Low light photography

[Tim_Lookingbill]

<blockquote>

<p>It's interesting to see how the quality of the color fix has improved throughout this thread.</p>

</blockquote>

<p>Not sure about that, Tom. Looking at all of them including mine I'm beginning to side with Mike's assessment on this particular image.</p>

<p>I think the image should retain the red colored light with just definition (luminance) detail in the facial features brought out more. IOW just remove the red blooming like what happens in over exposed sunlit red flowers.</p>

<p>There's something that just doesn't look "appealing" about trying to make the subject appear as if it's being viewed under full spectrum white light. I see either too pinky red skin or my jaundiced results. </p>

<p>This is a tough one to fix IMO.</p>

[sketch_tbhotmail.com]

<p>Right, well how about this fix then, I ran it through CEP Cross Balance again, this time keeping the red light while removing the blooming.</p><div></div>

[tom_mann1]

<p>Tim, you have brought up an exceptionally interesting point which I would love to discuss, but don't want to derail this thread, so, in a moment, I'm going to start a new thread titled something like "Heavily gelled - missing hue and saturation variability". Since even this discussion has gotten quite technical, I'm putting that thread in the Digital Darkroom forum.</p>

<p>Tom M</p>

<p>PS - Is everyone getting the weird spam messages among the email notifications of thread activity. Arghhh!</p>

[tom_mann1]

<p>Tim (and others) - I just posted the thread that I mentioned above in "Digital Darkroom", but made the fatal mistake of using an image from another photo.net participant as my example image. </p>

<p>I thought this was allowed under the TOS of photo.net, but apparently, any image that you tweak for demonstration purposes has to be from the same forum in which it was posted, and I didn't remember where I got this one, so Jeff had to pull my thread. It will take me a fair amount of time to generate all the examples I made up based on a new image, but I'll start over in a day or two and try again. :-(</p>

<p>Tom M</p>

[Tim_Lookingbill]

<blockquote>

<p>PS - Is everyone getting the weird spam messages among the email notifications of thread activity. Arghhh!</p>

</blockquote>

<p>I'm not nor have ever gotten any spam in all my personal email box accounts related to Photo.net activity. Of course I don't have email notification turned on when posting in threads. I just go through every forum category that have threads I participate in and just look for red highlighted thread titles and/or click on "New Responses".</p>

<p>I agree, Tom, this is an interesting topic. Don't know how the OP is going to be able to apply these edits demo'ed here, so I see these discussions as more of a confirmation as to what can be done in post. The OP may get discouraged by the difficulty and just decide to toss the image and reshoot with better WB settings.</p>

<p>Ryan, that last version is pretty much what I had in mind, maybe slightly shift the hue a bit toward orange/yellow or reduce saturation in the red so it looks like a colored light that's from this earth.</p>

<p>I think why this image is so difficult compared to the green image in the linked thread I and others participated is that the green image had more variety of different color objects including neutrals to fool the eye into seeing white light and they weren't isolated spot lights.</p>

<p>The red light image has only the man's face, his blue jean jacket and a dark blue banner to tell us there's some type of spot light on the subject that doesn't put a red cast anywhere else in the dark surround. I think the surrounding elements apart from the face needs to be lightened to help the eyes distinguish that this is a red spot light, not some processing color error just on the face.</p>

[tom_mann1]

<p>Here, I've tried yet another approach to color correcting this very difficult image. This is the result of two applications of the cc50c filter in the Tiffen DFX package, followed by a cc50b (partial opacity) from the same package. I followed this with a bit of levels adjustment and noise reduction, but this time, I didn't do any spotting of the blemishes on his face.</p>

<p>What do you think?</p>

<p>Tom M</p>

<p>PS - I still haven't gotten back to re-starting the other thread that I promised, but will do so ASAP.</p><div></div>

[Tim_Lookingbill]

<p>Tom, you still lose the tertiary hues in the skin required to make skin look like it's lit by a white full spectrum light. Now it just looks like desaturated monochrome pinkish skin. It just doesn't look right.</p>

<p>The image below I used in a discussion on flatbed scanner's use of fluorescent tubes and what it does to tertiary hues in skin captured and scanned from film which supports Mikes point about the narrow spectrum of the red light's affect on the look of skin when attempting to edit back the lost tertiary hues. No matter the sophistication of edits applied, the skin ends up looking plastic or ashen.</p><div></div>

[scott_ferris]

<p>Tim,</p>

<p>I don't agree, you have completely different exposures. A very simple curves adjustment that took 2 seconds, literally, instantly vastly improves the range of hues in the "not in sunlight" image. I'm not saying it is perfect (though for 2 seconds on a tiny jpeg it isn't bad), or that images captured in a narrow wavelength environment can all be saved, but you are making too much of your example image.</p><div></div>

