Measuring and visualizing light falloff in the frame?

I am trying to compare light falloff patterns and amounts across the frame between various lenses and softboxes. I am hoping to get falloff value in f-stops and IREs across the frame and also visualization of it. What is the best tool / way to do this, please?

A light meter. Unless you mean in post.

A light meter. Unless you mean in post.

I meant in post from an actual recording frame. Light meter gives me a rough idea for softbox's falloff but won't tell me anything about lens, what happened in frame once lens, softbox, and everything else combined in frame. Nor can I use it for shots that have already been taken.

Read this free guide:

"Using Waveform Monitors as Artistic Tools for Color Grading" by Steve Hullfish
https://download.tek.com/document/2PW_28619_0_HR.pdf

It will tell you everything you possibly need about using scopes for color. I've been doing this for more than 40 years, and even I learned a few things from this article. It's possible to use scopes as a "sort of" light meter, particularly in terms of key-to-fill ratio and so on.

Read this free guide:

"Using Waveform Monitors as Artistic Tools for Color Grading" by Steve Hullfish
https://download.tek.com/document/2PW_28619_0_HR.pdf

It will tell you everything you possibly need about using scopes for color. I've been doing this for more than 40 years, and even I learned a few things from this article. It's possible to use scopes as a "sort of" light meter, particularly in terms of key-to-fill ratio and so on.

Thank you! But how I can use that to measure and visualize light falloff in the recorded frame? I guess I am looking for "Cliff's notes version / approach for dummies".

I'm hoping to get something like this:

Are you going to do this for every stop? Lens fall off usually improves the more you stop down. From memory it'll also depend on focussing distance. The closer you focus the less fall off.

Unless you're doing reproduction work for fine art or something you're going down a rabbit hole.

You can easily generate an aggressive grade for that, plus the False Color effect.

Are you going to do this for every stop? Lens fall off usually improves the more you stop down. From memory it'll also depend on focussing distance. The closer you focus the less fall off.

Unless you're doing reproduction work for fine art or something you're going down a rabbit hole.

No, I'm not going down that deep, I just need to compare certain setups now and again in the future. Like, for example, what happened with illumination across frame if/when I changed/moved softbox.

You can easily generate an aggressive grade for that, plus the False Color effect.

Good idea, using False Color didn't cross my mind, thank you What you would suggest to do to generate an aggressive grade, please?

If you're just worried about moving the box you don't need to go to all that much trouble. We're back to the light meter.

If your camera supports spot metering that'll be good enough. Put a long lens on it to have the smallest possible spot.

I think he wants to experiment with lenses too, Nick.

By aggressive grade I meant increasing contrast. In your case, note the amount for future tests. Then make a 'creative' version of False Color or try one of those already present in B&W steps, they might already be what you want.

If you're just worried about moving the box you don't need to go to all that much trouble. We're back to the light meter.

I'm not worried just about box, I used that just as an example. With light meter I don't know how much lens adds on top of that, I can't quickly determine where in the frame I am, and would need many measurements (more time) to get some kind of visual interpretation.

If your camera supports spot metering that'll be good enough.

That is a good suggestion to get a quick rough idea but would also require lots of time to get approximate visual interpretation of falloff. If I just divided frame in 5 spots vertically and 5 spots horizontally I would need to take 25 measurements, write them down, and plot them later.

By aggressive grade I meant increasing contrast.

Thank you, I will give that a try.

The waveform monitor will show the falloff but you will need to work out what a camera stop is equivalent to when using the scope scale on the left. There are a few choices of scale available.
I would do a bracketed exposure test using different f stops to determine where the stop settings line up on the scope scale, then use that as a reference for your light falloff setup.

Screenshot 2023-06-03 at 4.49.38 pm.png (115.47 KiB) Viewed 918 times

The waveform monitor will show the falloff ...

Waveform does show falloff but it is 2D one, left to right across the frame, not 3D one as you look at frame in front of you.

Convert the data so it shows luminance, then convert to pure log representation and apply quantization/posterization using false color or something similar. Since pure log means each stop takes same sized range in data, it becomes a matter of figuring out how big that range is on that specific encoding. Should be relatively easy to figure it out, for example by introducing +1 exposure change before log conversion and comparing the values. Then use this range scale in posterization as level width. Once you get this thresholded result you can bend and stretch it as you like to visualize the result.

Convert the data so it shows luminance, then convert to pure log representation and apply quantization/posterization using false color or something similar.

Thank you! This sounds like most accurate way to go about. Please forgive me for needing, as an absolute illiterate, too much handholding but how one converts the data to show luminance and then converts to pure log representation?

To correlate stops and IRE you can create a synthetic middle grey within the camera color space/timeline then use the HDR Wheels global exposure tool to check the offset from the midpoint.

I'm new to Red so I wanted to see how data was allocated within Log3G10. This synthetic test puts mid grey at 33% with 11 stops +/-. Log3G10 is allegedly 10 stops over with mid grey at 33%. The test confirms, 10 over, 8-ish under.

All you'd need to do is create a mid grey square and subtract low exposure value from the high exposure value to calculate falloff. You couldn't do both at once. You'd need to calculate the lens falloff from the center of the light beam and calculate the light's falloff at an aperture/focal distance where the lens had the least falloff.

Good Luck

Log3G10.jpg (181.92 KiB) Viewed 782 times

To correlate stops and IRE you can create a synthetic middle grey within the camera color space/timeline then use the HDR Wheels global exposure tool to check the offset from the midpoint.

Thank you!

The waveform monitor will show the falloff ...

Waveform does show falloff but it is 2D one, left to right across the frame, not 3D one as you look at frame in front of you.

The left to right is the horizontal falloff and the vertical thickness of the trace is the vertical falloff.

The waveform monitor will show the falloff ...

Waveform does show falloff but it is 2D one, left to right across the frame, not 3D one as you look at frame in front of you.

The left to right is the horizontal falloff and the vertical thickness of the trace is the vertical falloff.

Thank you for teaching me I will give it a try

Thank you! But how I can use that to measure and visualize light falloff in the recorded frame? I guess I am looking for "Cliff's notes version / approach for dummies".

Did you read it? Scopes are enormously useful for frame-by-frame color/luma matching problems. The standard advice in the 1970s and 1980s was: "match the scopes exactly, and the picture will generally be matched as well." This is especially true for exposure variations. Best case, you can reduce the problem significantly, to the point where it's practically unnoticeable.

Peter Cave's scope shots above illustrate the problem perfectly: make the line FLAT, and there should be no visible exposure change. I've dealt with this kind of thing in car beauty shots against white seamless backgrounds, and even when they're going to use VFX to make the white background perfect, I try to smooth it out in advance to help reduce the time and effort for the composites done by other people.

BTW, if the camera is moving through 3-dimensionsal space -- like a dolly or crane move -- this will be exponentially more difficult.

If the scopes don't serve you, try this. I made a quick test with a Zeiss C/Y 18mm f4. A great lens, but with considerable vignetting. Shot full sensor size on a Sony A7 IV at f4 (left) and f8 (where vignetting is gone even for a discerning viewer). Contrast increased to 1.8, Offset down to 12, saturation 0. False Color 10 Bands, set to display percentage:

Falloff_measured.jpg (125.76 KiB) Viewed 519 times