How Important Is The Histogram . . . Really?

[Scottelly]

I just shot this photo of a red flower:


As shot, exported from SPP 6.4.0 to a level 11 compressed JPEG

This is how the histogram looked, upon reviewing the image:


Crop from shot of the camera.

Here's the whole scene/shot:


This is the way the scene looked from above/behind the camera.

As you can see, I was shooting at ISO 50. The histogram makes it look like I am under-exposing the scene. If I had pushed the histogram to the right, I would probably have blown the red flower. Here is how the histogram looks in SPP, with warnings:


See how the histogram looks? How about those warnings!?!?

Here is how the photo looks after I made some adjustments, to try to reduce the warnings:


After adjusting with a -.5 saturation and switch from Sunlight to Auto white balance.


Settings, after adjustment, showing the reduced warnings.

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

Which version of the red flower photo do you like better?

Here is the raw file, just in case you would like to play with it yourself:

Loading…

sbkart.mpsdigitalmarketing.com

--
Scott Barton Kennelly

http://www.bigprintphotos.com

 

[xpatUSA]

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

To me, Scott, that says it all . . .

 

[MauriceRR]

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

It's significant, of course !!! but you need to know how it works ! a rgb system has a histogram a bit weird, beacause it overlaying the 3 histograms of the 3 channels.

If one of the channel you have no information (for saturated colors near the twoo other colors), you will have a clipped histogram. It's show the limit of the system. It's not a LAB channel, or a TSL model. It's working with RGB values. And it's working with the values of a choosen space, that's mean a clipped histogram can be clipped easily in srgb, less in prophoto.

By the wayn, the histogram on your camera is very probably just the histogram of the thumbnail, or a quick one not calculated with all finals rgb values.

For info, check the histogram of that :

http://jeromemichel.fr/up/tests/gradient.jpg

Moreover, you can convert it in prophoto, and look again the histogram, nothing to see with the one before in srgb.

Which version of the red flower photo do you like better?

first one. The second one is desaturated on my screen, and the tint has changed.

 

[Kendall Helmstetter Gelner]

Looking at the histogram it looks properly exposed to me because you can see a significant amount of red near the top end, and a bit of other color channel data as well. It's just that the scales make it a bit hard to see because of the stuff to the left, but if you zoomed into the image at all you'd see pretty quick from the revised histogram the exposure is good.

If you had overexposed it I'm pretty sure you'd see a good bump of red channel clipping at the end, another way the histogram helps...

 

[Tiger1]

The problem with ALL cameras is that NONE of them actually show a true RAW histogram. A program like Rawdigger displays the true RAW histogram.

The ultimate camera would have this feature and expose the image so no clipping occurred either side. Of course that would mean that you could manipulate the image and not introduce noise or other artefacts with this manipulation.

 

[Daedalus2000]

After adjusting with a -.5 saturation and switch from Sunlight to Auto white balance.


Settings, after adjustment, showing the reduced warnings.

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

Which version of the red flower photo do you like better?

I think the problem is that SPP will show you a warning if only one of the colors (RGB) is at zero, it does not mean that the image is underexposed. In this case SPP says that there is no Green in the flower, and that is why it gives you the warning. You can ignore it in this case

 

[Scottelly]

Settings, after adjustment, showing the reduced warnings.

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

Which version of the red flower photo do you like better?

I think the problem is that SPP will show you a warning if only one of the colors (RGB) is at zero, it does not mean that the image is underexposed. In this case SPP says that there is no Green in the flower, and that is why it gives you the warning. You can ignore it in this case

Wouldn't it be nice if we had more control over what happens, when the warnings are there? I mean wouldn't it be nice, if we could click a check box for each of the three colors, to eliminate/activate warnings for that color? Wouldn't it also be nice to be able to check a check box or switch a switch, which would change the warnings to only appear when the raw data is clipped? Hmmm . . . it seems like SPP could have a long way to go before it is perfect, huh?



--
Scott Barton Kennelly

http://www.bigprintphotos.com

 

[Scottelly]

O.K. So I just discovered that I can actually limit the warnings to any channel I want . . . or intensity (whatever that is). I either forgot about that or never knew about it.

