Exposure steps with LW camera?

#4

Interesting question. That’s not so easy at it sounds, you know? If we’d want to go mad with how the real thing works, we may want to take into account shutter speed, aperture size, film speed (ISO/ASA) and film grain. In Lightwave we don’t have film speed parameter and shutter speed and aperture size don’t affect the exposure of the image. I think this is not too bad if Lightwave consider the camera as a human eye, but with the addition of the Real Lens Camera, this becomes necessary and consequent. A good starting point to emulate is V-Ray Physical Camera, I think. Would be good to have something similar within LW as an additional option for people who needs more accuracy in this regard (though this is not for eveyone and is not necessary in all cases). However we can begin to look for some guidelines to help us to have a reference for this within LW. According to LW parameters by default, shutter speed is 0.01667 (50%mblur) and aperture size is f/4, wich is about 10 EVs. If we assume global Light Intensity as it was our film speed parameter - let’s say ISO 100 or considering 100% mblur, ISO 200 (working in LCS the S-curve should be added later as post-processing), we can begin to approximate a shutter speed and an aperture size more or less consistent with the final exposure and film grain we gonna look for in post. I said more or less because this still seems arbitrary in LW (…thinking about HDRI lighting now), but it’s at least a reference for consistency. In such a case we might use global Light Intensity parameter to vary the exposure value. This must be made in log2 increments as f-stops (let’s say for underexposing 100% - 94.36% - 88.71% - 83.07%, etc or for overexposing 100% - 107.64% - 115.29% - 122.93% and so on). Though changing the exposure in post is simpler, one of the advantages of playing with this within the 3D package (Render Globals=>Light Intensity) is the high decrease of digital grain (noise) in the raw output render.

Gerardo

#5

hmm hmm hmm
I understood about half of your post. But do agree with you if we already have Real Lens camera we could use some more parameters with it. Another cool thing would be to have Raw format export for controling exposure later. Doing HDR does help, that’s for sure, but it doesn’t seem accurate to real life (at least not for me). Really good thing about HDR rendering is that you can use programs like Photomatix to tone map your renders.

Cheers!

#6

Sorry to not be clear (my English is terrible). When I say raw output render, I’m referring to a linear render just straight LW (without any gamma encoding). The S-curve and proper exposure can be added over this raw image later in post (if we save it in some FP format), but it’s better than a ‘common’ RAW file since we have a lot more dynamic range to play with.

Gerardo

#7

Mine is even worse, heh :slight_smile:

I was refering to all the tech stuff in your post.

Thanks!

#8

…Doing HDR does help, that’s for sure, but it doesn’t seem accurate to real life (at least not for me).

That’s because changing the exposure doesn’t affect any other aspects of the image. That’s more similar to changing the film speed. Exposure’s values increments are however accurate.

I was refering to all the tech stuff in your post.

Sorry for that too. Technical terminology is inevitable to keep short some posts:

EVs= Exposure Values (amount of light that hits the film)
LCS= Linear Color Space (by working in linear workflow)
S-Curve= S-shaped curve or D–logE curve (film density vs. log of exposure)
Log2= logarithm (base 2) - don’t worry about a log10 conversion, take a look at this calc

Btw, Blochi explains that terminology in a very understandable and practical way in his HDRI HandBook (very good reading for CG professionals and photographers) :slight_smile:

Seeing how LW works now, we can suppose that it has all requirements to implement a physical camera without any problem. 'Till then, people won’t be able to solve all that within LW only (in several cases is even desirable to solve all that in post). But in both cases and as you have noticed, we need observation (and if it’s possible, some understanding) about how these things work in reality to make some consistent adjustments in LW and post.

Gerardo

#9

I wrote a simple lscript that puts fstop 4.0 at 0 exposure and then exposes brighter and darker with fstop changed. It also has manual mode to raise and lower exposure in stops. I haven’t polished or finished it, but it works as a proof of concept.

#10

That’s sounds like an interesting lscript, Jezza :thumbsup: That’s a prove that is possible to have something like that as a built-in feature in LW. Do you think is possible to choose the f-stop to set up the 0 exposure (a kind of offset) and tie this with mblur amount accordingly (or vice versa)? if that’s possible (and though is still arbitrary), any change of these parameters would be even more consistent. Btw, what LW parameter are you using to vary the exposure value?

Gerardo

#11

So that’s like bracketing in LW. hmmm.
Seems very interesting. What I would like to do (that was my original idea) is to take these 3 images and combine them in something like Photomatix into HDRI to see the final result and tone mapping possibilites of it!

Any chance we see that script.

Cheers.

P.S. Jezza, sorry for not giving you any feedback on passport but I no longer work there.

#12

getting to this is on my to-do list.

#13

Inside the Image Processing tab:

Add Image Filter > HDR Exposure

Were you looking for this?

#14

What Akademus wants is to adjust images exposure globally, in steps increments. Globally means by affecting highlights and blacks in the same way (like when we change the f-stop or shutter speed of a real camera). HDRExposure works by way of an S-curve filter (similar to Curves filter in PS), so that we can expand or compress the ratio between highlights and blacks, in the limited range of our monitor space, by using whole dynamic range of an HDR image.

Gerardo

#15

Pretty much. I want to see image rendered in exposure steps like in real camera so I can combine them and get the HDRI like i would from real life photos so I can tone map them. I bet these results would be something quite interesting in CG world (or a total miss/mess)

#16

akademus, I haven’t forgotten my script, but for what you want it is extraneous. LW can render directly to HDR, saving you the steps you mention above.

#17

If it’s about to change the exposure only to tonemap, yes (Artizen and other apps can do both), but if we want to change the exposure and affect, at the same time, mblur (for the shutter speed), or simulate film speed, etc. A plugin or script would be great :slight_smile:

Gerardo

#18

I was merely wondering will the results with such processing give different (or probably more realistic) results than tone mapped HDRI directly exported from LW FP renderer. That was my goal.

#19

We can make a kind of logarithmic tonemapping within LW, or by using the local operator method, or both.

Gerardo

#20

Well, the auto-mode one is unlikely to happen due to time constraints. So here’s a simple one with manual stops control.

http://jezza.lwidof.net/cgtalk/fstopExpose_IF.zip

#21

Hey! thanks, Jeremy. I guess the same thing would be faster as a plugin compilation, right?

Gerardo

#22

Yep, you betcha. I’ll post here if I ever make it compiled.

#23

Just to report the script is uncalibrated 1/2 f-stop.

Gerardo