0.0 it uses full Exponential which is a mode that saturates as brightness approaches 1.0 in any 32 bit color channel. At a burn value of 1.0 Linear is used which just keeps adding up the brightness without adjusting. The perceived effect should be duller in contrast at lower burn values, but more saturated in the bright areas and linear more contrast and sharper bright areas.

 

Maybe the best description I have heard in that 0.0 works more like the human eye and 1.0 works more like a camera. I would submit that while both are fair game, as is any number in between, you should just stick to Linear, a burn value of 1.0. Don't clamp your colors and then you can use 32 bit compositing to re-balance any areas that are out of range.

[jervistuazon]

Ive tried using curve in 0.35 burn value but i didnt like the result compared to linear. Thankyou for your help and tips, ill stick to linear and adjust it in post process. Thankyou.

[jervistuazon]

Thankyou for the deatiled explanation, i guess linear is the best way to go, ill search on the net how to adjust properly oversatured areas. Thankyou again for your help.

[jervistuazon]

I've read on the internet that turning this option doesn't produce a physically correct render, but i think this is the simplest method to avoid overexposed areas. Thankyou for the suggestion.

[Corey Beaulieu]

Clamping does not "avoid" overexposed areas. It simply replaces an actual value above your clamped value with a value equal to your clamped value For example if your highlight is 1.35 in the Red Channel, it will be recorded as a 1.0 and the actual information will be lost. 32-Bit can handle working with the values above 1.0 and correct these back into range so I don't see why you would want to clamp it.

 

If you want to avoid overexposed areas then adjust your lighting and/or camera.

[jervistuazon]

Thankyou for the input, may you recommend a workflow on how to work an overexposed area.

[Corey Beaulieu]

This, the above, is not good advice. Shoot for mediocrity he says and leave the rest to those with time and money.... Boo.

 

Light, as in all things 3D has/needs a sense of scale. The camera is that sensor to lights scale.

 

The camera is setup to mimic real world functionality for many reasons not the least of which is that Vray has provided us with a Sun and Sky model that more or less mimics reality and the same exposure triangle that drives us in reality should and does drive us in 3D.

 

While not all lighting scenarios are setup to use the Environment as the primary light source consider the ones that do and begin by establishing an exposure (using the F-Stop, Shutter Speed, and ISO settings) that keeps the brightest area in range (0.0-1.0). I find it easiest to use a white-ish override and then turn off "Force Clamping" in the Vray Frame Buffer. This will tell you once you start going over 1.0 and where. A spot or two may be okay, but an large area is not.

 

Your initial image will be over contrast-y and have some overly dark areas. From here you just continue to add you artificial lights in order to fill out the scene with light. You don't want to over light a scene as it will flatten things out, but you do want to create a hierarchy of strength while keeping most if not all between the values of 0.0 and 1.0. On your artificial lights it is wise to use the the direction feature in the light objects.

 

It takes time at first, but realistically, a few times through, you will be able to light a scene with very nice precision/balance in just and hour or two. The more you can mimic real world things the better as you can then compare your scenario to reality and make educated decisions on how to improve.

 

One thing to remember with light is that is dies over distance. This means that a light close to a wall will create a burning hotspot and fade quickly. A light farther away will fade far less quickly, but will be higher in power and less likely to burn.

 

There is so much here. It's a study of real versus art/passion (how we desire to shape the real), but the more you do to relate to the real, the more your scenes will become realistic and lighting and cameras are at the top of the priority list.

 

If you don't have a "guru" here is perhaps the best one; http://www.learnvray.com/

 

Sorry if this is unclear. Your question is innocently crazy-complicated. Good Luck and ignore spacecase.

[jervistuazon]

Thankyou for the tips and suggestions, I appreciate them all it helps me to become flexible and understand vray. I also aspire to have raw render which are nice and almost no post process needed like from evermotion tutorials but in my current situation I dont have the hardware muscle and flexible time and got to make due of all possible work arounds.