Blown reds & Adobe RGB

[kevm14]

when I first used aRGB on my D70, also noticed the deep green grass
in our backyard to be more yellowish. It looked nice, more warm,
liked it - but to tell the truth, sRGB was closer to reality in
that aspect.

Now there is explanation for that as well: you can see a little
more shift of deeper greens towards yellow on the D50 aRGB Imatest

I noticed this as well when playing with color modes for shots with grass in them. Does anybody feel like we should have to even discuss this? The picture should represent the scene. Further tweaking can and may occur in PP but it shouldn't be this much of an issue!!

 

[THKTHK]

I think it is the same (at least these D50 tests perfectly match what we can experience on D70)
--
Henrik

 

[THKTHK]

I did a compare on a picture where I've found a huge difference. Only the colour space was changed in NC, nothing else (Mode II and Mode III seem to have more contrast, so a little curve adjustment would be better for them. Especially the face is too dark in Mode III)

The red coat in reality is like on ModeII, not so clean red as in Mode III or Mode I.

On the other hand, grass was nowhere near this yellow in life like on Mode II, Mode Ia seems to be spot on (but then again, I like this warm yellow cast, so finally kept aRGB)



--
Henrik

 

[kevm14]

Well this sucks! We have to choose between greens and reds??

 

[jfriend00]

This is probably the only / main reason why I will leave my D70s
for the D80 or D200 at some point. I love the D70s to death
especially the size / weight etc. I have several shots into the sun

• or similar shots where it way blown out because of the electronic

shutter chip deal taking up space on the CCD...

Yes, the D70 can suffer "blooming".

But, that characteristic of the camera has absolutely nothing to do with the discussion about color modes and blown red detail. Blooming only has to do with taking a picture of extremely bright objects (usually the sun) and it is completly solvable by using a slower shutter speed (something about 1/250) and a correspondingly smaller aperture.

Because the D70 uses an electronic shutter for certain shutter speeds, it can do 1/500 flash sync. The other cameras you mentioned cannot do flash sync that fast. It's a set of tradeoffs. I've found it very easy to learn that my shutter speed should be around 1/250th when taking pictures that include the sun and I never have issues with blooming any more. This is completely learnable if you choose.
--
John
Gallery: http://jfriend.smugmug.com/portfolio

 

[kevm14]

Ok, nothing I do seems to change anything.

The only thing that makes a noticable difference is when I select different color modes in NC. Changing the CS2 color space doesn't seem to have an effect on anything, nor did checking that box in NC.

I didn't even see a difference in the gamut warning indicator when I imported into aRGB space in CS2.

I am still able to get the results I got before - simply select Color Mode II and send to CS2. Even though I didn't have CS2 warn me when swapping between color modes, that didn't matter, because NC was sending the Mode II image as sRGB anyway. So with or without that dialog warning, as long as I select Mode II in NC, my reds are fine.

But I'm still not sure what to do with JPEGs. I suppose open each one in CS2 and save in sRGB. Or open in NC and save as sRGB. Either way I'd need to convert.

And that still doesn't address this green-yellow shift I see in grass.

At any point should I be changing my windows color profile settings? My current profile is one I made with Adobe Gamma, tailored to my monitor, and I think it's sRGB. I have many choices, including Adobe1998. But I haven't tried changing that stuff yet.

 

[THKTHK]

Ok, nothing I do seems to change anything.

It does for me!

Before selecting this "always use this profile" option, the only way I could produce sRGB output from NC was to select ModeI or ModeIII - with those nasty reds. ModeII was fine - but of course only in "Adobe-aware" applications, so I had to use PS for converting these files to sRGB.

Now I can safely select ModeII in NC, save JPEG right from NC - and the output will still be in sRGB space, but with nice reds! So no need for PS...

But I'm still not sure what to do with JPEGs. I suppose open each
one in CS2 and save in sRGB. Or open in NC and save as sRGB.

In Nikon View you can select multiple pictures, right click, "copy and resize as JPEG". If "convert images to sRGB" is selected in Preferences, I think it should work as an aRGB-> sRGB convertor even for JPGs.

--
Henrik

 

[TQ]

bookmark(nt)

 

[Frank C.]

I bet Mode II also got the color of the kid's pants correct
( nice purple color ).. right?

 

[adventure_photo]

That's why you shoot RAW. It's not even an issue if you shoot RAW.

 

[kevm14]

Before selecting this "always use this profile" option, the only
way I could produce sRGB output from NC was to select ModeI or
ModeIII - with those nasty reds. ModeII was fine - but of course
only in "Adobe-aware" applications, so I had to use PS for
converting these files to sRGB.

Now I can safely select ModeII in NC, save JPEG right from NC - and
the output will still be in sRGB space, but with nice reds! So no
need for PS...

Ohhh. See, I never had a problem making sRGB JPEGs from aRGB ones as I've been using CS2 the whole time. But yes, I see that since checking that checkbox, I am now able to make conversions with NC. I never will, of course.

But I'm still not sure what to do with JPEGs. I suppose open each
one in CS2 and save in sRGB. Or open in NC and save as sRGB.

