Photometrics / exposure control?

[RevitGary]

I have a good idea about how to use Photometric lights etc. ( Just started I know I am missing something. )

 

My question is...

 

If photometric lights are used to create a realistic lighted environment and you use photographic exposure control. How do you know the lights you are specifying will light the area properly.

 

I can put a 100 watt bulb in a big room, if I crank up the exposure it will look like a nicely lit room. We know this is not a realistic view of the room with a 100 watt bulb.

 

How do you render a room with photographic exposure control and verify the lighting you are specifying is appropriate for the room in a real world situation?

[luckymutt]

Exposure control is only tone-mapping.

Photometric lights in Max will calculate correctly, given that the model is to scale and the materials are made with real values in terms of reflectance, etc.

Exposure control will just tone-map the image in the frame buffer to control the light levels.

 

In your example of your room with a 100 watt bulb, by cranking it up it is indeed not a "realistic" view. However, if in a real room lit this way, if you used a camera and set the exposure the same way, you will wind up with an over-exposed image (at least compared to what you actually see) that is also not a realistic view.

 

Photographic exposure control was designed to tone-map the frame buffer in a manner that is consistent with actual film cameras. In fact, Zap Anderson (the guy at mental images who wrote the code for the exposure control(and the A&D materials, the SSS shaders, and a ton of other cool things))actually used formulas from Kodak when he put this together in order to create the accurate behaviour and translation to pixel values.

 

Now going back to your room with a 100 watt bulb, set the exposure value to something like 10, and you will be closer to what you might actually expect to see in the real world or a real photo.

Edited March 14, 2009 by luckymutt
sepinllg

[luckymutt]

Also, if you want to see the effect of your lighting in terms of real world values, you can set the exposure control to "Pseudo Color." This will show you the behaviour of your lights (luminace or illuminance) in false color as cd/m^2 and give you a clear idea of the distribution in your scene.

Edited March 14, 2009 by luckymutt

[AJLynn]

Cranking up exposure control is not at all unrealistic. It's how you photograph indoors with artificial light and no flash.

 

I just took a picture of my bookcase with my digital camera. I set it to ISO 1600 and f1.8 and it metered at 1/80 of a second.

 

If I took a shot of the same bookcase only I put it outside in direct sunlight and used the sunny 16 rule my f-stop would be 6.3 stops slower, my ISO would be 4 stops slower and my shutter speed would be .5 stops faster for a total difference of almost 11 EV. Both shots are correct, and both lighting situations appear normal to the eye, which has remarkable exposure control.

[luckymutt]

Man, I sometimes hate the word "realistic"

I took Gary's use of the word to mean something like: "how well lit a large room would appear to be lit by a simple 100 watt bulb as we are used to seeing it"

 

We can place your bookcase in the corner of a 16'x16'x12' room that is only lit by a single candle in the far corner. By adjusting the exposure we can get a nice "realistic" look to the image, however, it would be "unrealistic" in that we would not normally percieve a single candle to throw off that much light.

 

All of these examples go to show that exposure control is just tone-mapping the image.

In fact, you can turn exposure control entirely off, save out a floating point image, and adjust the exposure in post and get the same range of results for your desired look.

 

So to Gary's questions: lighting analysis tools will tell you the real world values of how your lights are affecting a space

 

Once the lighting is appropriately set, use exposure control to, well... expose the image to taste.

[AJLynn]

My point is that while using exposure control to make your lighting brighter is considered by many people to be cheating, incorrect or unrealistic, it's not. The question is, a realistic what? A realistic "properly" exposed photo, a photo underexposed for a particular reason, or a "what the eye sees" sort of thing. Which isn't really possible since you can't duplicate the eye's dynamic range compensation or the combination of foveal and peripheral vision.

 

Whatever image you create, including the one your mind generates as your eyes move around a room lit by one light bulb, is a representation, not reality. It's your choice of what to do with it. Taking a spot meter off an 18% gray card, using a wide angle lens and taking a timed exposure on a tripod to put your exposure within the characteristic range of the film, then burning or dodging the print for finish, is entirely correct and has been done for decades. So why is using a wide angle lens in Max and jacking up exposure control to get the brightness that looks about right then Photoshopping it to get it just right in print wrong?

 

By the way, the eye also does exposure compensation, and can get it wrong - the eye's aperture does not adjust instantly to drastically changed lighting conditions.

[RevitGary]

Thanks to all.

 

The point is ... I want to use max to verify that the lights be specified in the room will light the room properly. And I want a rendering that will accurately represent the lighting in a room as if I was standing in the room, without a camera.