 

[Scottelly]

Here's another photo I shot at ISO 100. I actually increased the saturation, causing the blue channel to blow. You can see in the following screen capture that the blue is blowing in various places all around the flower. At no adjustment there is almost no blue blown . . . only green. To me that means I got the exposure just right. (That's my inexperienced interpretation, anyway.)


A tiny bit saturated (+.1)


This shows the settings I used to export and the blue channel warnings.

What do you think of this exposure? What would you have done differently? (if anything)

Here is the raw file, just in case you want to play with it:

Loading…

sbkart.mpsdigitalmarketing.com

--
Scott Barton Kennelly

http://www.bigprintphotos.com

 

[TN Args]

Settings, after adjustment, showing the reduced warnings.

To me the warnings are B.S. I guess sometimes they are a good indicator, but like the histogram on the back of the camera, they don't always mean anything significant.

Which version of the red flower photo do you like better?

I think the problem is that SPP will show you a warning if only one of the colors (RGB) is at zero, it does not mean that the image is underexposed. In this case SPP says that there is no Green in the flower, and that is why it gives you the warning. You can ignore it in this case

Wouldn't it be nice if we had more control over what happens, when the warnings are there? I mean wouldn't it be nice, if we could click a check box for each of the three colors, to eliminate/activate warnings for that color?

You can do that in SPP by selecting the histogram channel of choice.

Wouldn't it also be nice to be able to check a check box or switch a switch, which would change the warnings to only appear when the raw data is clipped?

?? If warnings are ON, it does exactly that

Hmmm . . . it seems like SPP could have a long way to go before it is perfect, huh?

Are you sure you know it?

 

[SigmaChrome]

Here's another photo I shot at ISO 100. I actually increased the saturation, causing the blue channel to blow. You can see in the following screen capture that the blue is blowing in various places all around the flower. At no adjustment there is almost no blue blown . . . only green. To me that means I got the exposure just right. (That's my inexperienced interpretation, anyway.)


A tiny bit saturated (+.1)


This shows the settings I used to export and the blue channel warnings.

What do you think of this exposure? What would you have done differently? (if anything)

Here is the raw file, just in case you want to play with it:

http://sbkart.mpsdigitalmarketing.com/photos/SigmaSD14/SDIM7842.X3F

I don't know if anyone has mentioned this, but when Sigma released the firmware update that added ISO 50, they also added a warning that the amount of headroom would be substantially reduced. They recommended using manual metering and not using it for high contrast subjects.

Biological reds are by far the most difficult to meter correctly and I would not normally attempt to shoot a bright red flower on an SD14 at ISO 50, but I have done so successfully - as have you in that first shot.

In extreme exposure situations the histogram will often let you down. I suggest you use it only as a guide and base your exposures on readings that are taken from the light reflected directly off your subject - i.e. get in close. Check the exposure (and the histogram) and bracket if necessary. BTW, bracketing does not have to be used to make an HDR image, but to help you get the best possible exposure.

--
Regards,
Vitée
Capture all the light and colour!

http://www.pbase.com/vitee/image/98008547/small.jpg

Galleries by Vitee

www.pbase.com

 

[TN Args]

Looking at the histogram it looks properly exposed to me because you can see a significant amount of red near the top end, and a bit of other color channel data as well. It's just that the scales make it a bit hard to see because of the stuff to the left, but if you zoomed into the image at all you'd see pretty quick from the revised histogram the exposure is good.

If you had overexposed it I'm pretty sure you'd see a good bump of red channel clipping at the end, another way the histogram helps...

You mean like this?




SDIM7849 is overexposed

Notice all the little white spots:




Warnings OFF




Red channel warnings ON

See how all the white spots turn red when the Warnings ON fills them with a red (unfortunate) clipping shade.

IMHO the photo was overexposed, given the importance of red in this flower.

I saw SigmaChrome's comment about the use of ISO 50 possibly exacerbating this. Interesting.

