selmslie Loc: Fernandina Beach, FL, USA

You can't explain it and you really don't understand the topic at all. You can't even get past step #1.

As we all know by now, once you start to repeat yourself you have run out of ideas. You have lost the argument.

Ysarex Loc: St. Louis

All I need to know about step #1 is that your instructions are wrong. And I assume you followed your own wrong instructions and did it wrong. That says it all.

Doing it right: "For people that buy my books on the pro and prosumer DSLRs, I used to supply a UniWB balance file.... Unfortunately, every camera needs a different file (and in some cases, every firmware revision, too)..." -- Thom Hogan.

And doing it wrong: "The easiest way to set your camera up is to pick a custom white balance based on the display of the UniWB target below, preferably on a calibrated monitor but that's not critical." -- Scotty https://www.uglyhedgehog.com/t-692202-1.html#12141373

selmslie Loc: Fernandina Beach, FL, USA

What says it all is that you can’t even understand anybody’s instructions and do it with your own camera. It’s beyond your comprehension.

Come back when you have something new and constructive to add to the conversation.

Stop repeating yourself. You are making a fool of yourself.

Ysarex Loc: St. Louis

The photo below was taken before you created this thread in which you do it wrong and give wrong instructions. Guess you're just wrong about everything.

That's three strikes right?
Wrong instructions -- strike one!
You followed your wrong instructions -- strike two!
Wrong about me -- strike three!

You're out.

(Download)

selmslie Loc: Fernandina Beach, FL, USA

Maybe you managed to sett a UniWB based on one of the sets of instructions (which one?).

But how do you know your UniWB setting is actually effective? Does it give you anything constructive to add to the discussion about #2 and #3?

What does it do to the camera's JPEG histograms?

Do you think that UniWB is useful for determining exposure? If so, tell us all about it. If you can't then you have missed the point.

Do you have the remotest idea what is going on in these two examples: https://www.uglyhedgehog.com/t-692202-1.html#12141805 and https://www.uglyhedgehog.com/t-692202-1.html#12142228 RawDigger is looking directly at the raw file so there was no cheating on my part. If you are really smart you should be able to figure it out. Otherwise you are not as smart as you think you are.

No, I'm not wrong about you.

Ysarex Loc: St. Louis

Looks like I did.

Well, not your incorrect one.

You can test it. But first you have to actually set one.

selmslie Loc: Fernandina Beach, FL, USA

Why do you think it is incorrect?

Is it because it's different from the one that the other methods came up with? Do they agree with each other? Do you really know that the basis for their theories is correct?

I actually did set one. You just don't understand what it does and how it works.

Since you don't understand the two samples I presented above (https://www.uglyhedgehog.com/t-692202-1.html#12141805 and https://www.uglyhedgehog.com/t-692202-1.html#12142228) you have no way of knowing why my method might actually be more correct than the others. Study my exhibits some more. If you still can't figure them out I will explain them to you. But you will have to actually use your rational processes, not the internet.

But set #1 aside. How about #2 and #3? Do you have anything to say about them? They are the real reason for this thread.

Have you ever heard of Occam's razor?

Ysarex Loc: St. Louis

Already answered.

DaveyDitzer Loc: Western PA

Sad. I was hoping to learn something in this track, especially since I shoot Df. Then the schoolyard shouting match began and here we are.

selmslie Loc: Fernandina Beach, FL, USA

I can appreciate our frustration.

My purpose in starting this topic is to show that UniWB is a really useless concept.

There are much easier ways to arrive at a proper exposure, especially if you record the raw file and develop the image on your computer:

1. No metering at all - In broad daylight Sunny 16 (1/ISO seconds at f/16 or equivalent) works fine and it's easy to guess the increased the exposure if the clouds roll in.
2. There are many scenarios where the meter will not work. You are better off with an educated guess. See the Exposure value article.
3. Chimping - If you can make a reasonable guess at the correct exposure and the image looks reasonable you probably don't need to worry.
4. Matrix metering - The camera can usually arrive at a reasonable exposure.
5. Watching for blinkies - Highlight warnings are a simple indication that you are in danger of raw overexposure.
6. Watching the histogram(s) - They are a good indication and they can warn you if you are too close to the right edge.
7. Live luminance histograms and Zebra warnings can tell you if you are on the right track before you trip the shutter.
8. Bracketing - When all else fails and you can't arrive at the right exposure there is no shame in taking a series of shots and working the one that looks best.

