M Color

[wlaidlaw]

Advertisement (gone after registration)

For those who do not have the facilities, I am in the process of uploading a bright sunlight CC24 Gretag Macbeth DNG taken at 5000ºK fixed WB and the corresponding DCP profile generated by ColorChecker Passport. This was taken with a Zeiss Contax Vario Sonnar 28-85mm, which is a very high contrast lens. My upload speed to Dropbox is quite slow and it says it will take 4+ hours to upload the 29MB DNG file. I will post with link when complete.

 

Wilson

[wlaidlaw]

Dropbox was lying about the upload time it is done. You will find here https://www.dropbox.com/sh/n8ktnjd30daqcsc/erfebIraHe two .DNG files for north light and sunlight and two matching .DCP files. The Northlight was at 7500ºK and the Sunlight at 5000ºK

 

Wilson

[dwbell]

Thanks for that, great stuff.

 

Can you juuuuuust squeeze a tungsten one in Wilson?

[01af]

This was taken with a Zeiss Contax Vario-Sonnar 28-85 mm, which is a very high-contrast lens.

It's a very flare-prone lens, apparently ... at least at the 28 mm setting. So the test chart shot is useless, as some of the chart's patches are contaminated with flares which are not entirely colour-neutral.

 

So—can you re-do the sun-lit shot with a less exotic lens, please? And yes, a tungsten shot taken at 2,850 K (or thereabouts) would be useful, too, to build a dual-illuminant profile.

[wlaidlaw]

It's a very flare-prone lens, apparently ... at least at the 28 mm setting. So the test chart shot is useless, as some of the chart's patches are contaminated with flares which are not entirely colour-neutral.

 

So—can you re-do the sun-lit shot with a less exotic lens, please? And yes, a tungsten shot taken at 2,850 K (or thereabouts) would be useful, too, to build a dual-illuminant profile.

 

Olaf,

 

It was taken at 85mm. I think the lens is actually equally flare prone at 85 as it is at 28mm. The problem was also where the sun was, the white painted walls of my house and that my proper Contax W1 sun hood is in the post from Japan. I bought a tulip hood for it while the lens was on the way to me from Italy, having totally forgotten in the years since I last had a Vario Sonnar, that the front rotates with focus - doh!

 

I deliberately wanted to use a very high contrast lens. I will therefore use my 50mm ZM Planar instead, for which I have a decent lens hood. Tungsten is a bit of a problem, as I don't have any tungsten illumination in my UK house. The nearest I can get is halogen which will be around 3100º - 3250ºK - is that OK?

 

Wilson

[wlaidlaw]

Thanks for that, great stuff.

 

Can you juuuuuust squeeze a tungsten one in Wilson?

 

See my reply to Olaf,

 

Wilson

[wlaidlaw]

Advertisement (gone after registration)

Tomorrow, if it is still sunny. I will go and use the walls of our stables as a mount for the Gretag Macbeth chart, which are dark brown wood, so less light being thrown back at the camera. That should solve the flare issue together with the 50 Planar, which is very flare resistant anyway.

 

Wilson

[wlaidlaw]

If you go to the same public Dropbox as in my post above, you will now find a Halogen DNG file and its matching DCP profile. Ho-hummmm! I set the M240 to 3000ºK, which should be about right for halogen GU10 50W bulbs. Now I don't know if the batch I am using at the moment are very yellow or the way that the M240 handles a 3000ºK colour temperature is wrong but the DNG is extremely yellow. To correct to neutral, I have to lower the colour temperature to 2550ºK, which is some way below normal tungsten temperature. All a bit odd.

 

Wilson

 

PS my old Sixti colour temperature meter seems to finally died so I cannot say for certain what the real colour temperature falling on the CC24 chart was.

[wlaidlaw]

I have a bit of dilemma trying to redo the sunny CC24 charts. What do I do for colour temperature/white balance? Normally I would say just set the number on the camera to 5000ºK but after last night, I now have little confidence that the camera is actually setting to the number you dial in. When you look at the image in either ACR or LR, the colour temperature number is showing up quite a bit lower (about 10%) than the number you have dialled into the M240. Auto white balance is out as you have no idea what the camera is going to set and it is not repeatable. Similarly if you use one of the named WB settings, you don't know what you are getting.