 

[Scottelly]

Looking at the histogram it looks properly exposed to me because you can see a significant amount of red near the top end, and a bit of other color channel data as well. It's just that the scales make it a bit hard to see because of the stuff to the left, but if you zoomed into the image at all you'd see pretty quick from the revised histogram the exposure is good.

If you had overexposed it I'm pretty sure you'd see a good bump of red channel clipping at the end, another way the histogram helps...

You mean like this?


SDIM7849 is overexposed

Notice all the little white spots:


Warnings OFF


Red channel warnings ON

See how all the white spots turn red when the Warnings ON fills them with a red (unfortunate) clipping shade.

IMHO the photo was overexposed, given the importance of red in this flower.

I saw SigmaChrome's comment about the use of ISO 50 possibly exacerbating this. Interesting.

Here is the way that histogram looks after reducing the saturation by .5 and with slight reductions in exposure (only .1 reduction) and contrast (also only .1 reduction):


As you can see, the red channel was no longer blown after the adjustments.

I figure that's only a slight adjustment, and well within the raw file's range of possible adjustment. I guess I could have under-exposed the shot a bit by setting the shutter speed to 1/160 instead of 1/125, but that's but a slight difference, which I didn't figure was necessary at the time. Maybe in the future I will bracket when shooting red flowers, pure green stuff (i.e. the green grass of a golf course), and pure blue stuff (i.e. a blue car, boat, or building).

I find it surprising that the ISO 100 shot, for which I used twice the shutter speed (meaning half the open time - 1/250), turned out to be exposed perfectly (wasn't it?), when the ISO 50 shot was not. The ISO 50 shot seems more limited, though if the camera is truly non-AFE, it should actually have a greater dynamic range at ISO 50 than at ISO 100, right?

I'm thinking I'll have to experiment in a studio setting with strobes and some red flowers . . . and possibly some other red things, just in case the differences in exposure were due to the sun sinking closer to the horizon instead of some difference in the camera . . . though you'd think the exposure would have been less in the ISO 50 shot, since that was the later one, and it was late afternoon, when I shot the photo. (It was a clear sky, so I don't think it was a cloud cover issue.)

--
Scott Barton Kennelly

http://www.bigprintphotos.com

 

[Zone8]

... what's a Histogram Scott? (Only kidding, in case someone sees RED!) I've never bothered but then, I do everything manually including exposure to ensure it's "right".

 

[borissimo86]

I think that the in-camera histogram is not a RAW histogram, that it uses the JPEG for histogram information. Maybe someone here can confirm. I thought that it would help to use UniWB but I have not been able to get it working on my SD1 Merrill. I can get blue and green channels pretty closely aligned, but the camera refuses to set custom white balance for targets with extreme casts so the red channel always trails by at least a stop. Only getting half-way there makes it not worth sacrificing the custom white balance slot, IMO.

 

[xpatUSA]

Allow me to help . .

I think that the in-camera histogram is not a RAW histogram, that it uses the JPEG for histogram information. Maybe someone here can confirm.

That is correct.

I thought that it would help to use UniWB but I have not been able to get it working on my SD1 Merrill. I can get blue and green channels pretty closely aligned, but the camera refuses to set custom white balance for targets with extreme casts so the red channel always trails by at least a stop. Only getting half-way there makes it not worth sacrificing the custom white balance slot, IMO.

I tried UniWB on an SD10, following Guillermo Luijck's full instructions to the letter, including the use of ShowImage to make a screen target. It actually worked quite well as intended but of course the LCD colors looked funny and the (by modern standards) small LCD didn't help much with viewing the histogram.

As a regular user of RawDigger please note that the red channel is always at least a stop low. In the meta-data there is a X2 factor which I think is used by SPP to 'equalize' channel responses and that may be why the red can be noisy.

 

[digidog]

I think that the in-camera histogram is not a RAW histogram, that it uses the JPEG for histogram information. Maybe someone here can confirm.

Yes. Or let's put it this way: it's either a match for the data converted to JPEG or it's a match for the raw (if so, I don't know what camera does this).