In Nikon View you can select multiple pictures, right click, "copy
and resize as JPEG". If "convert images to sRGB" is selected in
Preferences, I think it should work as an aRGB-> sRGB convertor even
for JPGs.

And I could make a batch process for CS2 where it opens, converts and saves. Of course now I'm saving a JPEG twice instead of once, so one could make an argument for JPEG Fine instead of Normal, and now we've nearly doubled the filesize.....we're almost back to RAW again. Damn.

 

[knickerhawk]

This is a confusing topic filled with myth and mystery. Check out this link (and his follow-up) for a good explanation of the different color modes and what's going on behind the scenes:

http://www.earthboundlight.com/phototips/nikon-color-modes.html

Note in particular that COLOR MODE IS NOT THE SAME AS COLOR SPACE. Indeed, Capture and NX allow you to change color space from aRGB to sRGB for the different modes (I thru IIIa). The color space has little impact on the difference (for most images). That's because most images don't have big differences in color gamut as recorded by the camera. Color modes (especially III and IIIa) boost saturation in certain colors in an effort to mimic color-intense films such as Velvia.

Thus, it should come as no surprise that when you apply III or IIIa in Capture you'll get very saturated reds perhaps to the point of appearing to have a blown red channel or color shifts. As your example proves, however, the RAW data itself is not blown. Look at this as the (usually refundable in RAW conversion) price you pay for gettting those vivid colors in Capture.

--
My photos: http://www.pbase.com/imageiseverything/root

 

[segarci1]

That explains a lot.

Thanks.
--
sg

 

[THKTHK]

I forgot to mention, but the purple pants are also only correct in ModeII!

--
Henrik

 

[THKTHK]

it is an interesting link indeed, so as the next part of it:

http://www.earthboundlight.com/phototips/color-modes-on-computer.html

Oddly one part of it isn't similar to what I see on my pictures: "changing from Mode I to Mode II (or vice versa) causes a change in brightness and some color shift on screen. However, if you save the resulting images as jpegs and compare the results, there isn't any. Both Modes I and II are essentially the same"

The whole point of this topic here is there is a HUGE difference in reds (and somewhat less in greens) between colour modes Ia and II, even in the saved jpegs! Basically the same difference one can see already in Nikon Capture, before converting to JPG. I do not know what setting could cause his Mode I/Mode II images became the same after saving

--
Henrik

 

[FRiC]

Personally, I thought the problem doesn't seem to be too much red, but rather lack of blue and green. In Mode Ia, the lack of blue and green makes all the red into something like (255,0,0), so you just get a red blob instead of any detail.

When changing to Mode II, stronger blue and green are introduced, and you get the correct shade of red. But at the same time, other colors get changed too...

 

[echelon2004]

sometimes clips red channel too easily.

As the color modes have no effect on the sensor data and raw files, and if the electronic shutter have no impact, what causes the red channel clipping in the raw data?

I agree that it is a VERY minor problem, resulting in nothing like the images here, but still, it is maybe as much as 1/10th of a stop too much red data.

Any ideas?

Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

 

[JohannesS]

sometimes clips red channel too easily.

As the color modes have no effect on the sensor data and raw files,
and if the electronic shutter have no impact, what causes the red
channel clipping in the raw data?

If the red channel clips in the raw data that is caused by overexposure of the red channel.

Johannes

 

[echelon2004]

I should have worded that differently
Anyway, it is easily compensated for, but still rather strange I think=)
--
Anders

Some of my pictures can be seen at;
http://teamexcalibur.se/US/usindex.html

event photography and photo journalism

 

[Louie M]

Personally, I thought the problem doesn't seem to be too much red,
but rather lack of blue and green. In Mode Ia, the lack of blue and
green makes all the red into something like (255,0,0), so you just
get a red blob instead of any detail.

When changing to Mode II, stronger blue and green are introduced,
and you get the correct shade of red. But at the same time, other
colors get changed too...

Looks like Mode Ia, IIIa conversions have different models for "red" than Mode II/aRGB.

Here's an image (D50) with the original JPG (Mode IIIa sRGB). Below it are the same image converted from raw via Adobe Camera Raw into aRGB, sRGB colorspaces.
Differences (not-leveled, layer difference) on the right.



Looks like a lot of the texture/"detail" is lost by losing so much of the green/blue data.

Looks like Mode Ia, IIIa are models for color conversion into sRGB space [the sRGB tag is misleading per other articles people have posted; is space for color; seperate from the mode which dictates way colors are process into JPG]. For red, it looks like they tend to blow out the red and drop green and blue (at least in reddish colors).

ACR seems to have the same model for converting raw into aRGB and sRGB spaces. [Difference between the lower left two images is nada]. Of course, the deltas could be difference between ACR and the Nikon bodies (but based on the samples posted, that sounds unlikely). A better control would be to see the same thing done for a conversion via Nikon Capture which supposedly applies the same algorithms for conversion of raw to jpeg in Mode I,Ia,II,III, and IIIa as the camera.

It'll also be interesting to see if the D200 (mode I, Mode III) shows the same issue. Note: the D80 definitely shows the same behavior, so it's not the electronic shutter.

Yet another reason to postprocess (either shoot mode II or raw around reds), sigh.