Tuesday, September 25, 2012

In Search of Perfect Color--the New Tone Curves


I shot this a while back at the Villas Zoo during its hundredth anniversary celebration. The Young Shakespeare Players were putting on a performance of The Winter's Tale in the zoo's meeting room.  I caught these performers in a  darkish corner and ended up with a not particularly exciting ISO1600  snapshot.



But after some very quick editing I turned the snap into this.


How quick? Here is my workflow.
1-loaded the snap into RT's default mode
2-set the highlights with a single drag using curve 1 of the new dual tone-curve set in the Film-like mode


3-set the shadows with another single drag using curve 2 in the Weighted Standard mode.


4--adjusted the exposure slider to move the histogram to the left for saturated blacks in the actor's hair and costumes. 

5-dragged noise reduction sliders to roughly the settings I used for my last blog post.

6-developed and then cropped the finished image in Irfanview. Why Irfanview?  Because in my excitement to see a final image I ignored the clutter surrounding my intended image.

Total working time now that I know what to do --two or three minutes per image  Be less if I batch process the other ISO1600 snaps I shot before the play began. Not quite a one click clean up but RT is getting close. Plus the results are-- pure magic!

Two things you learn when you first take up image processing:
1-unless you have developed perfect technique, usually after decades of  professional photography, images straight out of the camera improve with an S tone curve.
2- RGB colors will shift when you apply your tone curve.

Applying a S-curve multiplies the color number in all three color channels by whatever numbers are created by the tone curve.  Consider the RGB color  75,128,175 as an example. With a curve fixed at the  grey point, 128, the red value will go down, the green will stay the same, and the blue will go up creating a different color. Since images and tone curves are unique designing a reasonable algorithm that can follow and corrects these color shifts becomes near impossible .

LAB mode is different. Tone curves in the L channel doesn't shift color. Unfortunately, correcting skin tones and the like with LAB's A and B channel curves is also near impossible.  But with the RT's four HSV tone curves, near impossible is moving in on routine.



For those not familiar with HSV terminology the hue is the color.  The chroma or saturation is the intensity of the color. The value is the brightness of the color. You can read off these numbers in RT's left history pane.

RT now has four different ways of creating a tone curve: Standard, Film-like, Saturation and Value Blending and Weighted Standard. I'll leave it to the RT programmers to explain the inner workings of the new algorithms in a promised tutorial because I suspect I'll get the details wrong. Whatever the details since you can mix  and match the methods with dual curves you don't lack for ways to create excellent images.

EDIT: For more info on the four tone curve types you will find an updated manual at https://docs.google.com/document/d/1DHLb_6xNQsEInxiuU8pz1-sWNinnj09bpBUA4_Vl8w8/edit?pli=1


First why dual curves? In my example I created an S-Curve by combining two curves. This may seem like a needless complication--my first thoughts--until I saw how well and fast it works. And if you don't care for this new fangled innovation, set curve2 to linear and ignore it.

Below is a test image that one of the developers, Michael, posted. The differences are subtle and are best viewed at original size but in the dual curve version the skin tones blend in more smoothly as the lighting slides from highlights into shadows.


I haven't tested all the combinations in any systematic way but here are some tips I picked up from issue 1529 in the google code where many test images were posted.

Two standard curves (above) work well on skin tones.

Brightening with Film-like adds saturation and gives richer colors. So does switching to a wide gamut working space like ProPhoto. (I've tested this). But if you like your colors less vivid Saturation and Value Blending is the best choice.

Weighed standard for darkening and Film-like for lightening--my example--are a good combination for people pictures.

All methods work for landscapes. Selecting a mode depends on taste

Conventional S-curves work best with Saturation and Value Blend

Whatever curve modes you settle on, this unique feature is a major addition to RawTherapee's tool kit.

This version of RT 4, build 138, can be found at: http://www.visualbakery.com/RawTherapee/Downloads.aspx   An updated manual is at: https://docs.google.com/document/d/1DHLb_6xNQsEInxiuU8pz1-sWNinnj09bpBUA4_Vl8w8/edit?pli=1  


Thursday, September 20, 2012

Teen with Fedora--A Second 'First Look'

In my original post on  RAW Therapee's new noise reduction routines I had excellent results with the image of Charlotte under Lord Shiva but far worse results with some other images. The noise reduction part still worked but thick black vertical lines appeared in the converted.images.

Turns out the noise reduction algorithm broke the image into segments called tiles. It worked its noise reduction magic on each tile individually but sometimes the tiles didn't rejoin smoothly. A tricky and hard to find bug, judging from the number of posts that appeared in the forum while they were fixing the problem.

That bug has been mostly swatted--noise reduction still doesn't work on JPGs-- but with code change set 135 we can now test and work with RAW files.  ( found at http://www.visualbakery.com/RawTherapee/Downloads.aspx  )



ISO 6400  f8.0  1/25 sec  EV -0.7

This snap was taken at Old World Wisconsin in the back of a barn during a demonstration of 19th century cow milking.  The lighting wasn't insanely difficult but as you can see from the camera settings I was pushing my D7000 to get this shot.

So how did RT do when I tried to improve the image.


YUP!  RT noise reduction works,

The noisy image to the left is without any noise reduction, just a +0.7 exposure correction to compensate for the -0.7 correction I did in the camera to move the shutter speed up to 1/25 of a second. The NR settings of the much cleaner image on the left are shown below.



I adjusted the Chrominance slider first. Use a 400 % box on something with little detail like the girl's cheek. Find a setting where all the color speckles disappear and then add a bit more.  This slider works on the A and B LAB channels that only contain low resolution color information so you are not blurring detail. With my D7000 at ISO 6400 about 2/3 of the noise was chromatic so that was a no fiddling around adjustment. Nice




For the other two sliders I picked a region with fine detail, the wood grain on the 2x4, and adjusted them to match the detail in my conversion to the detail in the jpg out of the camera. The Luminance slider controls the strength of the noise reduction. The Luminance Detail slider sets the threshold on  another algorithm that does its best to seperate real detail from random noise.  In this image Luminance 70 and Luminance Detail 50 worked well.




So where are we at now.  The left top line profile is from a RT conversion with no noise reduction.
The one below that is from the camera's jpg where I used the D7000's normal noise reduction setting.  It also is very close to the noise I measured when I used RT's default settings.

The top right profile is from my detail preserving image.  And finally the insanely low noise ISO 6400 profile below it is from when I went all out with a Luminance of 90 and a Luminance Detail of 6 to see how low noise I could go.

How bad was that image.  When I displayed it side by side with my tuned image I didn't have to pixel peek hard to see I had lost detail.  But if I had emailed it to Fedora Girl I'm sure she wouldn't have screamed "WHERE IS MY DETAIL" and immediately sent it into the bit bucket. We RT users are after the very best quality but in this world drowning in imagery, I'm afraid most viewer don't notice.

So great job RT folks and especially Emil who designed the algorithms and did the coding.