Face exposure / skin tones

[dancook]

Advertisement (gone after registration)

Nevermind you're using the Q2, not Q

Edited April 9, 2019 by dancook

[paul.bridges.3388]

10 hours ago, dancook said:

Guinness might be a giveaway, but looks dark hair, light skin, looks like an Irish beauty to me. 

Ain’t that a fact 😏

[paul.bridges.3388]

14 hours ago, evikne said:

I do that. Always. In my wallet I always carry I credit card size WhiBal card, and for a bit more serious shooting I use an ExpoDisc WB filter. To me, there is always a subtle, yet (to me) important difference from the camera's AWB. And it saves a lot of time in post.

Totally get that and genuinely interested in understanding how this would work,  practically speaking.

How would your white balance card technique work in the pub situation, like the above?

Acknowledging this is a somewhat routine family snap...it’s still precisely the kind of photo that I bought the Q2 for.  It’s these family candids that I frame the most....and so want to get right.

Thanks for all your help!

[wda]

Paul, going back to my original post, in LR Develop module I pressed 'W' and my cursor changed shape ready to touch a neutral grey tone, if there is one. In your picture I touched a white pixel on the beer mats. Had I used a calibrated grey card I would have eliminated more colour contamination. Sometimes, such as a pub scene, that is too clinical and I would tweak the result to make it subtly warmer to match my memory of the ambient lighting. If that messes up facial colour I would use a brush (K) on the face to remove the false colour from the face.

I also have a WhiBal,  sadly no longer available in the UK, which I also use to set a custom white balance in camera. (Check white balance options in the camera menu) That method is good for situations where the lighting is not changing.

Colour fidelity is alien to less serious  photographers. But for others there are excellent video tutorials to help one's understanding. Just Google search and take your pick. Once you develop your own techniques, you will begin to appreciate the difference colour management can make.

[evikne]

3 hours ago, paul.bridges.3388 said:

Totally get that and genuinely interested in understanding how this would work,  practically speaking.

How would your white balance card technique work in the pub situation, like the above?

Acknowledging this is a somewhat routine family snap...it’s still precisely the kind of photo that I bought the Q2 for.  It’s these family candids that I frame the most....and so want to get right.

Thanks for all your help!

In a pub situation like the above, I would just hold the WhiBal card in front of me, in the same light as the main subject, and take a picture of the card. Be aware of light reflections on the card. If the gray color is overexposed, the picture will be useless. Usually I take this reference picture after I've taken the ordinary pictures. It all takes less than five seconds. When back home, I press "W" in LR and click once on the card in the picture with the WB pipette to "reset" the gray color in the picture. Consequently all the other colors will fall into place too. Then I just copy the white balance settings from this picture over to the other pictures from the same place.

It is also possible to take the reference picture first, and then set a custom WB in the camera, but this is a less spontaneous method which I rarely use. The advantage is that you don't need the reference picture when coming home: Just select WB "As Shot" in LR and you're all set.

You can read more about WhiBal here: http://michaeltapesdesign.com/whibal.html

I think ExpoDisc is even more precise. It picks up the light from a wider angle than WhiBal, and the result is in a way an average color of the ambient light. It works in a similar way, but the reference picture shall be taken through the filter in the opposite direction: from the subject's viewpoint and towards the place you took / are going to take the other pictures. If used in advance, the ExpoDisc can also help you setting a correct exposure.

You can read more about how it works here: https://www.expodisc.com

Edited April 9, 2019 by evikne

[Marcel Bertran]

All of the comments have a degree of merit, even going back to the, “take a spot/Ctr weighted reading off of your 🖐 & open 1stop if you don’t have an 18% gray card or exposure aid device.”  Skin tones & facial structures are far from consistent, especially when you consider the multitude of exposure & ambient variables of light, makeup, reflection, white balance etc. Cameras, lenses, sensors & meters have unique fingerprints & signatures. Dedicated  photographers learn these characteristics & adapt their capture & post processing workflow accordingly. To expect a JPG grab shot, even in the hands of an experienced photographer with acceptable in camera settings, to produce other than mediocre results, is really pushing your photo God’s prayers to the extremes. Most photographers don’t have the luxury of a studio environment to hopefully produce consistent results ... we learn & adapt are photographic techniques to a degree of personal acceptance & utilize the tools that are now available, in particular RAW & Post Processing AP’s, to fulfill are interpretive subjectivity.

“Follow Your Heart and Your Pathway Can Be Illuminated By Your Creative Vision and Your Passion Can Serve As Your Compass”

 

 

[leicamario]

Advertisement (gone after registration)

Grab a Fuji and you know cameras can produce magic with in-camera jpeg-processing.

