Do Curves throw you a curve?

arodney

New PDF just up that is quite useful to those who believe there's an issue with how Photoshop and Camera Raw/Lightroom's so called Master Curve are problematic:

http://www.luminous-landscape.com/essays/Curves.shtml

pathfinder

I have been following this discussion on the Yahoo Applied Color Theory group discussions, and am looking forward to Mark Segal's findings.

Thank you for posting this, Andrew. I am sure it will be of interest to many of our readers here on the Finishing School forum.

windoze

pathfinder wrote:

I have been following this discussion on the Yahoo Applied Color Theory group discussions, and am looking forward to Mark Segal's findings.

Thank you for posting this, Andrew. I am sure it will be of interest to many of our readers here on the Finishing School forum.

what i have been doing now for awhile is learning how to use ACR in CS3. So my understanding is that people who know how to use CR will often skip the contrast slider in the basic tab and go to the tone curve. I guess the parametric curve is prefered over the point curve because you get more control to affect "parts" of your image. Now regarding the parametric curve: if i got this right - the "darks" slider affects the shadows, and the "Shadows" slider affects the blacks?? That sounds a little mismatched if you ask me ( but dont ask me )
Anyway - so i can play with adjusting the "midtones" to get a better contrast to an image and is all "subjective" or is there a science to this with real numbers that we should use ( suggested values not to exceed )?

when i make adjustments using the parametric curve - you can see changes being made to the histogram on top - what am i looking to do?
Lastly, can u very briefly summarize that arfticle that Andrew refers to - its somewhat over my neural tolerances....




troy

Mark Segal

windoze wrote:

..................I guess the parametric curve is prefered over the point curve because you get more control to affect "parts" of your image. Now regarding the parametric curve: if i got this right - the "darks" slider affects the shadows, and the "Shadows" slider affects the blacks?? That sounds a little mismatched if you ask me ( but dont ask me )
Anyway - so i can play with adjusting the "midtones" to get a better contrast to an image and is all "subjective" or is there a science to this with real numbers that we should use ( suggested values not to exceed )?

when i make adjustments using the parametric curve - you can see changes being made to the histogram on top - what am i looking to do?
......................


troy

There is no technical preference between the two kinds of curves in the Tone Curve tab of CR 4.x. Use whichever you like best. The only important difference between them that I know is that the point curve can be detached from the upper left and lower right corner points allowing you to more exactly remap the data to the black and white limits of the tonal range. The attractive idea behind the parametric curve is that its use is quite intuitive for photographers who think of luminosity in terms of "zones", there being four here" shadows, darks, lights and highlights.

By switcing over the to Point Curve interface and activating the eyedropper, you can mouse over the various tonal zones of the image (depressing Control and the left mouse button on Windows) and see where these zones in the image relate to the curve. Then you revert to the parametric curve and set the three zone demarcation buttons immediately under the curve box accordingly. The area to the left of the left-most button is shadows and it should contain the shadow material - not only black. The areas to the left and right of the middle button are the darks and lights respectively, and the area to the right of the right-most button is for the highlights. By moving these buttons and the curve sliders for each of the four zones you can shape this curve any way that gives you what in your mind are the best results achievable from this tool.

What you are looking to do with these adjustments is to give the image the contrast and brightness you think is most suitable for that image. There is no cookbook, and there are no magic numbers. "Season to taste". The one general idea of which you may wish to be mindful while you do this is to avoid clipping shadow and highlight information, but even that isn't cast in concrete if it suits the desired effect of the image do allow some clipping.

pathfinder

Mark,

Welcome to dgrin.

Would you comment on the utility of ARC 4.1 in correcting milder color casts in jpgs or tiffs? I understand that jpgs or tiffs have already been "color corrected" by the 'in camera' processing, but frequently with less than stellar results in mixed lighting conditions, or with specular reflections being chosen as white points by the camera.

Is this really more harmful to the image than taking it into Photoshop and correcting it with individual R,G,B curves there? I know this question relates to your pdf on the Luminous Landscape mentioned in Andrews link above, but that was directed towards RAW files rather than jpgs or tiffs, wasn't?

Mark Segal

pathfinder wrote:

Mark,

Welcome to dgrin.

Would you comment on the utility of ARC 4.1 in correcting milder color casts in jpgs or tiffs? I understand that jpgs or tiffs have already been "color corrected" by the 'in camera' processing, but frequently with less than stellar results in mixed lighting conditions, or with specular reflections being chosen as white points by the camera.

Is this really more harmful to the image than taking it into Photoshop and correcting it with individual R,G,B curves there? I know this question relates to your pdf on the Luminous Landscape mentioned in Andrews link above, but that was directed towards RAW files rather than jpgs or tiffs, wasn't?

Hi Pathfinder,

Thanks for the welcome.

I have no experience adjusting JPG or TIFF images in a raw converter, so I cannot speak to that question from any personal experience. Nor do I know whether edits in Camera Raw to JPG or TIFF images are non-destructive in the same sense that edits to raw images are non-destructive within Camera Raw. Purely guessing here - I IMAGINE the editing process for JPG and TIFF is handled with instructions saved as meta-data and that these instructions can be amended non-destructively. Also I IMAGINE that once a JPG is saved with those instructions lossy compression impacts the image data, but I emphasize that I don't know, and I could not find any such insight in the Camera Raw CS3 User Guide or the on-line Help files. Sorry I cannot be more helpful on this point. Maybe we should do a web-search and see what we can come-up with! This is all pretty new stuff.

Mark

arodney

I've played a bit using metadata corrections for JPEGs. Its just OK and not anything like what we get using this on Raw data. Severe white balance issues are just as hard, maybe worse than using Photoshop. Where you gain an advantage is speed in applying all the edits to a group of similarly ugly images. If the image needs pretty aggressive tone and color work, most certainly cross color cast issues, I move directly into Photoshop. Note that's not something that happens often. I did have a recent experience getting an award and someone used my 5D, shooting me. I didn't notice the camera got set to JPEG from Raw until I shot about 30 images and fixed the situation. Those JPEGs under low room light at high ISO looked really poor and were very difficult to correct in Lightroom. The Raws were a piece of cake. So if you're shooting Raw, render the quality from the start in the converter. If you have JPEGs or TIFFs, render them as best you can from the start (film or setting the camera for JPEGs), if you do end up with a turd, Photoshop is the place to fix what hopefully are rare problem images.

arodney

Mark Segal has written a significant follow-up piece to his essay from last month titled Do Curves Throw You A Curve? You can download this PDF update directly here.

http://www.luminous-landscape.com/pdf/Curves2.pdf