M8 vs DMR, very different colors in Capture One

[pascal_meheut]

I've noticed some strange colors with the M8, especially in red/violet and skins tones. I shoot some pictures of my son with both cameras, using ambient light and flash and it was obvious.

So to be sure, I also shot a GretagMacbeth ColorCheck DC chart, on tripod, both cameras in Raw, lowest possible ISO.
I used C1 LE 3.7.6 to process the raw files, did the WB on the same gray spot. And the difference is still here, both in the ambient light and in the flash shots: oversaturated reds.

Here is the DMR file :

[pascal_meheut]

And now, the M8 file. As you can see, the background is redish as the skin tones. Look at the M2 to M4 and D7 to J7 patches for instance.

Anybody has an explanation? I had such oversaturated reds at first when using ACR with the DMR, 1 year ago but then C1 came and "nailed the colors" as Guy Mancuso said.
Now, ACR still gives horrible colors with the M8 files but C1 seems not as good as it could be too.

[fuchs]

Hello Pascal, did you notice the same oversaturation in jpeg files? Or does the in-camera processing give better color results off the box?

Best
Ed Albesi

[jrgeoffrion]

Originally Posted by pascal_meheut

I've noticed some strange colors with the M8, especially in red/violet and skins tones. I shoot some pictures of my son with both cameras, using ambient light and flash and it was obvious.

So to be sure, I also shot a GretagMacbeth ColorCheck DC chart, on tripod, both cameras in Raw, lowest possible ISO.
I used C1 LE 3.7.6 to process the raw files, did the WB on the same gray spot. And the difference is still here, both in the ambient light and in the flash shots: oversaturated reds.

Here is the DMR file :

Pascal, maybe try the M8 profile from v3.7.5 to see if it solves the magenta issues.

[pascal_meheut]

I just did, following your advice. No differences or very minor ones.

[robsteve]

Have you tried Lightroom?

[marknorton]

There's a comment on the C1 3.7.5 release note that they've turned up the red for the M8 in this version. Maybe they've gone too far?

[pascal_meheut]

To answer the questions:

- no, I did not tried in jpeg. I could but I do not shoot jpeg
- yes, I tried LightRoom and ACR 3.6. Same magenta cast, but even stronger.

[guy_mancuso]

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.

[DaveEP]

It looks like P1 have some work to do. It is to be expected. I have noticed some odd colour casts from C1Pro before.... I wonder if some one there has a badly calibrated monitor

[newyorkone]

Originally Posted by DaveEP

It looks like P1 have some work to do. It is to be expected. I have noticed some odd colour casts from C1Pro before.... I wonder if some one there has a badly calibrated monitor

But he is seeing the magenta cast in EVERY raw software tool that he has tried...not good.

What I'm wondering is:

1. Is this just a bad sample? Need to get others with early M8's to test as well.
2. How much of this is hardware and how much of this is software?

This reminds of the magenta cast issue that plagued early D1X bodies.

[guy_mancuso]

Pascal re process the M8 file with the generic R9 DMR camera profile if you would and post that

[DaveEP]

Originally Posted by newyorkone

But he is seeing the magenta cast in EVERY raw software tool that he has tried...not good.

What I'm wondering is:

1. Is this just a bad sample? Need to get others with early M8's to test as well.
2. How much of this is hardware and how much of this is software?

This reminds of the magenta cast issue that plagued early D1X bodies.

This is true! Sorry, I missed that part. Bummer.

[sreidvt]

I've been using 3.75 and I haven't noticed any strange color casts. The pictures I posted in the reviews looked good to me, as have the many others I've processed. Could it be that specific camera?

Cheers,

Sean

[jrgeoffrion]

Originally Posted by pascal_meheut

I just did, following your advice. No differences or very minor ones.

[zanydave]

I think this problem may have already been addressed in Version 3.7.6 which is now available for download.

Dave

[osera]

Maybe Pascal could trade around a raw file with another M8 owner, just to test his hardware/software pathway. Not to many M8's now, but lots more over the next few days.

[Jamie Roberts]

Originally Posted by guy_mancuso

Pascal re process the M8 file with the generic R9 DMR camera profile if you would and post that

An even weirder request: re-process the file using the IMACON profile for the DMR. Ok, I know this will be messed in the shadows, since it won't have CO's special bit mojo for gamma, but try it anyway (if you had FlexColor installed for the DMR, it will be there--gotta love the way C1 lets you experiment).

I mention this because I did this to a couple of the "bootleg" M8 files and was quite surprised by how much better Caucasian skin tones, especially, looked. No more magenta.

(FWIW, I'm not prone to doing suff like this, but I am prone to accidents. I was changing the output profile and I actually applied the Hassy input profile by accident, then, why not? the DMR default. An immediate--and better--difference in red / yellows, and greens!! They don't have that "yellowy" look CO often creates)

I'd love to see how the GM chart comes out with that, lacking my own M8 right now...

Anyone want to pay Magne Neilson of ETC to develop some better generic profiles?

[zanydave]

Version 3.7.6

Improved camera support

•

Leica M8
Colors:

Red saturation has been increased a little to ensure a deeper rendering of pure red
colors in e.g. sunshine

Skintone reproduction has been optimized by decreasing the saturation and making
a more yellowish look
Film curves:

Film extra shadow: Has been optimized to give a better overall look

Dave

[newyorkone]

Originally Posted by osera

Maybe Pascal could trade around a raw file with another M8 owner, just to test his hardware/software pathway. Not to many M8's now, but lots more over the next few days.

Good idea...

Also, it is a stretch but perhaps the computer or something else is a factor. Maybe Pascal could provide the DNG file to see what others come up with.