Value of using Color Calibration in a non-studio, set and forget environment

[tippingpoint]

Lately I have been learning a bit around color calibration using something like X-Rite ColorChecker. I ended up looking into this topic as through my photographic journey over the last 5ish years, I came to realize that color rendition is very important to me. In fact, it has been one of the main causes for my obsessive gear flipping, always trying to find the perfect camera.

I came across a source which was talking and demonstrating how you could simply use color calibration to actually get great colors, minimizing the need for tweaking white balance and turning up the vibrancy, saturation, hue sliders. In this particular case it was about making Sony colors better. Speaking of color preferences from all the cameras I've had so far, here's my subjective ranking: Panasonic S1/S5 > Leica CL > Ricoh GRIII > Nikon Z6 > Sony A7c.

To the actual topic and question: I understand that the benefits of using a ColorChecker passport fully articulate if you take a reference shot of the passport for each scene, ie. for a certain lighting condition. This sounds like it's best used for studio work or a fixed "professional" setting, in which you fulfill a certain job. Now I am a hobbyist, mostly shooting street, travel, architecture, landscape kind of things. This means I am mostly out and about with my camera for many hours with highly fluctuating scenes and rarely a possibility to accurately take a reference shot in order to calibrate the colors for every shot.

Is there still value in getting a ColorChecker and approximate better colors by using it once in a set and forget manner? More specifically, is there value in creating a few profiles in distinct settings which replicate most common use cases (bright sun light, evening, indoor) and using these as better alternatives to the in-camera or Lightroom profiles?

Hope someone can relate to this and give some insight around this topic.

Thanks!

 

[Erik Kaffehr]

Lately I have been learning a bit around color calibration using something like X-Rite ColorChecker. I ended up looking into this topic as through my photographic journey over the last 5ish years, I came to realize that color rendition is very important to me. In fact, it has been one of the main causes for my obsessive gear flipping, always trying to find the perfect camera.

I came across a source which was talking and demonstrating how you could simply use color calibration to actually get great colors, minimizing the need for tweaking white balance and turning up the vibrancy, saturation, hue sliders. In this particular case it was about making Sony colors better. Speaking of color preferences from all the cameras I've had so far, here's my subjective ranking: Panasonic S1/S5 > Leica CL > Ricoh GRIII > Nikon Z6 > Sony A7c.

To the actual topic and question: I understand that the benefits of using a ColorChecker passport fully articulate if you take a reference shot of the passport for each scene, ie. for a certain lighting condition. This sounds like it's best used for studio work or a fixed "professional" setting, in which you fulfill a certain job. Now I am a hobbyist, mostly shooting street, travel, architecture, landscape kind of things. This means I am mostly out and about with my camera for many hours with highly fluctuating scenes and rarely a possibility to accurately take a reference shot in order to calibrate the colors for every shot.

Is there still value in getting a ColorChecker and approximate better colors by using it once in a set and forget manner? More specifically, is there value in creating a few profiles in distinct settings which replicate most common use cases (bright sun light, evening, indoor) and using these as better alternatives to the in-camera or Lightroom profiles?

Hope someone can relate to this and give some insight around this topic.

Thanks!

I would be skeptical.

The way I would think, it may make sense to use a grey card to get WB right.

Making a full profile may be error prone, but may make sense in 'bad light'.

As long as the spectrum of the illuminant is by and large continuous, my guess would be that just getting WB right is quite enough for good reproduction of color.

Best regards

Erik

 

[D Cox]

Lately I have been learning a bit around color calibration using something like X-Rite ColorChecker. I ended up looking into this topic as through my photographic journey over the last 5ish years, I came to realize that color rendition is very important to me. In fact, it has been one of the main causes for my obsessive gear flipping, always trying to find the perfect camera.

I came across a source which was talking and demonstrating how you could simply use color calibration to actually get great colors, minimizing the need for tweaking white balance and turning up the vibrancy, saturation, hue sliders. In this particular case it was about making Sony colors better. Speaking of color preferences from all the cameras I've had so far, here's my subjective ranking: Panasonic S1/S5 > Leica CL > Ricoh GRIII > Nikon Z6 > Sony A7c.

To the actual topic and question: I understand that the benefits of using a ColorChecker passport fully articulate if you take a reference shot of the passport for each scene, ie. for a certain lighting condition. This sounds like it's best used for studio work or a fixed "professional" setting, in which you fulfill a certain job. Now I am a hobbyist, mostly shooting street, travel, architecture, landscape kind of things. This means I am mostly out and about with my camera for many hours with highly fluctuating scenes and rarely a possibility to accurately take a reference shot in order to calibrate the colors for every shot.

Is there still value in getting a ColorChecker and approximate better colors by using it once in a set and forget manner? More specifically, is there value in creating a few profiles in distinct settings which replicate most common use cases (bright sun light, evening, indoor) and using these as better alternatives to the in-camera or Lightroom profiles?

Hope someone can relate to this and give some insight around this topic.

Thanks!

I would be skeptical.

The way I would think, it may make sense to use a grey card to get WB right.