 

 

"Now going back to your room with a 100 watt bulb, set the exposure value to something like 10, and you will be closer to what you might actually expect to see in the real world or a real photo."

 

1. Does this hold true for a room that is lit with rows of fluorescence? Is 10 a good interior setting in general or a good setting for a 100 watt bulb?

 

 

"took Gary's use of the word to mean something like: "how well lit a large room would appear to be lit by a simple 100 watt bulb as we are used to seeing it"

 

2. Yes this is what I meant exactly

 

 

" lighting analysis tools will tell you the real world values of how your lights are affecting a space"

 

3. ok Ill start looking for these .. any pointers on where to find good information, books or tutorials on this?

[AJLynn]

I'm not just saying this because I'm a theorist:

 

It is impossible to make a representation of the appearance of a space "without a camera". Cogito ergo sum.

 

But yeah, what you really want is lighting analysis, e.g. pseudocolor exposure or whatever Max is using now. You can use a standards book for the required illumination for various spaces. Also, with appropriate energy modeling you'll be able to calculate "daylight factors" for LEED use. Max sort of does this but isn't perfect.

[luckymutt]

Is 10 a good interior setting in general or a good setting for a 100 watt bulb?

No.

"10" was just arbitrary to show that no matter how the light is actually calculated, exposure control can raise or lower the levels of the lighting to make an aesthetically pleasing image based on the light distribution. It is all scene dependent and determined by artistic sensabilities. (arch-viz is a creative effort, after all)

 

Lets take the extreme example of AJLynn's bookcase in my large room lit by a single candle...

using Pseudo Exposure control, the image turns out completely blue (luminance from the candlelight at those distances results in (damn near)0.0 cd/m^2 on the bookcase)

However, by using exposure controls, we can raise the lighting levels across the scene, in effect raising the relative luminance value intensities, to make a nice looking image.

...pseudocolor exposure or whatever Max is using now...with appropriate energy modeling you'll be able to calculate "daylight factors" for LEED use. Max sort of does this but isn't perfect.

Besides the "Pseudo color" exposure, there is also the seperate "Lighting Analysis" tool in Max 2009 Design.

Neither seemed perfect to me either, however from what I have read about the upcoming version of Max, it will include a newer tool that can cert for LEED.

 

As far as creating a rendering that looks as though you are standing in the room without a camera...not really possible in a measurable way...that's where the artistic sensibilities come into play with exposure control. There are variations in the way that any one person percieves light than another, so just expose to what you feel as "proper," I suppose. Make it sexy. If anyone disagrees and tries to call you out on it, show them the analysis results If they insist it should be brighter/darker, adjust the exposure in post (you did render out to a floating point image, didn't you? Good. Now you don't need to re-render)

[AJLynn]

All true but still not quite there on the theory. Exposure control does not raise or lower the levels of the lighting - it's not a lighting multiplier - it is a property of the camera that raises and lowers the camera's sensitivity to the lighting that's been put in the scene. (Even if in mental ray they put it in the wrong place so it looks like a property of the scene.)

 

This is fundamental to the theory argument - it is impossible to make something that "looks as though you are standing in the room without a camera" because it is impossible for you to know what something looks like without use of a camera. Anything you can use to represent the image in a way that is comprehensible to a human relying on vision is conceptually homeomorphic to a camera, including but not limited to an actual camera, a human eye or in some respects the setup you do when you hand draft a perspective. In all cases the information received is interpreted, filtered and converted into a useful format, making the amount of light useful only inasmuch as it is necessary that it be within the range that the film/sensor/rods and cones can interpret.

 

Everything you see or photograph is always subject to tone mapping and exposure control.

 

This is why I quote Descartes - it's pretty much the same argument.

[luckymutt]

All true but still not quite there on the theory. Exposure control does not raise or lower the levels of the lighting - it's not a lighting multiplier - it is a property of the camera that raises and lowers the camera's sensitivity to the lighting that's been put in the scene. (Even if in mental ray they put it in the wrong place so it looks like a property of the scene.)

I am not saying that exposure control is a lighting multiplier. I've said from the beginning it is about tone-mapping (and being such, you do not even need to apply it within Max).

I am using the term "levels" in it's tone-mapping use. Simliar to "gain." Or "curves." Or "offset." Or "exposure."

 

Not quite sure I see how applicable Descartes' ponderings of his existance leading to confirming his very existence is in this discussion, but...ok, man.

Personally, I prefer:

Sum ergo cogito.