Sparkle Test

[MiniChris]

I've noticed several members who always manage to get a good sparkle to their water highlights without overpowering the scene. The image below has the PP on the left and a section from the RAW file before any PP. I think in some ways I have it sort of, almost, maybe, but not quite spot on. Any help will be greatly appreciated.

[Dave Humphries]

I've noticed several members who always manage to get a good sparkle to their water highlights without overpowering the scene. The image below has the PP on the left and a section from the RAW file before any PP. I think in some ways I have it sort of, almost, maybe, but not quite spot on. Any help will be greatly appreciated.

Hi Chris,

I think you've just put too much contrast into it in PP, I prefer the detail in the rocks in the unprocessed section. Maybe a bit less exposure (at shooting time) would help too.

Here's one I did back in June 2009, I don't suggest it is any better though


Nikon D5000 + Sigma 18-250mm VC: 1/4000s, f/8, iso800 (2897)
F11 and click image to see at 1,024px × 641px

[Peter Ryan]

Hi Chris,

It does look a little to much contrast. You still have the sparkle in the original and I prefer the more natural colours.

[MiniChris]

Lowered all the + values I had and detuned the saturation and contrast...much more true to the natural colors I saw.

[Peter Ryan]

I agree Chris. It looks more natural now. I was wondering if you added a small increase in sharpening the sparkles might sing. I am not sure how it would go but maybe worth a try.

[MiniChris]

I agree Chris. It looks more natural now. I was wondering if you added a small increase in sharpening the sparkles might sing. I am not sure how it would go but maybe worth a try.

Already has a bunch but applied a tad more and this is what it does...

[Peter Ryan]

Hi Chris,

On my screen the reeds are sharper with more contrast in the water and sparkle. The water flow appears more dynamic.

[MiniChris]

I think it is a good start in understanding how this part of sharpening works. I am doing a ton of study in layer sharpening and in channels within the layers...sometimes I have to go back to the history menu to figure out where the heck I am..

Thanks for your continued input. This is an exciting challenge and given I've only had CS5 for a little over a month, I think (he says humbly), I've come a goodly way...but for every door I open successfully, there are six doors behind it beckoning me with siren voices..."try me, try me, try me...I don't think it will ever stop.

[arith]

Here is a quick attempt to show you how I do it;
in photoshop duplicate the background with 7 layers.
Set top layer to 10% luminosity
Klick eye on the one under it
Set multiply 100% on layer 5 and adjust Gauss blur to get sparkle
Set up to 4 Screen layers
So it looks like this with levels on top

Levels 8-1-255
Luminosity 33%
Pin Light 100% (with eye open)
Multiply 100% with Gauss 11 pixels
Screen 100%
Screen 100%
Screen 100%
Screen 100%

But Pin sometimes turns things pink, then try difference or exclusion and play around. cheers

[MiniChris]

A tiny tweak in your suggestion and changed output size a bit...but, I am of want to know why so many layers of screen @ 100%. I tried it without all four and it didn't work nearly so well...so am curious why?

[arith]

Hi Chris, I found this method if you can call it that while messing about with the Orton method; Multiply always darkens while Screen increases brightness. I noticed that a Pin light layer makes white spots lighter or at least stand out, but it is worth messing about with different layer blends and amount of Gaussian blur.
Even just putting a plain colour layer over the top and seeing what happens gives some interesting results with different blends, my favourite is Difference and Exclusion, a bit like Difference, both at say 50% opacity, they make the sky darker and haze hazier with an appropriate amount of Gaussian blur, say 70 pixels.

[MiniChris]

One method I've tried with reasonable success when working with metallic surfaces is color overlay at 100%, then changing opacity as suited. This is especially good if the original intent is to present the final piece as a B&W image.
I did a little more research into layer blends, etc., last night and will spend a considerable amount of time learning them. Thanks for the inititive to explore further.

[arith]

Cheers Chris, I found this although I can't get my head round what pin light is doing.

http://graphicssoft.about.com/gi/o.h...nd-highlights/

[arith]

Tried that method

original


Color 100%
Level on Pin Light 100% 43-0.86-211
base


Color 100%
Pin Light 100%
Multiply 100% with Gauss 9.3 pixels
Screen 100%
Screen 100%


The Color blend appears to correct pink which means I've got a lot of work to do.