[antigallican]

Just expose for the brightest thing in the scene - in this case her cheek or under eye. Everything else will come back in LR. I had to be told this after a lifetime of b+w analogue photography. In digital you can recover quite a bit of detail in shadows but never in blown highlights, unfortunately.

[FHPdoc]

34 minutes ago, antigallican said:

Just expose for the brightest thing in the scene - in this case her cheek or under eye. Everything else will come back in LR. I had to be told this after a lifetime of b+w analogue photography. In digital you can recover quite a bit of detail in shadows but never in blown highlights, unfortunately.

This is really a critical point, I think, because film and digital exposure strategies are fundamentally different, are they not?

In film, we expose for the shadows because pure black completely erases the emulsion from the film base during development, while the emulsion for highlights remains, allowing the potential for at least some recovery.  With digital, we expose for the highlights as John describes above, though I'm afraid I don't understand the underlying technology of digital sensors to explain why.  Can somebody else provide the technical explanation for this?

[gotium]

4 hours ago, antigallican said:

Just expose for the brightest thing in the scene.

This is what I do with the Q. If there’s no time, I just leave it on -.3 or -.7 exposure comp and hope for the best - opposite of film. 

Having said that, I like the look of the blown highlights and the flattened facial features in your example, and I’d leave it alone. Or convert to B and W and enjoy the tones.

Edited April 10, 2019 by gotium

[leicamario]

I worked with an expo disc in my professional life. Man, thats not what I want to do at home.

We can discuss it all day and night, but the skin tones/Auto-WB in this price range could be better.

[cnj]

6 hours ago, FHPdoc said:

In film, we expose for the shadows because pure black completely erases the emulsion from the film base during development, while the emulsion for highlights remains, allowing the potential for at least some recovery.  With digital, we expose for the highlights as John describes above, though I'm afraid I don't understand the underlying technology of digital sensors to explain why.  Can somebody else provide the technical explanation for this?

The digital situation is an (if you forgive the term) analog to the film situation you describe.  Despite referring to raw files as "digital negatives", digital sensors record the (positive) signal they receive in each of red-light, blue-light, and green-light.  As long as the value for each colour stays below the maximum for the sensor, there's meaningful data and you can always (effectively) increase the gain on the signal to pull more data out of lower values which also increases noise.

 

If your red channel is at the maximum value for the sensor (let's say 255), and you continue to expose the scene, there's no way for the sensor to record the extra light, so if you had a gradient from 130 to 150 and then double the exposure, you now have a flat 255 to 255 over-exposed area.  If you expose the scene long enough, eventually all three channels (RGB) become over-exposed (maxing out at 255) and you end-up with pure white.  (Hopefully there's a helpful explanation in there, albeit a little rushed on the example.)

[FHPdoc]

2 hours ago, cnj said:

The digital situation is an (if you forgive the term) analog to the film situation you describe.  Despite referring to raw files as "digital negatives", digital sensors record the (positive) signal they receive in each of red-light, blue-light, and green-light.  As long as the value for each colour stays below the maximum for the sensor, there's meaningful data and you can always (effectively) increase the gain on the signal to pull more data out of lower values which also increases noise.

 

If your red channel is at the maximum value for the sensor (let's say 255), and you continue to expose the scene, there's no way for the sensor to record the extra light, so if you had a gradient from 130 to 150 and then double the exposure, you now have a flat 255 to 255 over-exposed area.  If you expose the scene long enough, eventually all three channels (RGB) become over-exposed (maxing out at 255) and you end-up with pure white.  (Hopefully there's a helpful explanation in there, albeit a little rushed on the example.)

Thank you--this is very helpful!  Where do you suggest one go to learn more about what you have described?  I'd like to see how this connects to histograms, plus learn about what happens at the bottom end of the tonal spectrum.

Thanks so much again!

- Scott

[cnj]

21 minutes ago, FHPdoc said:

Thank you--this is very helpful!  Where do you suggest one go to learn more about what you have described?  I'd like to see how this connects to histograms, plus learn about what happens at the bottom end of the tonal spectrum.

Glad it was helpful.  Depending on how technical you want to get, there are a lot of information if you get the terms right (unfortunately, my first few searches returned some rather dated pages dating back to when CCDs were still dominating the high-end).