[Tim_Lookingbill]

<p>Scott, you misunderstood my point I made with the posted image. It wasn't to demonstrate exposure.</p>

<p>I used that image to show (bring out) tertiary color detail brought about by the effect sunlight, a full spectrum light source, has on skin tone detail as a comparison to what the narrow spectrum of the red spot light does to the skin tone of the subject in the OP's image.</p>

<p>I think I should've cropped out the darker version to prevent the confusion. The darker version was deliberately made that way by applying a curve in order to accentuate the contrast in the captured tertiary color detail in the skin shown in the bottom 400% view ACR screenshot.</p>

<p>You do know what tertiary color detail is, right? I've been noticing these type of effects brought out by hot lights, flash and sunlight compared to how this detail looks lit by fluorescent and other spiky low spectrum type lights. It's just an example to demonstrate this effect.</p>

<p>The amount of the variation in hues seen in the detail varies due to the spectra of the light source. I'm not saying you won't get skin detail with other lights. It's the variation of hues that will be limited. Gas type sodium vapor street lights are some of the worst in this regard delivering almost monochromatic results when photographing under them.</p>

[scott_ferris]

<p>Tim,</p>

<p> You are right, I did misunderstand your image. But I think it is misleading, it is just a personal rendition of what you think you see, it isn't even a comparison of what you actually get for others to judge. Not as valuable as the image initially implies, though, of course, you don't misrepresent it.</p>

<p>As for asking if I am familiar with tertiary colours, yes, when referring to RGB and azure (210º), violet (270º), rose (330º), orange (30º), chartreuse (90º) and spring green (150º), I am very familiar with them.</p>

<p>But tonal sensitivity is not limited to capture light spectrum but also film emulsion colour layer sensitivity, digital sensor bayer array colours etc. One of the most powerful tools in the digital arsenal is the ability to generate custom camera profiles to tweak the individual colour response curves. Using the X-Rite Passport has speeded up my colour important work (art reproduction) immensely.</p>

<p>It is a very interesting area, one that like most aspects of photography can be analyzed to death on an ever steeper and never ending curve to gain slightly more accuracy.</p>

<p>This narrow spectrum live venue lighting is an area I have a great interest in, I shoot some live shows and in the last couple of years budget lighting schemes have utilized a vastly inferior lamp and light board techniques that have made photography far more challenging.</p>

<p>Here is a before of one of my red band light captures. Even more dramatically bad than John's initial post! How is one supposed to get publishable images from these lighting schemes?</p><div></div>

[scott_ferris]

<p>Here is the image as sold.</p><div></div>

[tom_mann1]

<p>Tim, Scott, etc. - I think you might be interested in a new thread I just started on a closely related topic:<br>

http://www.photo.net/digital-darkroom-forum/00ZjDp</p>

<p>Cheers,</p>

<p>Tom M</p>

[link]

<p>I used a technique not yet mentioned here (I think :).</p>

<p>Just looking at this image, the red light means that almost all the detail on the face is in the red channel. I used the channel mixer (as an adjustment layer) and then curves to move the detail from the red channel into the blue and green channels.</p><div></div>

[tom_mann1]

<p>BG, you are right, nobody had suggested using the channel mixer. It's an interesting idea & certainly produced results that moved the image in the right direction.</p>

<p>I obviously don't know the exact algorithm that Adobe used in the channel mixer, but I strongly suspect that this approach is equivalent to converting the face to B&W, and then just colorizing the resulting BW image, say, by placing a color adjustment layer above it and setting the blend mode to color. </p>

<p>The problem with approaches like this is that they provide no point-to-point variability in hue or saturation over the face, only luminosity changes, ie, the exact issue that Tim L correctly raised earlier in the thread.</p>

<p>Thanks again for reminding us of the channel mixer. I must confess that I have not been using it as much as I used to, so a nice application of it always is appreciated.</p>

<p>Cheers,</p>

<p>Tom M</p>

 

[Tim_Lookingbill]

<p>Scott, that's not the definition of tertiary colors. The following page will tell you what I'm talking about...</p>

<p>http://profeditor.wordpress.com/2010/03/24/40/</p>

<p>In reference to skin as shown in the my 400% close crop, the varying hues of beige, oranges, tans are considered colors next to each other on the color wheel that are mixes of primary and secondary colors of slightly varying luminance and hue. Tertiary colors are alternating transitional colors that make up an overall dominant color (in this case skin color) that add depth and richness. A type of richness that's also desirable when capturing precious gems and stones as well as expensive dyed fabric and vibrant mult-colored bird plumage.</p>

