[LensBeginner]

I'm trying to get a more solid understanding of how the histogram works.
Reading here
Expose Right
it would seem that the span of the histogram is about five stops, but DR of the average DSLR is about 12EVs (which should amount to the same thing).
Has it something to do with linear vs logarithmic representation?

When metering a scene, I sometimes "get a feel for it" by metering one or two points on the subject, then one or two in the highest light I want to retain, and one or two in the deepest shadow.
Is there a way to calculate exactly, and not just by feel, how to compress all this data, eg. metering the top highlight and then underexposing by the "distance" from 18% gray and the right side of the histogram (which should be between 2.5 and 3EVs, but why?)?
Thanks in advance.

[Na Horuk]

I'm not sure. I think the problem with the histogram is that it is poorly labelled, and that it is not representative of the raw file, but of the jpeg (which is compressed, has smaller bit depth). So if you shoot raw, you can sometimes recover things you cannot recover in the jpeg.
And I don't think there is any way to automate that kind of metering. You can try doing things that are approximately right, like EV+/- and highlight correction. I don't think you can ever get that mathematically perfect ETTR. But on the other hand, you don't need to. From what I have seen, ETTR can be great, but it has its problems as well, like purple fringing more obvious, oddly coloured highlights, posterization due to lack of data in the extreme highlights..

Also, I remember some time ago there was a big debate over this, and that Pentax should improve the in-camera histogram so that it would accurately represent the raw, by colour channels, with precise labeling of the cutoff values

[LensBeginner]

Last post I made got lost somehow... oh well...
On mobile presently, so short version.

I'm looking into rawdigger trial presently and UniWB (but not fond of the fact that you basically lose an usable ooc jpeg...)
I'd like to find a scientific, reproducible way to do this, I believe it should be possible...

If not possible with histogram, then it should be doable with spot metering only, after all it spits out an exposure that aims for a fixed luminance value ("middle" gray, even though that varies between cameras)

Also, how do you label a given exposure with an absolute value?
EV values ore only relative, not absolute if I understood it correctly.

[tuco]

... but DR of the average DSLR is about 12EVs

That is Engineering DR (SNR 1:1) not photographer's DR. In reality most people do not use the bottom 2 or 3 stops. It's full of noise and color shifts.

[wildman]

Is there a way to calculate exactly, and
not just by feel, how to compress all this data, eg. metering the top
highlight and then underexposing by the "distance" from 18% gray and
the right side of the histogram (which should be between 2.5 and 3EVs,
but why?)? Thanks in advance.

I think you may be over-thinking the problem.

First, assuming you are shooting RAW and are competent at processing
RAW, you are not trying for a perfect RAW exposure but rather an
optimal exposure. By "optimal" I simply mean capturing the
maximum amount of data possible within the limits of the sensor. As a
practical matter, due to human perception, this usually means metering
so you bring the highlights just to the point of burning and let
everything below that fall where it may.

After that everything else is processing the RAW files to normalize the
tonal range to the point where you have a perceptually accurate final file.

This, of course, is a practical answer not a theoretical one.
I leave theory to others.

Take a look at this - it will explain things better than anything I can say.

https://www.pentaxforums.com/forums/32-digital-processing-software-printing/2...ml#post3122334

[LensBeginner]

I think you may be over-thinking the problem.

First, assuming you are shooting RAW and are competent at processing
RAW, you are not trying for a perfect RAW exposure but rather an
optimal exposure. By "optimal" I simply mean capturing the
maximum amount of data possible withing the limits of the sensor. As a
practical matter, due to human perception, thus usually means metering
so you bring the highlights just to the point of burning and let
everything below that fall where it may.

After that everything else is processing the RAW files to normalize the
tonal range to the point where you have a perceptually accurate final file.

This, of course, is a practical answer not a theoretical one.
I leave theory to others.

Take a look at this - it will explain things better than anything I can say.

https://www.pentaxforums.com/forums/32-digital-processing-software-printing/2...ml#post3122334

Thats what I do, but I'm looking for more dependable ways to expose properly for the highlights, especially in M mode (histogram refers to wide open so it's no use).
I'd rather take less shots and ge it right or about right the first time.
Can't trust your eyes to find a "middle gray" item and meter it... esp in high dr scenes like a sunset/panorama...
It can be helpful when the GF is pressing you to move on...

[SpecialK]

I'd rather take less shots and ge it right or about right the first time.
Can't trust your eyes to find a "middle gray" item and meter it... esp in high dr scenes like a sunset/panorama...