I have never seen any example of a scene where UniWB provided any benefits over using a normal white balance so it doesn't really matter whether it is easy or difficult to use.

Ysarex Loc: St. Louis

But instead you gave out incorrect instructions for how to set up UniWB and followed them yourself.

Ysarex Loc: St. Louis

Read the article by Thom Hogan that quixdraw posted: https://www.dslrbodies.com/cameras/nikon-and-dslr-camera-faq/what-is-uniwb.html

UniWB is used by raw only shooters to add as much precision as possible to the task of setting an exposure. Odds are they're going through the trouble because they're interested in exposing the camera sensor to it's full capacity -- a risky business since clipped highlights are forever. Why bother? They probably was maximum DR in their captures. So they want to get as close to the sensor's clipping threshold as they can.

UniWB is another tool then to help improve the odds and get a precise result. What it will do is alter the camera's histogram display and/or the camera's highlight clipping warning. Both of those exposure aids are typically calculated by the camera from the processed JPEG image and not from the raw data. As Thom Hogan explains in that article white balance coefficients have already been applied to that processed data and will influence the result.

What the UniWB user discovers is the camera histogram and/or highlight clipping warning without the WB coefficients applied can be more accurate indicators of the raw file exposure. You really have to not want anything to do with the camera JPEG. Setting UniWB trashes the camera JPEG.

The value of setting and using UniWB is highly camera dependent. It's going to be more or less useful depending on your specific camera model. For example your Nikon DF: The DF is a DSLR and you can't get a histogram or highlight warnings live. You have to take the photo and chimp it. Different cameras provide better/worse histogram displays on those dinky little LCD screens. If they're barely readable what good are they. Some cameras only provide a luminosity histogram.

The highlight clipping warning may be more useful but again there are specific camera model issues. Does your camera start the clipping warning if one channel is clipped or (as many do) does it wait for at least two clipped channels? Your DF will chimp highlight clipping and that might be useful in conjunction with UniWB, but like any exposure aid you'll need to do careful testing.

First step is to set an accurate UniWB as a custom WB on your camera. The red, blue WB coefficients read out in the EXIF data and you want to make sure they're both 1.0xxxx.

selmslie Loc: Fernandina Beach, FL, USA

That would be great if it were true. But the devil is in the details.

I came up with a quick preset without using the sample image. The coefficients are not as close to 1.0 as you might wish but they are a lot closer than they were before.



We are primarily concerned about what happens to the raw file as you approach its upper limit and how well the JPEG will track it. What happens at lower levels is irrelevant since the JPEG uses different gamma values throughout its range.

I plotted the JPEG green values and the log(raw,2) values adjusted to match the scale of 0-255. The exposures were made in 1/3 stop increments from EC+1 through EC+3.3 where I knew that the raw file would reach its limit.



As you can see, the JPEG distribution for UniWB is similar to the Daylight WB plot but it's shifted up throughout most of its range - uniformly until it nears the end.

What might be surprising is that the Daylight version reaches it's upper limit of 255 at about the same time as the raw file reaches its maximum which is actually about 0.1 stops below the theoretical maximum of 16383. The UniWB version, as it did with the preset based on the magenta target, is premature by about 1/3 stop in predicting the upper limit.

So the question of whether the UniWB, in either form, is more accurate or correct than Daylight WB is in doubt.

Ysarex Loc: St. Louis

Got it. You still can't set UniWB on your camera.

selmslie Loc: Fernandina Beach, FL, USA

I don’t need to. The results would be about the same. They might be even worse if the green channel moves further away from the Daylight version.

UniWB remains a lot of effort with no payback.