Some good background, although not directly related to your concern about histograms, but explains how most cameras capture colour (Fuji uses a variation on the theme, and Sigma's Foveon X3 is quite different) : https://www.cambridgeincolour.com/tutorials/camera-sensors.htm

The histogram is essentially showing what fraction of the scene (y-axis) is at a certain brightness level (x-axis, from perfectly dark "0" on the left to max exposure "255" in my example above).  A quick search brought-up B&H with a good primer: https://www.bhphotovideo.com/explora/photography/tips-and-solutions/how-read-your-cameras-histogram and Cambridge in Colour again with a little more detail: https://www.cambridgeincolour.com/tutorials/histograms1.htm

You probably want to read-up on noise and ISO invariance to understand what's happening at the bottom end on the digital spectrum.  https://www.cambridgeincolour.com/tutorials/image-noise.htm covers types of noise; there are plenty of other posts.  ISO invariance essentially means that the camera records the same noise at ISO 200 as ISO 400, all other exposure settings aside: 1/200s, f2, ISO 200 pushed one stop in post will look the same as 1/200, f2, ISO 400 as captured by the camera.  Most cameras add additional processing noise such that the ISO 200 pushed image will have additional noise than natively capturing at ISO 400.

In addition to the noise issue, if you think of the 255 steps of colour (per channel in my earlier example): if you're under exposing the whole scene, it means you're only using a fraction of the potential values.    While *over* exposing means you hit a hard ceiling and clipping with digital capture, *under* means you have less signal to describe changes in colour or luminosity and you end-up with flat, blotchy areas.  Consider what your monitor looks like if you were to switch from 16-bit colour to 256-colours (assuming that's still a setting most computers support; otherwise consider converting a JPEG with smooth colour transitions into an indexed image like a GIF, which has limited colour information).  Playing around with levels and curves in post help adjust how the captured data is interpreted, and Ansel Adam's zone system also shares some similarities.  I fear all of this is heading way off-topic though.

To bring it back to the larger discussion, correctly setting white balance before taking the picture will help both your auto-focus system, and you looking at  post-capture histograms.  This is because white-balance changes how the value in each channel is interpreted, and when correcting for the appropriate white-balance, may cause a colour shift such that a channel which *hadn't* been over-exposed before is now over-exposed.  Depending on your software, this might be recoverable and is a reason for getting white-balance right initially.  That said, I still expose to the right, develop to the left, and usually fix white balance in post 😕  AWB has come a long way.

[FHPdoc]

9 hours ago, cnj said:

Glad it was helpful.  Depending on how technical you want to get, there are a lot of information if you get the terms right (unfortunately, my first few searches returned some rather dated pages dating back to when CCDs were still dominating the high-end).

Some good background, although not directly related to your concern about histograms, but explains how most cameras capture colour (Fuji uses a variation on the theme, and Sigma's Foveon X3 is quite different) : https://www.cambridgeincolour.com/tutorials/camera-sensors.htm

The histogram is essentially showing what fraction of the scene (y-axis) is at a certain brightness level (x-axis, from perfectly dark "0" on the left to max exposure "255" in my example above).  A quick search brought-up B&H with a good primer: https://www.bhphotovideo.com/explora/photography/tips-and-solutions/how-read-your-cameras-histogram and Cambridge in Colour again with a little more detail: https://www.cambridgeincolour.com/tutorials/histograms1.htm

You probably want to read-up on noise and ISO invariance to understand what's happening at the bottom end on the digital spectrum.  https://www.cambridgeincolour.com/tutorials/image-noise.htm covers types of noise; there are plenty of other posts.  ISO invariance essentially means that the camera records the same noise at ISO 200 as ISO 400, all other exposure settings aside: 1/200s, f2, ISO 200 pushed one stop in post will look the same as 1/200, f2, ISO 400 as captured by the camera.  Most cameras add additional processing noise such that the ISO 200 pushed image will have additional noise than natively capturing at ISO 400.

In addition to the noise issue, if you think of the 255 steps of colour (per channel in my earlier example): if you're under exposing the whole scene, it means you're only using a fraction of the potential values.    While *over* exposing means you hit a hard ceiling and clipping with digital capture, *under* means you have less signal to describe changes in colour or luminosity and you end-up with flat, blotchy areas.  Consider what your monitor looks like if you were to switch from 16-bit colour to 256-colours (assuming that's still a setting most computers support; otherwise consider converting a JPEG with smooth colour transitions into an indexed image like a GIF, which has limited colour information).  Playing around with levels and curves in post help adjust how the captured data is interpreted, and Ansel Adam's zone system also shares some similarities.  I fear all of this is heading way off-topic though.

To bring it back to the larger discussion, correctly setting white balance before taking the picture will help both your auto-focus system, and you looking at  post-capture histograms.  This is because white-balance changes how the value in each channel is interpreted, and when correcting for the appropriate white-balance, may cause a colour shift such that a channel which *hadn't* been over-exposed before is now over-exposed.  Depending on your software, this might be recoverable and is a reason for getting white-balance right initially.  That said, I still expose to the right, develop to the left, and usually fix white balance in post 😕  AWB has come a long way.

Wow--GREAT info--thank you so much again!!  Now I'll go do my homework...

[cnj]

22 hours ago, cnj said:

…correctly setting white balance before taking the picture will help … your auto-focus system…

Sorry, I meant automatic exposure rather than focus there, in case anybody was confused.