 

So should I set it manually? I do have an Expodisc, which is a handy device for when you are out in the boondocks but I am not sure how accurate it is. I could use a grey card but mine is down in France and I would have to drive over to Park Cameras or Marrutt to buy a new one. Then the issue is, what colour temperature is the M240 generating from either the Expodisc or grey card? Is it correct?

 

OK guys what is the best thing to do?

 

Wilson

[Peter H]

Wilson, does it produce consistent differences in recorded colour temperature for identical shots compared with the M9, say, or another camera with which you're familiar? Could you use another camera as a means of calibrating the M? Or is this too simplistic?

[dwbell]

I have a bit of dilemma trying to redo the sunny CC24 charts. What do I do for colour temperature/white balance? Normally I would say just set the number on the camera to 5000ºK but after last night, I now have little confidence that the camera is actually setting to the number you dial in. When you look at the image in either ACR or LR, the colour temperature number is showing up quite a bit lower (about 10%) than the number you have dialled into the M240. Auto white balance is out as you have no idea what the camera is going to set and it is not repeatable. Similarly if you use one of the named WB settings, you don't know what you are getting.

 

So should I set it manually? I do have an Expodisc, which is a handy device for when you are out in the boondocks but I am not sure how accurate it is. I could use a grey card but mine is down in France and I would have to drive over to Park Cameras or Marrutt to buy a new one. Then the issue is, what colour temperature is the M240 generating from either the Expodisc or grey card? Is it correct?

 

OK guys what is the best thing to do?

 

Wilson

 

As far as I'm aware it doesn't matter what colour temp you set for the chart image. It's irrelevant. Nor does it have to be WB corrected before you create a profile from it. What is very important is that none of the colour channels are clipped, so preferable to take an ambient meter reading on the card with a light meter rather than let the camera pick an exposure.

 

Others should jump in if I'm incorrect on the WB question -> 0laf?

 

edit: Not that Wilson couldn't have done this himself, but here's an M240 Dual Illuminance profile, need some test dng's for it now?

https://dl.dropboxusercontent.com/u/624443/dwb_LEICA%20M%20%28Typ%20240%29%20Dual.dcp

[wlaidlaw]

Wilson, does it produce consistent differences in recorded colour temperature for identical shots compared with the M9, say, or another camera with which you're familiar? Could you use another camera as a means of calibrating the M? Or is this too simplistic?

 

Peter,

 

The problem is believing the figures you get shown. Obviously other than setting a particular numeric colour temperature on the M, you have no idea of what colour temperature the camera is choosing. Then you don't know if the camera is setting the colour temperature accurately. I think it may not be. I feel the images are too warm. I think that means that it is setting the colour temperature too high. Now when you import say the very yellowish halogen DNG CC24 image I took last night, with the camera set to 3000ºK, ACR shows the image as taken at 2750ºK. Capture One shows it as taken at 2917ºK - so what is it? ACR does not have a WB correction eye dropper but Capture One corrects it to 2425ºK. By eye and looking at the RGB numbers, ACR would need to adjust to 2350ºK.

 

Wilson

[dwbell]

<<Subjective alert!>>

 

Using the linked dual luminance profile above, and Jono's Beta camera winter "dog walk" etc

images I'm seeing very well tamed reds, slightly "bolder" blues compared to what I'm used to with the M9 and a remaining yellow sensitivity as per M9. Especially on skin tone.

 

I wouldn't call them unusable by any means. I'd need to learn the subtleties (that's my job as a photographer, right?) but as a starting point I'm not too concerned. I'd welcome more dng files with skin in to play with though.

 

This MacBook Pro hasn't been calibrated for over a month and it's very bright in the office at the moment.

[wlaidlaw]

As far as I'm aware it doesn't matter what colour temp you set for the chart image. It's irrelevant. Nor does it have to be WB corrected before you create a profile from it. What is very important is that none of the colour channels are clipped, so preferable to take an ambient meter reading on the card with a light meter rather than let the camera pick an exposure.

 

Others should jump in if I'm incorrect on the WB question -> 0laf?

