Digital metering with histogram

[Chris_T]

As a film shooter, I'm fascinated by the discussion in the thread "Shooting to the right and raw conversion, Why shoot to the right":

http://luminous-landscape.com/forum/index....showtopic=17706

To me, ETTR seems very similar to how I would expose transparencies, i.e. meter and preserve highlight details. With an in camera histogram, I thought that ETTR would be easier and more accurate. But the thread seems to suggest that there is more to ETTR than simply relying upon the jpeg histogram. So, do the seasoned film shooters have to change how they meter digitally? Here are some more specific questions:

- In a film camera, I rely upon the exposure indicator (i.e. the analog scale) to adjust the exposure settings. If a jpeg histogram is less than ideal, why not fall back to the exposure indicator?

- Often a film camera's meter needs some calibration for accuracy. The following article suggests something similar for "calibrating" the histogram. Is it really that simple?

http://visual-vacations.com/Photography/Di...Photography.htm

[rdonson]

As you've already seen folks like to dive to the quantum physics level in this debate pretty quickly.

You've also seen the debates over an in-camera histogram built on 8 bit JPEG data and that the histogram in Photoshop is really only 8 bit.

If I may summarize, your basic question is, how do I get a feel for this?

A bit of experimentation on how your in-camera histogram relates to a good histogram in Photoshop.  Here's a link to a free 16 bit histogram for Photoshop from Reindeer Graphics that should aid in the adventure.

[John Sheehy]

A bit of experimentation on how your in-camera histogram relates to a good histogram in Photoshop.  Here's a link to a free 16 bit histogram for Photoshop from Reindeer Graphics that should aid in the adventure.
[a href=\"index.php?act=findpost&pid=126889\"][{POST_SNAPBACK}][/a]

What would you expect a 16-bit histogram to show that an 8-bit didn't show?  Other than the ability perhaps to zoom into the histogram, or show one more than 256 pixels wide with full detail, the 8-bit shows the full range of 16-bit RAW conversion.  Min and Max are still in the same place.

What is really needed is a tool to automatically create RAW RGB histograms from RAW files.  I believe someone made one based on DCRAW last year, but it would be limited to the cameras that DCRAW could read back then.  You have to know where black and clipping are in the RAW file (they aren't always at 0 and 4095 or 16383 or 65535), to properly set the blackpoint and whitepoint of the RAW file for a specific camera at a specific ISO.

[rdonson]

What would you expect a 16-bit histogram to show that an 8-bit didn't show?  Other than the ability perhaps to zoom into the histogram, or show one more than 256 pixels wide with full detail, the 8-bit shows the full range of 16-bit RAW conversion.  Min and Max are still in the same place.
[a href=\"index.php?act=findpost&pid=126904\"][{POST_SNAPBACK}][/a]

All true. I like looking at the wider histogram.

[Jonathan Wienke]

- In a film camera, I rely upon the exposure indicator (i.e. the analog scale) to adjust the exposure settings. If a jpeg histogram is less than ideal, why not fall back to the exposure indicator?

- Often a film camera's meter needs some calibration for accuracy. The following article suggests something similar for "calibrating" the histogram. Is it really that simple?

http://visual-vacations.com/Photography/Di...Photography.htm

It is, but of course it isn't.

It takes some time and thought to configure your camera to show the most accurate histogram it can, and judgment on your part to understand the histogram and its limitations and make decisions regarding whether certain highlights are really that important orn not. If you have a camera that only displays a luminance histogram, avoiding clipping in the red and blue channels when shooting certain saturated colors can be problematic, and requires some testing to see much you need to adjust exposure to avoid problems. When shooting flowers with my 1Ds, (which only has luminance histogram) I have to back off exposure as much as 2/3 of a stop to avoid clipping. With the 1D-MkII, I have the contrast set low and WB set to daylight, and the RGB histogram gets me close enough that I can make minor adjustments in ACR without compromising the image significantly.

I use the camera's internal meter to get me in the ballpark, and then fine-tune based on the histogram. I put the camera in aperture priority, and shoot a test shot, check the histogram, and either adjuse exposure compensation or switch to manual and dial in adjustments there. I can usually get dialed in perfectly with one or two test shots.

BTW, the link I think you meant to post is Digital Exposure And Metering Strategies

[Chris_T]

What we really need is to better understand the in camera's histogram, and how to use it, such as:

- Detailed descriptions of what an in camera's histogram is showing, from the manufacturers.

- Camera reviews devoting some verbage on the accuracy of an in camera histogram, or how to get the best out of it.

But I won't hold my breathe for either.

If I can't meter accurately (and quickly), what good does the high resolution, low noise, speedy focus do? For someone already familiar with how to meter a film camera, the digital metering compexity came as a surprise. This dog is getting too old to learn yet another new trick.

What is really needed is a tool to automatically create RAW RGB histograms from RAW files.
[a href=\"index.php?act=findpost&pid=126904\"][{POST_SNAPBACK}][/a]

[Chris_T]

I use the camera's internal meter to get me in the ballpark, and then fine-tune based on the histogram. I put the camera in aperture priority, and shoot a test shot, check the histogram, and either adjuse exposure compensation or switch to manual and dial in adjustments there. I can usually get dialed in perfectly with one or two test shots.

Being able to view digital test shots is a tremendous advantage. But for fleeting moments, this is irrelevant.

Aside from making test shots, have you changed how you meter digitally from meter film?

BTW, the link I think you meant to post is Digital Exposure And Metering Strategies
[a href=\"index.php?act=findpost&pid=126963\"][{POST_SNAPBACK}][/a]

Oops, I must have gotten the link wrong. Thanks for posting and correcting. If there are similar articles on this topic, please point me to them.

[Eric Myrvaagnes]

Hi, Chris,

When I have the time (and patience -- which I admit isn't that often), I use Jonathan's technique.

I have done some rudimentary experiments which have given me some "rules of thumb" to go by in certain situations that generally get me pretty close to the correct ETTR exposure for those fleeting moments. If a scene has a lot of bright, light areas in it (excluding specular highlights), I know that my camera will tend to underexpose, so I give a standard extra half-stop exposure. If the scene is overall on the dark side, I will give a half-stop less exposure. For most "normal" scenes, my meter seems to do a pretty good job used normally.

But whenever I have time to do so, I check the histogram and redo the shot if adjustment is needed and the scene is still worth shooting.

So basically I'm using the same skills as "before digital", but now I check with the histogram whenever there is time to.

Just my 2 cents.

Eric

[Jonathan Wienke]

Being able to view digital test shots is a tremendous advantage. But for fleeting moments, this is irrelevant.

Not if you take a test shot before the fleeting moment happens. Most situations let you take some shots and wait for the action.

If not, you use the camera meter and hope you guessed right, just like with film.