Making a full profile may be error prone, but may make sense in 'bad light'.

As long as the spectrum of the illuminant is by and large continuous, my guess would be that just getting WB right is quite enough for good reproduction of color.

Best regards

Erik

I think most, perhaps all, current models of camera can give good colour, at least for general photography, without the user doing anything elaborate. Reproduction of artworks is more critical.

Setting WB is often a matter of artistic judgement: if the photo was taken in warm light (such as on a winter afternoon), do you keep a warm bias in the image, or do you correct it to match midsummer daylight ? Or partially correct ?

Older cameras are a different matter. Fortunately I had saved raw files of the majority of photos that I took with the Sony NEX-5N. As I still have the camera, it was easy to make a custom input profile for Adobe Camera Raw by taking a photo of the Color Checker. Re-processing the NEX-5N raw files with this profile gave a marked improvement in colour quality.

I also saw an improvement in the colour of DNG files from the Sigma sdQH, though this was less marked.

Don Cox

 

[Sdiver2489]

The x-rite profiler software is capable of making a “dual illuminate” profile which allows for a general purpose DNG profile to be made that will give you a better starting point than the adobe standard profile. This seems to be what you want.

 

[ggbutcher]

Lately I have been learning a bit around color calibration using something like X-Rite ColorChecker. I ended up looking into this topic as through my photographic journey over the last 5ish years, I came to realize that color rendition is very important to me. In fact, it has been one of the main causes for my obsessive gear flipping, always trying to find the perfect camera.

I came across a source which was talking and demonstrating how you could simply use color calibration to actually get great colors, minimizing the need for tweaking white balance and turning up the vibrancy, saturation, hue sliders. In this particular case it was about making Sony colors better. Speaking of color preferences from all the cameras I've had so far, here's my subjective ranking: Panasonic S1/S5 > Leica CL > Ricoh GRIII > Nikon Z6 > Sony A7c.

To the actual topic and question: I understand that the benefits of using a ColorChecker passport fully articulate if you take a reference shot of the passport for each scene, ie. for a certain lighting condition. This sounds like it's best used for studio work or a fixed "professional" setting, in which you fulfill a certain job. Now I am a hobbyist, mostly shooting street, travel, architecture, landscape kind of things. This means I am mostly out and about with my camera for many hours with highly fluctuating scenes and rarely a possibility to accurately take a reference shot in order to calibrate the colors for every shot.

Is there still value in getting a ColorChecker and approximate better colors by using it once in a set and forget manner? More specifically, is there value in creating a few profiles in distinct settings which replicate most common use cases (bright sun light, evening, indoor) and using these as better alternatives to the in-camera or Lightroom profiles?

Hope someone can relate to this and give some insight around this topic.

Thanks!

I would be skeptical.

The way I would think, it may make sense to use a grey card to get WB right.

Making a full profile may be error prone, but may make sense in 'bad light'.

As long as the spectrum of the illuminant is by and large continuous, my guess would be that just getting WB right is quite enough for good reproduction of color.

Best regards

Erik

Agree with Erik, the single most significant thing you can do to get good color is to get white balance right, and having a white reference in at least one image from the scene is the best reference, even as just the starting point for artistic departures.

Now, I've messed with all sorts of camera profiling in the past few years, and IMHO the most significant upgrade to your camera profiling is to make a LUT profile from measured spectral data for your camera. If you can find a SSF dataset someone else has measured for your camera, this can be quite easy to do, but if not, it's a royal pain in the anterior regions. After a lot of messing with spectroscopes and special software I finally have such for each of my cameras, and even then I only use those profiles for taming extreme colors.

Probably the most significant upgrade to your camera profiling that is relatively easy to do is to make a dual-illuminant DCP for your camera, and using a raw processor that can properly apply such a profile. If you can't find an already-made DCP, you can make one by shooting a ColorChecker under daylight and then under a 3200K light (stay away from those LED things, they don't usually provide a continuous spectrum) and using software like Lumariver or dcamprof to make a proper dual-illuminant DCP. Oh, and if you have camera SSF data, you can do the above without target shots, just two profiles from the SSF data with the respective illuminants.

In fact, for the largest flexibility and control in making camera profiles I commend Lumariver. It'll do all the above and also give you controls over skewing the matrix/LUT that goes in the profile to do things like film emulation.

A lot to digest, but you asked...

 

[tippingpoint]

The x-rite profiler software is capable of making a “dual illuminate” profile which allows for a general purpose DNG profile to be made that will give you a better starting point than the adobe standard profile. This seems to be what you want.

That’s it! Totally fine with a somewhat better starting point for colors, even if it doesn’t mean it’s 100% hands off for changing conditions. Thanks for the answer

 

[tippingpoint]

Lately I have been learning a bit around color calibration using something like X-Rite ColorChecker. I ended up looking into this topic as through my photographic journey over the last 5ish years, I came to realize that color rendition is very important to me. In fact, it has been one of the main causes for my obsessive gear flipping, always trying to find the perfect camera.