[antigallican]

Obviously lots of people reading this thread will understand the 'expose to the right' thing, but just in case anyone doesn't there is a good YouTube by Sean Tucker called 'Protect your highlights: A lesson for Light and Life.' I think he's a rather talented explainer of this stuff - I especially liked him showing the sensitivity of the human eye versus cameras on transparent cells and then moving them around to show how that works. It's between 6 & 7 minutes in.

 

Edited April 13, 2019 by antigallican

[jaapv]

On 4/11/2019 at 4:02 AM, cnj said:

Glad it was helpful.  Depending on how technical you want to get, there are a lot of information if you get the terms right (unfortunately, my first few searches returned some rather dated pages dating back to when CCDs were still dominating the high-end).

Some good background, although not directly related to your concern about histograms, but explains how most cameras capture colour (Fuji uses a variation on the theme, and Sigma's Foveon X3 is quite different) : https://www.cambridgeincolour.com/tutorials/camera-sensors.htm

The histogram is essentially showing what fraction of the scene (y-axis) is at a certain brightness level (x-axis, from perfectly dark "0" on the left to max exposure "255" in my example above).  A quick search brought-up B&H with a good primer: https://www.bhphotovideo.com/explora/photography/tips-and-solutions/how-read-your-cameras-histogram and Cambridge in Colour again with a little more detail: https://www.cambridgeincolour.com/tutorials/histograms1.htm

You probably want to read-up on noise and ISO invariance to understand what's happening at the bottom end on the digital spectrum.  https://www.cambridgeincolour.com/tutorials/image-noise.htm covers types of noise; there are plenty of other posts.  ISO invariance essentially means that the camera records the same noise at ISO 200 as ISO 400, all other exposure settings aside: 1/200s, f2, ISO 200 pushed one stop in post will look the same as 1/200, f2, ISO 400 as captured by the camera.  Most cameras add additional processing noise such that the ISO 200 pushed image will have additional noise than natively capturing at ISO 400.

In addition to the noise issue, if you think of the 255 steps of colour (per channel in my earlier example): if you're under exposing the whole scene, it means you're only using a fraction of the potential values.    While *over* exposing means you hit a hard ceiling and clipping with digital capture, *under* means you have less signal to describe changes in colour or luminosity and you end-up with flat, blotchy areas.  Consider what your monitor looks like if you were to switch from 16-bit colour to 256-colours (assuming that's still a setting most computers support; otherwise consider converting a JPEG with smooth colour transitions into an indexed image like a GIF, which has limited colour information).  Playing around with levels and curves in post help adjust how the captured data is interpreted, and Ansel Adam's zone system also shares some similarities.  I fear all of this is heading way off-topic though.

To bring it back to the larger discussion, correctly setting white balance before taking the picture will help both your auto-focus system, and you looking at  post-capture histograms.  This is because white-balance changes how the value in each channel is interpreted, and when correcting for the appropriate white-balance, may cause a colour shift such that a channel which *hadn't* been over-exposed before is now over-exposed.  Depending on your software, this might be recoverable and is a reason for getting white-balance right initially.  That said, I still expose to the right, develop to the left, and usually fix white balance in post 😕  AWB has come a long way.

Actually, the White Balance remark is only valid for JPG images. When shooting raw (as you should) the White Balance is defined in post-processing. It will affect your in-camera histogram, though.

[wslam]

On 4/8/2019 at 5:48 PM, leicamario said:

Indeed, the scene is hard to calculate correctly with center weight AE.

Still, the skin tones are, most of the time, too blueish or yellowish. Nobody else with that problem?

Totally agree. Metering I can deal with with EV compensation. But the color is difficult to tweak. Side by side with my M10p, the Q2 files colors consistently look like they have a yellow cast. I only shoot RAW

[Advertisement]

Hallo,

Look here: Leica Q

[PhotoCruiser]

vor 3 Stunden schrieb wslam:

Side by side with my M10p, the Q2 files colors consistently look like they have a yellow cast. I only shoot RAW

This can be your raw processor who renders the colors different like the CaptureOne problem with the too red/yellow color rendering of the CaptureOne Leica Q2 camera profile.
Try to use other RWA processor and this will change to better or worst.

As i use only RAW files i never do white balance (not even underwater), i use either AWB or set WB to the current light condition.
I used initially the system with a gray card but i feel that the final result is not that better (if so) than editing the RAW file accordingly amd
it limit my possibility to shoot quick.
BUT, i am not a studio nor a professional photographer and i am rather exyperienced with editing photos as i started with Lightroom 1
Using the correct white balance on each shot will/should reduce editing time, but i enjoy editing so i'm fine with that.

Chris