<p>Lights of varying spectra from full spectrum (sunlight) to flash, the red light in the OP's sample all the way to sodium vapor street lights will reduce in varying degrees according to their spectra makeup the amount of different hues seen in these types of objects that may add or take away from this richness and depth (i.e. the monochrome/plastic look). </p>

<p>The available colors seen in the face of the drummer in the corrected version of your posted image shows only about three colors. At least that's all I see. It looks posterized. It doesn't matter if it's the correct overall skin color (which is all that can be done to the OP's image), there's still no depth and richness because it was lost at the point of capture by the narrow spectrum of the lights used.</p>

 

[scott_ferris]

<p>Tim,</p>

<p>That is exactly the definition of tertiary colours.</p>

<p>If you mix two primaries you get a secondary, if you mix a primary and a secondary you get a tertiary, if you mix RGB primaries and secondaries in equal amounts you get tertiary colours in the tint ranges I laid out. That your link chooses to call the tertiaries red-orange, red-purple etc just displays a lack of common colour name labeling. Countless colour pros from artists, house painters, auto paint suppliers, printers etc all know yellow-green as chartreuse, because that is what it is.</p>

<p>My point with the drummer was that John is not alone in fighting live venue lighting "advances", I know it is not a fully toned portrait, though it did sell. In keeping with Tom's new thread it also demonstrates the additional capabilities of RAW processing over trying to improve jpegs, there is no way that you could get as smooth a final image from the jpeg.</p>

<p>Here is a closer crop of the drummers face, for sure it has problems, but as was also pointed out earlier in the thread, if it is shot under dramatic stage lighting why try to neutralize it 100%? I needed the image to look dynamic and retain the feeling of the night, but the original capture has even more problems.</p><div></div>

[tom_mann1]

<p>Tim, I don't think the problem is limited to, or caused by loss of tertiary colors. While the tertiaries can be a symptom, IMHO, as per my other thread, it's much more general than that. It's due to a loss of variability in saturation and hue, no matter where around the color wheel this shows up.</p>

<p>Just my $0.02,</p>

<p>Tom M</p>

[scott_ferris]

<p>I agree with Tom 100%. The problem is response sensitivity and the elimination of hues and subtle tonality, the deficits are not limited to any range of hues in particular. </p>

<p>Funnily enough my above image dramatically illustrates Tom's ideas on RAW superiority for this very use. At that shoot I was working for the band, I had a friend shooting for the local press, he is a jpeg shooter for the fast turnaround times he needs but on this occasion it tripped him up, he couldn't save his images from dramatic blocking up and complete loss of tonality and hues.</p>

[Tim_Lookingbill]

<p>We're talking past each other. </p>

<p>You can redirect relevance of my points all you want. I stand by what I said as a legitimate observation about the effects of the spectrum of light's have on skin tone hues. I know what I see. </p>

<p>If not then why do so many photographers complain about the color rendering of fluorescent and LED lights which aren't as gelled as the images by both the OP and Scott's illustrate. I'm explaining why they see what they can't put their finger on about these types of lights. It has to do with the reduction of tertiary colors we humans rely on to tell the difference between tan colored plastic from tan colored skin.</p>

<p>You redirecting heavily gelled lights and their affects on the appearance of skin as something different from the points I put across on tertiary colors isn't going to change my mind. </p>

<p>I'm done.</p>

<p> </p>

[scott_ferris]

<p>Tim,</p>

<p>Nobody is trying to change your mind, or your point, that doesn't mean you expressed it in a particularly accurate way. You are the one hung up on a word you used out of context, the effects we all agree are seen in these conditions have got nothing to do with tertiary anything. </p>

<p>We are all talking about the same thing just in a different way. People are seeing a lack of hue and tonality, I say it like that, you insist on saying they are losing tertiary colors, even though the lack of primary and secondary hues can be just as bad.</p>

<p> </p>

[tom_mann1]

<p>Hi Tim - There was absolutely no intent on my part (and almost certainly not on Scott's part either) to redirect relevance. I think your observations about the loss of the tertiary colors in the vicinity of skin tones are right on the mark.</p>

<p>In terms of hue variability within a section of image, I think you are effectively saying that hue variability has been reduced from 30 or more degrees on each side of the primaries (ie, where the tertiaries are) to a much smaller value, and this variability is what allows us to distinguish plastic of a given color from skin of the same average color. I couldn't agree more, and this is exactly what my other post was about. </p>

<p>I just generalized that conclusion to say that the reduction in hue variability almost certainly also occurs everywhere around the color circle, not just in the vicinity of skin tones, but since we are so sensitive to skin tones, we can detect it most easily there.</p>

<p>Cheers,</p>

<p>Tom</p>