 

edit: Not that Wilson couldn't have done this himself, but here's an M240 Dual Illuminance profile, need some test dng's for it now?

https://dl.dropboxusercontent.com/u/624443/LEICA%20M%20%28Typ%20240%29%20Dual.dcp

 

I am never 100% convinced by the concept of a dual illuminance profile. It is very easy to do on the x-Rite Color Checker Passport application I have. You just drag the two DNG's for the two illuminances to the program and select dual illuminance profile.

 

I will get down to making some more CC24 shots after I have had lunch. Been baby sitting for my 20 month old grandson all morning (so now exhausted). My table says early afternoon sunlight 5000ºK, which is what I will set the camera at.

 

Wilson

[dwbell]

I'm very convinced by Dual Illuminance profiles. Sensors simply respond differently throughout the spectrum. I've had a lot of success with all my cameras profiling in this way.

 

I'll do shot specific profiles if it really must be bang on. And often compare them to the dual generic during post. The differences are usually very subtle indeed.

[wlaidlaw]

I'm very convinced by Dual Illuminance profiles. Sensors simply respond differently throughout the spectrum. I've had a lot of success with all my cameras profiling in this way.

 

I'll do shot specific profiles if it really must be bang on. And often compare them to the dual generic during post. The differences are usually very subtle indeed.

 

How far apart do you feel you should choose your illuminances? I can understand if say you have sunlit and dull day or say tungsten and fluorescent. However say sunlit and tungsten does not seem to make sense to me.

 

Wilson

[01af]

... the concept of a dual-illuminant profile. It is very easy to do on the X-Rite Color Checker Passport application I have. You just drag the two DNGs for the two illuminants to the program and select dual-illuminant profile.

Which illuminants is your X-Rite Color Checker Passport application asking for? The Adobe DNG Profile Editor wants 6,500 K and 2,850 K ... but isn't too sensitive when the actual colour temperatures are off by a few hundred Kelvin.Still, 5,000 K is a bit low for a good dual-illuminant profile. Of course, it'll make a perfect single-illuminant profile for shooting at, say, 4,000 K - 7,000 K.

[dwbell]

Sunlight shot is 5200K, the halogen one 2600K.

[dwbell]

How far apart do you feel you should choose your illuminances? I can understand if say you have sunlit and dull day or say tungsten and fluorescent. However say sunlit and tungsten does not seem to make sense to me.

 

Wilson

 

Not sure I understand? As 0laf points out you want 6500 and 2850 ideally. For my use I take bare summer mid day sunlight and my standard indoor tungsten shooting.

 

See;

X-Rite ColorChecker Passport Dual Illuminant Profiles - Part 1

for a short description.

 

Edit - Ok, I think I know what you're getting at, correct me if I'm wrong? A dual illuminant profile takes the warm 2850 and the cold 6500 values returned from the chart and creates an interpolated (inverse correlated color temperature method) profile specific to the WB on each shot. So using just one DI profile for your camera if you set WB to 2850 it will use the settings directly recorded in your test shot, similarly at 6500. Somewhere in between it will interpolate a response between the two. It probably isn't as accurate as taking a chart shot and creating a new profile for every occasion, as I mention above though, it's very very close and a whole lot easier! Hope that helps?

[Advertisement]

Hallo,

Look here: Leica M

[wlaidlaw]

Which illuminants is your X-Rite Color Checker Passport application asking for? The Adobe DNG Profile Editor wants 6,500 K and 2,850 K ... but isn't too sensitive when the actual colour temperatures are off by a few hundred Kelvin.Still, 5,000 K is a bit low for a good dual-illuminant profile. Of course, it'll make a perfect single-illuminant profile for shooting at, say, 4,000 K - 7,000 K.

 

Olaf,

 

X-Rite seems to leave it up to you. They say you can have a narrow range or broad range dual illuminance profile. I have just done a narrow range (sunlit - north light) to produce a dual illuminance general use daylight profile (uploading at the moment to Dropbox together with the DNG's and single illuminant DCP's, taken at 5000ºK with a Zeiss ZM50 Planar at f4) Interestingly ACR again shows these consistently as taken at 4600ºK. The files should be there in around 15 minutes.

 

Wilson