M8 vs DMR, very different colors in Capture One

[newyorkone]

Advertisement (gone after registration)

Pascal...perhaps try another colorspace.

[woodyspedden]

Hmmm Pascal , this is something i do not like at all. The DMR looks great but i agree there is a magenta cast going on here.

 

Guy and Pascal

 

Perhaps we should try another converter altogether and see what happens. I am going to try Silkypix since my experience with that converter shows it to have the best color rendition, other than C1 of all. You can download a free 30 day trial which could lead us to determine whether it solves the magenta cast.

 

Woody Spedden

[pascal_meheut]

First, for zanydave, I'm using C1 3.7.6 and said so in my first post.

 

Second and this is important, I do not have the same color cast using the flash, despite what I said earlier. With the flash, the DMR & M8 files are quite similar.

 

Third: I played with other profiles as suggested: R9 one, Flexcolor one, custom ones... Some changes, less saturation but still a cast.

 

 

 

Finally, those of you who want to play with the files can download the M8 one at:

 

http://www.photofaq.eu/temp/L1000337.DNG

 

and the DMR one at:

 

http://www.photofaq.eu/temp/L1002246.DNG

 

Also, here is are the M8 and DMR files when using flash, in this order:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[bradreiman]

doe anyone know which square in gm chart should be magenta? which should be red? in the first two examples dmr and m8 the m8 actually looks right to me. at least on my macbook pro. i have a good eye for color, its my carreer as a video engineer.

[bradreiman]

also-can you shoot a greyscale chart? for me a better indicator of color balance. only because its what im used to. others will im sure disagree....b

[adan]

In this chart, pure magenta should be 4-I (Four-Eye) - fourth row just below the SG logo.

 

Also the center-most squares (5-6, E-J) and the border squares ARE gray scales.

 

The hot cyans and magentas in Pascal's original image looks a lot like what I see in my betaware 0.2 M8 shots. Which makes me think the profile the camera is embedding in the DNG file is off a bit.

 

Pascal - have you tried doing your own RAW calibration?

[pascal_meheut]

Advertisement (gone after registration)

Red should be G4, magenta I4.

 

And here is a Kodak Gray Scale shot by the M8 under the problematic halogen light. Sorry for the blur but the tripod vibrated.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[pascal_meheut]

Pascal - have you tried doing your own RAW calibration?

 

I generated some profile from the ColorChecker chart in ProfileMaker but nothing more. Anyway, I do not understand why does the DMR has the right colors under both lights, i.e. flash and halogen whilst the M8 is correct only when using flash.

[adan]

Anyway, I do not understand why does the DMR has the right colors under both lights, i.e. flash and halogen whilst the M8 is correct only when using flash.

 

My understanding is that the RAW software makers (or at least the sophisicated ones) build different color profiles for a camera for different white balance settings.

 

My guess is that you are getting a bad embedded profile for daylight or auto white balance (due to some programming error in the camera, or some other reason) but a good embedded profile for use with "flash" white balance.

 

When I first got my Sony R1, ACR only had beta support for it, and the colors were off (too much magenta in blue skies, among other things). So I built and saved my own "R1" profile in ACR. Once Adobe updated to full support for the R1 in ACR 3.4, their profile improved. But I recent reprofiled the camera (partly because I am considering switching to the ProPhoto color space as my primary working color space), and found I could get even better colors by tweaking the RGB hues and saturations (my blue saturation is about minus-50 from Adobe's 'official' settings, and most of the other settings benefited from a little shifting)

 

From which I've decided that I will profile my M8 myself (and any other RAW-capable camera I get). What I consider to be good color is not necessarily what the Adobe or Phase One software guys consider to be good color.

[pascal_meheut]

Sure but once again, why is the DMR ok in both lights with a very close sensor and the same Raw program, i.e. C1 ?

[J_Brittenson]

Is it a pink cast, or oversaturated pinks, or reds that are too strong? If it's just oversaturated pinks during processing rather than a CFA sensitization issue, then that's pretty easily fixed through C1's Color Editor to edit the stock input profile (a copy of it). Looking at the GMB CC it doesn't look like a cast per se.

[adan]

"Sure but once again, why is the DMR ok in both lights with a very close sensor and the same Raw program, i.e. C1 ?"

 

Perhaps because the DMR's profiles are correct for both lights, while the M8 is (or may be) embedding an incorrect profile for daylight and a (more) correct one embedded for flash.

