Colour cast issues - can anyone help me edit this image?

[robert_k1]

<blockquote>

<p><em>Forget about CMYK mixes. Color temp perception will throw that formula right out the window as demonstrated in this thread and in other threads that correct skintone.</em><br>

<em>I have a 24,000 CMYK color patch formula guide that shows no difference in color change going from 5 to 10 percentage number increases in any of the channels except black and this applies only to skintone anyway.</em></p>

</blockquote>

<p>By "cmyk values", I do not mean working in a cmyk color space or using cmyk channels (although sometimes they do work better). Instead, I mean working in a rgb color space, but monitoring the cmyk color values in the Info Pallete. Skin tones are easier to understand, and to correct according to the *relative* cmyk color values. Read my explanations here:<br>

<a href="http://www.photo.net/digital-darkroom-forum/00Sm7r">http://www.photo.net/digital-darkroom-forum/00Sm7r</a></p>

[Tim_Lookingbill]

<p>Robert K,</p>

<p>I was also referring to working in an RGB working space as well and noting the info palette set to CMYK. I just don't agree that it is very reliable because of mixed color temps that can arise in images such as these. If shooting in a controlled studio setup maybe. </p>

<p>Examine and compare various areas of the skintone samples in the PDI color target and note the relationship between the amount of yellow compared to magenta between the very white pinkish baby and the slightly tan caucasian baby on the far right. It doesn't follow a predictable formula that can be applied to the image in this thread because you've got a white caucasian woman partially lit by the sun slightly low in the sky and bluish shade plus the blue sky canopy from above. If you put too much yellow to counter the magenta then she looks jaundice and if you pull back the magenta she looks ashen. Cyan has to be kept to a minimum or the face will look too dark.</p>

<p>What you have to do is to fool the eye's adaptation mechanism and apply a cool highlite and warm shadow effect which is why I used the Split Tone slider in ACR. This will allow a large amount of yellow to counteract the magenta while keeping the subject from looking jaundice. I don't go by CMYK formula mixes. I go by the preview and keep note of adaptation's effect.</p>

<p>You know the skintones in the PDI color target have a blue background which will affect your perception of how much yellow is really in the those skintones because of adaptation.</p>

[robert_k1]

<p>Tim, agreed with most of what you said. But skin tones' cmyk values are meant to be sampled under natural lighting. Even for a specific ethnicity/age, they are not precise or iron clat, but should be within a range. IOW, these values get you within the ballpark.<br>

Even when a face is lit under controlled natural lighting, it does not mean every square inch would have the same exact relative cmy values. It is therefore important to sample the values at the "right" spot. The DPI caucasian babies have rosy cheeks, which may result in m greater than y. I would sample their chests instead. In any event, even when the relative y and m values should vary somewhat, m should never be greater than y, and c should never be greater than either y or m. Sampling at the right point and following these rules will get you a long way.<br>

The image in this thread is certainly not shot under natural lighting. The possibility that she is sun tanned and/or has makeup on further complicates things. IOW, she has multiple skin tones on her face, arms, chest, etc. Again, I avoided sampling her cheeks or forehead, but sampled her chin, neck and arm instead. Once these were corrected to be within relative cmyk values' range, the skin tones looked "natural" without much of a sun tan. I further assumed and corrected the paint worn bare wood on the railing to be brownish or near neutral. That threw the skin tones' cmyk values off a bit, but still very much within the expected range.<br>

Yes, our eyes can easily be fooled by neighboring colors of a subject, especially after staring at the moniting for a long time. That's another reason to monitor the color numbers for sanity check. The bottomline is that after the cmyk values are corrected to be within range, more subjective tuning should not throw these values completely out of wack.</p>

[ysabel_mejia]

<p><strong>After:</strong><br>

<a href="http://i40.tinypic.com/jb3dy0.jpg">http://i40.tinypic.com/jb3dy0.jpg</a><br>

What I did was used Adobe Photoshop CS4 lightened your skin, used the smudge tool to make your skin like porcelain, whitened your teeth, made your eyes a bit more dramatic, thickened your lashes.</p>

<p>hope you like. enjoy!</p>

[robert_camillo]

<p>Thread is a little old now, but quite clearly there is some ordinary editing out there. Some of the photos are absolutely out of colour. I don't know if they have been posted in Pro photo colour space or what, But one photo goes from a strong green tinge to a strong orange. I don't have a colour profiled monitor or a wide gamut monitor but they are so varied its not funny. Personally the only photo I saws that I feel would reflect the actual skin tones would be from Greg Nixon. Most of the others were over edited or the colours were way out. Although for an edit <a href="../photodb/user?user_id=3671344">Sinh Nhut Nguyen </a> Has a pretty good looking edit</p>