Camera histograms - are they useful?

[pdk42]

When I was in Scotland recently with a photographer friend, we had a little discussion on whether the in-camera histogram is useful or not. He says "yes", I say "no". It would be interesting to hear your views. The discussion applies to raw shooting only, and probably mainly to landscapes where preservation of highlights is essential. Here's my case:

• Since we're talking raw, the brightness of the image can be corrected in PP later.
• To minimise noise, we want to increase the exposure as much as we can, but not to the point of clipping highlights (i.e. ETTR).
• Given that, what we need is something that tells us when highlights are clipped.

OK, so we need a histogram then? Well, maybe, but I contend that the histogram is too crude:

• Clipping is shown by the highlights pushing up to the right edge of the graph.
• But the difference between clipping and not clipping isn't that obvious.
• The amount of clipping is hard to assess.
• And you have little idea WHAT is clipping in the image.

So, what's the alternative?

Zebras, or "blinkies" (over-exposure indicators) of course!

• They not only tell you that you're clipping highlights, but they show you WHERE the clipped highlights are.
• So, you can choose to perhaps leave some clipped by a quick assessment of the EVF/screen. Like the sun for instance, but not the clouds.
• They overlay nicely on the image. By contrast, the histogram is some alien graph that's splatted somewhere you probably don't want it!

I rest my case!

I suspect he sticks to the histogram since he shoot Nikon Z and that doesn't support blinkies/zebras (although I understand there is a sort of kludge that can be used as a workaround).

 

[CharlesH]

I agree, and use zebras. I’ve always found histograms to be pretty useless to avoid clipping highlights. This is for photography; for video I use the waveform monitor, and sometime zebras plus the waveform monitor.

 

[AlanC]

I'd say both the histogram and the zebras aren't particularly useful because they're based on the output of the camera's JPEG processing and so aren't a true representation of the RAW data. Try using either with the camera first set to the vivid photo style and then to the V-Log style: you'll get two very different results.

And while it's not any more accurate, I prefer the "False Color" display on the Sigma fp L: this is like an extended version of zebras, with a colour coded overlay showing black clipping / just before black clipping / medium gray / medium gray + 1EV (standard exposure) / almost white clipping / white clipping. You can call it up to check the exposure by assigning it to one of the buttons so it doesn't interfere with composing the image (I find the motion of the zebras really annoying).

 

[pdk42]

I'd say both the histogram and the zebras aren't particularly useful because they're based on the output of the camera's JPEG processing and so aren't a true representation of the RAW data. Try using either with the camera first set to the vivid photo style and then to the V-Log style: you'll get two very different results.

And while it's not any more accurate, I prefer the "False Color" display on the Sigma fp L: this is like an extended version of zebras, with a colour coded overlay showing black clipping / just before black clipping / medium gray / medium gray + 1EV (standard exposure) / almost white clipping / white clipping. You can call it up to check the exposure by assigning it to one of the buttons so it doesn't interfere with composing the image (I find the motion of the zebras really annoying).

I always find it strange that none of the manufacturers give us a raw data over-exposure overlay/zebras.

 

[Pete_W]

I don’t use the histogram but I do use blinkies.

BTW, the histogram can be moved around on the screen if it’s in the way.

 

[RuleOfThirds]

On my GR3 I sometimes use Highlight weighted, just checked, S5ii has that metering mode also. On the GR3 sometimes it goes too dark, but S5ii has much more latitude.

I never used the histogram during camera exposure. Blinkies/Zebra’s yes!

 

[Oíche]

I always find it strange that none of the manufacturers give us a raw data over-exposure overlay/zebras.

Agree, is there a reason why they don't give definitive tools to measure the exact limit of the sensor regarding RAW? Do different RAW convertors make this difficult or impossible?

Histograms, blinkies, zebras etc are always uber consevative but that leaves a twilight zone for us trying to expose to the right as far as possible. I suppose you can only learn how far to push before completely losing highlights.