Bracket for those situations, and move on :-)

[wildman]

Thats what I do, but I'm looking for more dependable ways to expose properly for the highlights,

Not sure exactly what your problem is.
Generally I just spot meter on the highlights with EV set appropriately for the scene and when I look at the Hgram in Photoshop its pretty close to an ETTR exposure.
It works reliably enough for me and gives me the results I want.

Good luck...

[nomadkng]

Thats what I do, but I'm looking for more dependable ways to expose properly for the highlights, especially in M mode (histogram refers to wide open so it's no use).
I'd rather take less shots and ge it right or about right the first time.
Can't trust your eyes to find a "middle gray" item and meter it... esp in high dr scenes like a sunset/panorama...
It can be helpful when the GF is pressing you to move on...

This post to me sounds more like an amalgamation of the Zone System and ETTR philosophies. ETTR is post capture histogram dependent. The Zone system (as I understand it) meters off the highest EV portion of a scene and adds a 2? stop compensation for the 18% algorithm the camera uses per Wildman's comment

Not sure exactly what your problem is.
Generally I just spot meter on the highlights with EV set appropriately for the scene and when I look at the Hgram in Photoshop its pretty close to an ETTR exposure.
It works reliably enough for me and gives me the results I want....

The Zone system, or basic metering is PRE capture. The histogram then tells you if your calculations were correct and whether you captured as much data as you could for the given scene. (If you wanted to use the histogram in live view, you could use stop down metering, and then the histogram readout would be pre-capture, but that requires a lens with an aperture ring)

Keep in mind, digital images are all about signal to noise ratio and the amount of data that is stored for a given image. The ETTR philosophy has taken root based on the premise that more data is stored at the higher (right side) memory locations and the signal to noise ratio is better on that "side". This is the data you eventually utilize to create a finished image. Essentially, the more data you capture, the more latitude you have in post processing. It's also why a "properly exposed" ETTR file will actually be larger in size than an underexposed image. The histogram is not really a display of dynamic range, it is a display of the number of PIXELS and the amount of data stored in the spectrum from 0,0,0 (BLACK) to 255,255,255 (WHITE). It's not shadows and highlights, like film, its a colored pixel. It's why it's called clipping, when you exceed these numbers, because basically everything becomes 0,0,0 or 255,255,255. It's also why 14bit images (RAW) are far more forgiving than 8bit images (JPEG). You have inherently have more data to work with. There's a great article on Luminous Landscape that actually gets into the whole digital/computer-ese explanation that I'm trying to summarize here. It's why I rarely ever look at the LCD output except to read my histogram. (I think I'd actually use an option for a bigger histogram with no image)

The process and adjustments I make for my sunset exposures are ALL based on histograms POST capture. Especially sunsets, where the light changes with nearly every shot taken. It's shoot, chimp, adjust shutter speed, shoot again, chimp, adjust shutter speed. I also have parameters I require myself to stay within for aperture and shutter speed, depending on the subject. If my shutter speed gets too long/too short, I have to start adjusting aperture. Once I start getting above f16 or below f8, I have to start adjusting iso.

It's why I make a rule of being at my scene (for sunrise/sunset) 1 hour prior. I get set up, take test shots, adjust to a baseline histogram and then start shooting away. If you are being pushed to expedite the process, you should probably settle for a 5 exposure bracket and move on (or leave the GF at home?)

Bracket for those situations, and move on :-)

PS - If you regularly shoot sunsets, you MUST include 1, 2 & 3 stop Gradient (or Graduated, whichever you prefer) Neutral Density filters in your kit. GND will balance out the scene and properly expose for both highlights and foreground (shadows). They make a HUGE difference in final output. They make such a difference, you can actually see the improvement in the LCD display after just one shot.

[tuco]

...The Zone system (as I understand it) meters off the highest EV portion of a scene and adds a 2? stop compensation for the 18% algorithm the camera uses per Wildman's comment

There is a "Zone System" of metering and exposure for digital? For film, that's not the zone system. The zone system with film you typically place your shadows and develop for the highlights (expand, compress or average). But if you know your film and developing well enough, you could meter and place the highlights too. But that is not as easy to control as the opposite.

And those who only shot small format film rarely use the zone system of exposure thought you could use the metering system for average developing. Because when you compress or expand your highlights, you have to devote all the frames on the roll to that process and typically people had mixed shots on it (but not impossible). Medium format film with only 10 or 12 frames per roll is way more doable for the both the Zone System of metering and exposure. And of course sheet film is a natural fit. I practice the zone system with highlight compression technique with film today still.