I thought that it would help to use UniWB but I have not been able to get it working on my SD1 Merrill. I can get blue and green channels pretty closely aligned, but the camera refuses to set custom white balance for targets with extreme casts so the red channel always trails by at least a stop. Only getting half-way there makes it not worth sacrificing the custom white balance slot, IMO.

Some of us exposed optimally for very unforgiving media, transparency film, for years and professionally long before Histograms existed on cameras. When I shot national ads, shot for magazines, for annual reports, all on transparency film, I needed to nail exposure by 1/4 a stop or I didn't deserve to be paid! You absolutely do NOT need any Histogram on your camera to do this. Unfortunately, a lot of 'young folks' never learned fundamental exposure, something that's photography 101 without the lie of the Histogram as perhaps a crutch. Would having a raw Histogram be useful on camera? Sure. Must we have one? No. But don't look at a Histogram built from and based on a camera JPEG and expect it has bearing on the raw data or is necessarily to use one to optimally expose.

 

[reeluff]

Some of us exposed optimally for very unforgiving media, transparency film, for years and professionally long before Histograms existed on cameras. When I shot national ads, shot for magazines, for annual reports, all on transparency film, I needed to nail exposure by 1/4 a stop or I didn't deserve to be paid! You absolutely do NOT need any Histogram on your camera to do this. Unfortunately, a lot of 'young folks' never learned fundamental exposure, something that's photography 101 without the lie of the Histogram as perhaps a crutch. Would having a raw Histogram be useful on camera? Sure. Must we have one? No. But don't look at a Histogram built from and based on a camera JPEG and expect it has bearing on the raw data or is necessarily to use one to optimally expose.

Young folks do not have to care what old folks had to do back in the day, when better tools exist to do the job better, like histograms. Histogram is a tool and using one does not make it a crutch or a lack of knowledge. A raw histogram would be great, the "lie" based on the camera preview is good enough if you spend a little time understanding how the in camera histogram on your specific camera correlates with the raw data.

 

[MauriceRR]

Some of us exposed optimally for very unforgiving media, transparency film, for years and professionally long before Histograms existed on cameras. When I shot national ads, shot for magazines, for annual reports, all on transparency film, I needed to nail exposure by 1/4 a stop or I didn't deserve to be paid! You absolutely do NOT need any Histogram on your camera to do this. Unfortunately, a lot of 'young folks' never learned fundamental exposure, something that's photography 101 without the lie of the Histogram as perhaps a crutch. Would having a raw Histogram be useful on camera? Sure. Must we have one? No. But don't look at a Histogram built from and based on a camera JPEG and expect it has bearing on the raw data or is necessarily to use one to optimally expose.

Young folks do not have to care what old folks had to do back in the day, when better tools exist to do the job better, like histograms. Histogram is a tool and using one does not make it a crutch or a lack of knowledge. A raw histogram would be great, the "lie" based on the camera preview is good enough if you spend a little time understanding how the in camera histogram on your specific camera correlates with the raw data.

There is no raw histogram. a rgb histogram is made with rgb values. Raw is made of linear lab values. (and with a gama 1 I guess). I guess you can make one, but it would be no significant.

A RGB histogram is dependant of a rgb system AND a gamut profile (srgb/adobe/prophoto, etc).

 

[MauriceRR]

Let's play a bit.


prophoto file


srgb file

First serie is made in my screen profile, then converted in prophoto. If you have a good adobe rgb monitor, take a look.

My screen is a eizo cx240, only 97%aRGB gamut, but lot more in reds.

Second serie is made in my screen profile, but this time I attribuated a srgb profile, in consquence, if you have a standard/poor screen you will see diferencies, for each. But all the values are staggered in srgb, so there is a bit more of twisted colors.

Every histograms with corresponding output. Outputs in tiff 16 bits, the profile is in each filename.

PS1 : For info, I didn't trust color values from sigma software, I had right, specially in ARGB.

PS2 : I've seen a value with a lab color a:123 !!! , in prophoto tiff file, it's so close to the max that I would say there is false colors in this flower, and I would not really trust it like a good value.

PS3 ; download images, open them in photoshop, colormanagement in navigators are not great at all.

I hope you enjoy the work.