I agree with Paul though, the histogram is just too crude, too conservative and gives no specific information, especially with the sun in the frame as I like to do a lot hence why I consider sunstars, flare and ghosting important in lens choice.

I need to try zebras etc. with the S5ii, I'm trying to use it like a film camera and at 10/100 proficiency with it so far and not confident with it to do as I did before, the amount of options are a head melter. The iA and P modes are never going to be used, I never used Program modes and iA seems to be kind of the same in that the camera doesn't do what you want.

 

[CharlesH]

I'd say both the histogram and the zebras aren't particularly useful because they're based on the output of the camera's JPEG processing and so aren't a true representation of the RAW data. Try using either with the camera first set to the vivid photo style and then to the V-Log style: you'll get two very different results.

True, but I don't think it matters much. And V-Log isn't a good reference because the camera automatically switches to ISO 640 with V-Log; if you have a different ISO for Vivid the histogram/zebras will jump to ISO 640.

If you are shooting RAW you shoot Standard Photo Style. You can add a Photo Style in post processing if you want.
To put it in quantitative form I turned on the Luminance Spot Meter, set the camera to look at a uniform white lighted surface (my calibrated monitor) and adjusted the camera so Standard was 100% on the Spot Meter. I had zebras, mode ZEBRA2, set to 95% - so zebras were showing in this setup. Here are the Spot Meter readings for the selected Photo Styles (I have not included them all, but for all the styles the minimum was Flat at 93% and the maximum was Like709 at 102%).

Photo Style	Spot Meter	Zebras
Standard	100%	On
Vivid	100%	On
Natural	96%	On
Flat	93%	Off
Landscape	100%	On
Portrait	96%	On
L.Monochrome D	99%	On
Like709	102%	On
V-Log	+3.5 Stops	Off

Consider using zebras with RAW, with the Photo Style set to Standard. With this same white setup, and increasing iSO, the maximum Luminance Spot Meter reading is 109%. That is, the sensor saturates at 9% higher than in the table above. The maximum ZEBRA2 (or ZEBRA1) setting is 95%. So when you first see zebras you are 14% away from saturation. Now going back to the white screen, what does this mean? You can open the aperture two clicks to saturation (109%), or slow the shutter two clicks to saturation. Or just very slightly more than this. In practice you should be okay with one click more exposure after you see zebras. In practice, of course, it depends on what part of the image shows zebras, and it may be okay to saturate some things.

 

[CharlesH]

I spent some time today taking pictures, outside in the sunlight, of scenes that included some bright areas whereI coul be usde zebras to set the exposure without saturating the sensor. A first thing I noticed was an error from the posting last night. The maximum ZEBRA setting with the Photo Style set to Standard is 105% (not 95%). Unfortunately I had the Photo Style set to V-Log when I set ZEBRA, and the maximum when using V-Log is 95%.

So for photography today using the Standard Photo Style I worked with zebras at 100% and also at 105%. Either of these worked okay; at 100% zebra detection the exposure can be increased another 9% before saturation; at 105% zebra detection it can be increased another 4%. In each case I also used the Luminance Spot Meter to back up the zebras. For the things I was shooting the Spot Meter worked okay, but often areas in the image will be too small for a good Spot Meter reading, where zebras will pick up these smaller areas.

Historically I've used 95% zebras for photography. I'm now trying decide between 100% and 105%. I think I'll try 105% for a while and see how it goes.

 

[pdk42]

My own approach is to use flat profile with zebras set at 105%. I then try to ensure nothing important (usually the sky!) shows in the zebras. That works well for me.

 

[Markuswelder]

Not a fan of the histogram. I'm also a zebra stripes user, I shoot standard or natural JPEG picture profile plus raw, zebras at 105%, and don't care too much if single colours blow out a little bit. Like plain blue sky for example. There's no detail there, just plain blue, so let it blow out a bit.