I came across a source which was talking and demonstrating how you could simply use color calibration to actually get great colors, minimizing the need for tweaking white balance and turning up the vibrancy, saturation, hue sliders. In this particular case it was about making Sony colors better. Speaking of color preferences from all the cameras I've had so far, here's my subjective ranking: Panasonic S1/S5 > Leica CL > Ricoh GRIII > Nikon Z6 > Sony A7c.

To the actual topic and question: I understand that the benefits of using a ColorChecker passport fully articulate if you take a reference shot of the passport for each scene, ie. for a certain lighting condition. This sounds like it's best used for studio work or a fixed "professional" setting, in which you fulfill a certain job. Now I am a hobbyist, mostly shooting street, travel, architecture, landscape kind of things. This means I am mostly out and about with my camera for many hours with highly fluctuating scenes and rarely a possibility to accurately take a reference shot in order to calibrate the colors for every shot.

Is there still value in getting a ColorChecker and approximate better colors by using it once in a set and forget manner? More specifically, is there value in creating a few profiles in distinct settings which replicate most common use cases (bright sun light, evening, indoor) and using these as better alternatives to the in-camera or Lightroom profiles?

Hope someone can relate to this and give some insight around this topic.

Thanks!

I would be skeptical.

The way I would think, it may make sense to use a grey card to get WB right.

Making a full profile may be error prone, but may make sense in 'bad light'.

As long as the spectrum of the illuminant is by and large continuous, my guess would be that just getting WB right is quite enough for good reproduction of color.

Best regards

Erik

Agree with Erik, the single most significant thing you can do to get good color is to get white balance right, and having a white reference in at least one image from the scene is the best reference, even as just the starting point for artistic departures.

Now, I've messed with all sorts of camera profiling in the past few years, and IMHO the most significant upgrade to your camera profiling is to make a LUT profile from measured spectral data for your camera. If you can find a SSF dataset someone else has measured for your camera, this can be quite easy to do, but if not, it's a royal pain in the anterior regions. After a lot of messing with spectroscopes and special software I finally have such for each of my cameras, and even then I only use those profiles for taming extreme colors.

Probably the most significant upgrade to your camera profiling that is relatively easy to do is to make a dual-illuminant DCP for your camera, and using a raw processor that can properly apply such a profile. If you can't find an already-made DCP, you can make one by shooting a ColorChecker under daylight and then under a 3200K light (stay away from those LED things, they don't usually provide a continuous spectrum) and using software like Lumariver or dcamprof to make a proper dual-illuminant DCP. Oh, and if you have camera SSF data, you can do the above without target shots, just two profiles from the SSF data with the respective illuminants.

In fact, for the largest flexibility and control in making camera profiles I commend Lumariver. It'll do all the above and also give you controls over skewing the matrix/LUT that goes in the profile to do things like film emulation.

A lot to digest, but you asked...

Haha, yes that’s something to get my head around. Thank you for this detailed answer. I think I’ve got some homework to do

 

[Entropy512]

Probably the most significant upgrade to your camera profiling that is relatively easy to do is to make a dual-illuminant DCP for your camera, and using a raw processor that can properly apply such a profile. If you can't find an already-made DCP, you can make one by shooting a ColorChecker under daylight and then under a 3200K light (stay away from those LED things, they don't usually provide a continuous spectrum) and using software like Lumariver or dcamprof to make a proper dual-illuminant DCP. Oh, and if you have camera SSF data, you can do the above without target shots, just two profiles from the SSF data with the respective illuminants.

I think the biggest benefit here is the consistency - some raw processor software hides parts of the profile from the user, and in many cases, profiles are tweaked by the software provider to emulate the camera's JPEG behaviors.

Generating a dual-illuminant DCP that either does not have an embedded tone curve, or has a "common" one selected by the user (dcamprov/lumariver allow either of these use cases easily) gives you easy camera-to-camera consistency.

I know you're aware of this, but for the OP: https://rawpedia.rawtherapee.com/How_to_create_DCP_color_profiles describes the process used to generate the built-in color profiles used by RawTherapee. For some cameras I've just used the existing DCP profiles in RT, for others I generated my own using that process (and then submitted them to RT for inclusion, RT's A6300 and A6500 profiles come from test shots I submitted. A7M4 has been submitted but not yet included due to key developers being absent at the moment. Many users in the Mi Sphere 360 camera users group use my color profiles to compensate for the camera's built-in DNG profile metadata being severely broken.)

A side observation is that for certain highly saturated colors (such as things lit by blue LEDs), a common matrix profile can often lead to some pesky math errors (like negative luminance), so both Adobe and dcamprof will generate a matrix that is "desaturated" that is then followed by an HSV LUT that will undo the desaturation built into the matrix.

 

[Entropy512]

The x-rite profiler software is capable of making a “dual illuminate” profile which allows for a general purpose DNG profile to be made that will give you a better starting point than the adobe standard profile. This seems to be what you want.

I've found the X-Rite software to frequently give inconsistent results. dcamprof, while slightly harder to use if you're not comfortable in a command line environment, gives much better results and also makes it easier to handle things like highly distorted reference images (needed if your camera has a fixed fisheye lens...)