[photodesignch]

I think I am just a super lazy guy. I don't want to go through all the "intensive" calculation before I shoot photos. Because that took a lot of fun out of my photography. I use RAW mainly, so I really relay on RAW development and the nature of RAW capability to do the task. The following is my steps and thinking process :

RAW can capture 5~7 steps depend on your camera (sensor size). For example, Sony A7 can do about +7 to -7 total 14 stops. However, since digital sucks more at highlight, but doing better at shadow details, let's try to preserve more on highlight.

1) meter on a scene, and turn on highlight chipping tool.
2) while metering, look at the chipping area.. turn +EV dial till overexposed area is acceptable to determine the brightest point of the image for my taste (percentage of the overexposed sky in the entire composition of the image of my liking).
3) since RAW can preserve at max of +4 EV, so I start to move back and dial down EV step by step from the highest point, while watching the subject at front make sure it's not too dark. (Normally would be dial down around 2 EV from brightest).
4) take the shot in RAW. Highlight will be well preserved. Recover shadow to make image balanced.. DONE!

This method is not to look at histogram at all.. Since a lot of camera can do preview real time for brightness chipping. This make sure I don't overexpose too much or underexpose too much. I am using my brain to justify the "mean" between brightest and darkest in my scene by my own eyes..

[LensBeginner]

I think I am just a super lazy guy. I don't want to go through all the "intensive" calculation before I shoot photos. Because that took a lot of fun out of my photography. I use RAW mainly, so I really relay on RAW development and the nature of RAW capability to do the task. The following is my steps and thinking process :

RAW can capture 5~7 steps depend on your camera (sensor size). For example, Sony A7 can do about +7 to -7 total 14 stops. However, since digital sucks more at highlight, but doing better at shadow details, let's try to preserve more on highlight.

1) meter on a scene, and turn on highlight chipping tool.
2) while metering, look at the chipping area.. turn +EV dial till overexposed area is acceptable to determine the brightest point of the image for my taste (percentage of the overexposed sky in the entire composition of the image of my liking).
3) since RAW can preserve at max of +4 EV, so I start to move back and dial down EV step by step from the highest point, while watching the subject at front make sure it's not too dark. (Normally would be dial down around 2 EV from brightest).
4) take the shot in RAW. Highlight will be well preserved. Recover shadow to make image balanced.. DONE!

This method is not to look at histogram at all.. Since a lot of camera can do preview real time for brightness chipping. This make sure I don't overexpose too much or underexpose too much. I am using my brain to justify the "mean" between brightest and darkest in my scene by my own eyes..

As far as I can tell, on the K-01 the highlight/blacks clipping warning is worthless in M mode.
I can set 1/100s f/32 in a dark-ish room and get no warning (or just a touch of yellow spots - because of auto-gain I guess) and then shoot and get a black frame (obviously).

The shoot-chimp-shoot-chimp-shoot thing, I'd rather not do it.

I already shoot RAW, I'd like to be able to place the values I see wherever I want by way of calculation, yes just like in a zone system for digital.

[nomadkng]

There is a "Zone System" of metering and exposure for digital?

Digital Zone System

The Digital Zone System | OutdoorPhotographer.com

these were the two main exposes that came to mind from my past reading. I'm sure if I googled I could find several others, but I find these two websites/sources to usually be credible

[photodesignch]

As far as I can tell, on the K-01 the highlight/blacks clipping warning is worthless in M mode.
I can set 1/100s f/32 in a dark-ish room and get no warning (or just a touch of yellow spots - because of auto-gain I guess) and then shoot and get a black frame (obviously).

The shoot-chimp-shoot-chimp-shoot thing, I'd rather not do it.

I already shoot RAW, I'd like to be able to place the values I see wherever I want by way of calculation, yes just like in a zone system for digital.

Only if you would understand what my method was about. It's both zone system + histogram all together, only I don't have to look at histogram. I knew the exact range of my digital negative can reach.. so I expose to the right according to know many steps i could save by just looking at the overexposed warning.

Yeah.. I don't think Pentax's highlight warning system is that great. On Sony both zebra and focus peak is accurate at it's best! I can even view real time histogram right in the EVF. but that's derailing the topic.

[tuco]

Digital Zone System

The Digital Zone System | OutdoorPhotographer.com

these were the two main exposes that came to mind from my past reading. I'm sure if I googled I could find several others, but I find these two websites/sources to usually be credible

Um, well okay. But it seems more just using the histogram and calling it the zone system for the digital age to me.