 

The two (or more) profiles written into the M8 are independent from one another - one may be screwed up while the other is correct. They are simply data files - and data can get corrupted, or may have been incorrectly installed in the first place.

 

I.E. we are talking about 4 calibrations here - 4 separate data files stored in the two cameras:

 

DMR daylight color calibration file = OK

DMR flash color calibration file = OK

M8 daylight color calibration file = corrupted or simply inaccurate

M8 flash color calibration file = OK

 

I hate to make too many assumptions here, about a new camera. The issue LOOKS like a problem with color profiling or calibration, to me - but it may be something else entirely, or it may be just bad firmware in your camera only, or a certain batch of cameras.

 

If you make inkjet prints, and one of your printing profiles gets corrupted, you will get bad prints when you use that particular profile, and good prints using other papers and other profiles.

[ho_co]

Pascal--

Probably another wild hare, but in a thread referring to the S Reid M8 review at http://www.leica-camera-user.com/digital-forum/8430-sean-reids-m8-part-3-review-3.html#post83042 the poster remarks:

"In the M8 Part 3 high ISO comparisons, it appeared to me ... that the M8 color shifted toward magenta in some areas (e.g. potato) as the ISO was increased and as compared to the R-D1...."

 

Was the M8 picture made at high ISO? Would a lower ISO look better?

 

The blacks/whites/greys look neutral to me, and I think "color cast" implies an overall deviation from proper.

 

There is clearly something amiss--even the wood background has a funny color in the M8 shot. I'm sure you've checked your white balance and color space between the two cameras.

 

Good luck.

 

--HC

[bradreiman]

not trying to be contrarian at all but on my calibrated macbook pro all elements look correct including the blurry greyscale which to me is the best indicator of proper color balace. maybe it is a colorspace issue. are you guys on macs?

[andybarton]

Howard

 

Pascal was using the lowest possible ISO. (100 or 160)

 

Is it possible to set both cameras at the same ISO?

[ho_co]

not trying to be contrarian at all but on my calibrated macbook pro all elements look correct including the blurry greyscale which to me is the best indicator of proper color balace. maybe it is a colorspace issue. are you guys on macs?

Bradley--

Do you not see a difference between the M8 and DMR images at the top of the thread?

 

I agree, the greyscales look good; but it looks as if one part of the chart has very funny colors; and the two charts definitely don't look the same on my (Mac) monitor.

 

Very strange.

 

--HC

[ho_co]

Howard--Pascal was using the lowest possible ISO. (100 or 160)

 

Thanks, Andy. I should have re-read. Just grasping at straws and hoping that some simple hook will correct the problem.

 

--HC

[bradreiman]

hoco-to me both magentas look quite similar. the dmr chart looks like more chroma saturation or either the m8 is minus chroma. the magentas both look a little bit either -blue or +red but look the same, only difference being saturation. to me magenta usually looks less flat or glows a bit. if you reduce blue or add red the "glow" becomes less prevalent. im thinking of smpte bars for television. maybe it doesn't apply here....b

[pascal_meheut]

Thanks for your ideas guys, the are food for my thoughts. Changing ISO changes nothing, well at least in the "reasonable" range. Changing the camera WB settings neither as it could expected in RAW mode so there is embedded profile or something.

 

However, I did another test and noticed that the cast increases when the TC goes down.

Here is the example in this order: 2350K, 2450K, 2800K, 6100K. The first one are with my big halogen, one with the old lamp, then with a new one. 2800K is with a small desk halogen and 6100K is daylight.

 

I've noticed that for other cameras, C1 has often several profiles to handle the TC. Maybe the same is needed for the M8.

 

If someone here as a M8 and can do the test and shoot with a 2300K light, that would be interesting.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[bradreiman]

hoco-i take back my reply above i was looking at the flash charts. the normal charts i find the magentas very similar but the m8 chart is just more chroma saturated. greys look pretty neutral in both but m8 very slight elevated green low end. the last charts make me think this whole thing is a color temperature thing. magenta cast could be a case of to low kelvin value hence warmer and less blue. in the last charts to me all the magentas,greys,blacks and whites look really nice just different levels of contrast and chroma saturation. unless my monitor is calibrated wrong (i calibrated recently) i think they all look fantastic. maybe colorspace. maybe im calibrated to different colorspace on macbook pro but dont know. id rather see a daylight shot than a chart or a full greyscale "chipchart" to decide if there's a magenta cast though. i just dont see it.