RAW to Photoshop

[pwnage101]

I like to use photoshop for post-processing rather than raw converters such as ACR. I believe it gives me more control when making significant alterations of a photo. I heavily use masks, spead each step across multiple adjustment layers, and manipulate individual color channels. However, I get the feeling that I am losing image data when I bypass ACR (set everything to 0, linearize curves, and open in photoshop in 16-bit).

Here is an example of the typical workflow I follow. The following two images are final product, and layers I used to make final product, respectively.



There doesn't seems to be any problem with this particular example, but this image was very properly exposed. What if it wasn't? Does the image lose data by bypassing ACR as I've done? I am actually quite the stubborn perfectionist

And just a quick question: Why doesn't photoshop have luminance curves? RGB curves apply a formula to the three channels independently, right? I wish there were built-in curves that maintained the ratios between red green and blue within each pixel.

[Colin Southern]

Hi Troy,

Does the image lose data by bypassing ACR as I've done?

Yes - very much so. The general rule is "everything you CAN do in ACR you SHOULD do in ACR"; the main reason being that whilst in ACR the data is in linear gamma. ACR doesn't "apply adjustments" per sec - it tailors the conversion ... very powerful stuff.

Why doesn't photoshop have luminance curves?

It does - but you have to use LAB mode (which is is what Photoshop uses internally anyway). LAB takes a bit to get your head around, but once you do, you won't want to go back to RGB in a hurry!

[Pawel]

I've just recently started my adventure with photography, but I'm also trying to find best possible workflow with RAW and photoshop. I'm trying to do as much as I can in ACR and only then go to PS. I've also read that it's better to import DNG file as a smart object - that way you can always go back to original file in ACR, do some changes and they will automatically appear in your .psd file. But since I didn't played much with smart objects I cannot tell you exactly what advantages/disadvatages they have.

[Colin Southern]

But since I didn't played much with smart objects I cannot tell you exactly what advantages/disadvatages they have.

The advantage is that you can go back to ACR - adjust the image - and have the changes dynamically update in Photoshop; the disadvantage is that you can't do any destructive edits while it's a smart object.

So it's an excellent tool, but not one that needs to be used on every occasion. Normally, I don't have to open images as smart objects because I can get the image looking pretty right whilst still in ACR, and the very first thing in my Photoshop workflow (even before capture sharpening) is "dust bunny" removal, which is destructive (although you can even do that in ACR now).

[proseak]

Questions, really - what is ""dust bunny" removal"? I did google it, and found nothing that seemed relevant...

My other question is about the nature of RAW files - I'm fairly new to this, and have thus far just converted straight to 16-bit Tiff. What data am I losing? I've experimented with developing a few at different settings, ie "for" highlights, and "for" shadows but not found any advantage - it was either "there" in the Tiff, or not. Or have I missed the point?

proseak

[Colin Southern]

Questions, really - what is ""dust bunny" removal"? I did google it, and found nothing that seemed relevant...

Hi Peter,

"Dust Bunnies" are just "cute slang" for sensor dust spots (little furry things!)

My other question is about the nature of RAW files - I'm fairly new to this, and have thus far just converted straight to 16-bit Tiff. What data am I losing?

Potentially, quite a lot. RAW captures are linear gamma whereas when they're converted to TIFF (or any other 16 bit format) the gamma is converted to around 2.2 by default, meaning that there is a lot of compression & stretching of tonal ranges. When the image is manipulated further in Photoshop then levels are lost due to compression (resulting in loss of detail) whereas in other areas levels are stretched to the point where they may be so far apart that the steps between the levels become visibly obvious (banding).

Which comes down to a fundamental "rule" ... basically, "anything you CAN adjust within the RAW converter you SHOULD adjust within the RAW converter". This is something that a lot of people don't fully understand ... when you move the sliders in ACR you're not "adjusting the image" per sec - what you're actually doing is "tailoring the conversion" (it makes ALL of the adjustments in a single pass) - thus it's a lot less lossy.

In a "normal" image you probably won't see any difference - but when you have a tricky image (where the dynamic range is pushed to the limit) and you need to make big changes - then that's where RAW really comes into it's forte.

Does this help?

PS: This book is an absolute bible for anyone shooting RAW.

[proseak]

I can see that I have a lot of experimenting to do; I'm just making the transition from film - the quality of digital images that i had seen really, really put me off - Jpegs with blown highlights, poor colour(both appeal, and apparent depth) left me feeling that digi was quick, cheap, and convenient, with the emphasis on cheap. Until I discovered the difference RAW can make...even at 16-bit, I'm getting a dynamic range that would have needed hand-printing from film. And without the chemicals, at that.

Having glanced at your work, it's plain that I have a lot to learn.

Regards

Peter

[RonH]

Clarity Slider in ACR
This adjustment seems to be very effective and is highly rated in Scott Kelby's 'Elements 8' book. Even though our experts views are 'don't sharpen in ACR', is this particular slider commonly used within our CinC circle to add

[Dave Humphries]

Clarity Slider in ACR
This adjustment seems to be very effective and is highly rated in Scott Kelby's 'Elements 8' book. Even though our experts views are 'don't sharpen in ACR', is this particular slider commonly used within our CinC circle to add

I stick by "don't sharpen in ACR", that's because I don't consider "Clarity" as sharpening, although it is mask based and arguably works in a similar way, it is far more akin to the Local Contrast Enhancement (LCE) use of USM.

The "Real World" book on ACR/CS3 gives a workflow in PS to recreate Clarity (p74), it is USM; 15% at 100 radius, plus some messing about with layers and blending modes (I wish I understood those).

I do use ACR's Clarity, but not excessively, or even creatively , relying on 'doing' LCE later in my workflow with USM.

Have you tried Vibrance Ron?

Cheers,

[RonH]

... I do use ACR's Clarity, but not excessively, or even creatively , relying on 'doing' LCE later in my workflow with USM.

Have you tried Vibrance Ron?

Cheers,

G'day Dave ... playing CinC during Easter Thanks for your input.
I have just started to try out Clarity on some of my pics. Around 40% seems to offer some hope, picture dependant. Vibrance also played with a bit but I need to get out and about for some good RAW shots to try out all these new toys once the 'ball and chain around my ankle has been unlocked But our weather is at last on the improve.

Whilst I only have Elements 8, I am most impressed with what can be acheived. A heck of a lot still to learn and practice though and very time consuming .... "Hurrah, hurrah, ouch, I felt that dearest"

[Michael Bailey]

Just a quick thought on the question of Luminance curves...I believe you can achieve the same effect by using a regular RGB curve layer and applying the Luminance blending mode. As far as I've seen, if it's not exactly the same, it's awfully close.

MB

[Colin Southern]

Clarity Slider in ACR
This adjustment seems to be very effective and is highly rated in Scott Kelby's 'Elements 8' book. Even though our experts views are 'don't sharpen in ACR', is this particular slider commonly used within our CinC circle to add

Hi Ron,

Sometimes I use it on landscape images, but on portraits it's generally a train wreck. IMO it's probably just "a lazy man's content/creative sharpening"

[Colin Southern]

Just a quick thought on the question of Luminance curves...I believe you can achieve the same effect by using a regular RGB curve layer and applying the Luminance blending mode. As far as I've seen, if it's not exactly the same, it's awfully close.

MB

Hi Michael,

Your probably right. With LAB colour you'd typically access the L channel via a curves layer - but - that's usually a relatively minor adjustment ... the far more significant adjustment at that point is to steepen the A & B curves, and when you steepen one more than the other (eg to drive blues and yellows apart more than greens and magentas) then you get something that's VERY difficult (or impossible) to emulate in RGB mode. So manipulating the L channel is a relatively minor thing (well it is in my workflow anyway).

[Michael Bailey]

Hi Michael,

Your probably right. With LAB colour you'd typically access the L channel via a curves layer - but - that's usually a relatively minor adjustment ... the far more significant adjustment at that point is to steepen the A & B curves, and when you steepen one more than the other (eg to drive blues and yellows apart more than greens and magentas) then you get something that's VERY difficult (or impossible) to emulate in RGB mode. So manipulating the L channel is a relatively minor thing (well it is in my workflow anyway).

Well, Colin, you expose my ignorance about Lab, so I happily defer to your greater knowledge.

MB

[Colin Southern]

Well, Colin, you expose my ignorance about Lab, so I happily defer to your greater knowledge.

MB

Hi Michael,

I think a lot of people avoid (or don't know about) LAB; it's just another Photoshop tool often left unexplored. Often it's not needed - often its no better and no worse than RGB - but sometimes it's possible to do things with it that just can't be done easily in other colourspaces.

People know RGB because it's how the eye sees - what they often don't know is that although the eyes see with an RGB "mechanism" the brain actually processes that information in an "LAB" (or "AB" anyway) kind of way, using colour oponantcy (sp?).

Dan Margulis' book "The Canyon Conundrum" is the industry standard reference text if you're interested in learning more (he also has a good video presentation on www.kelbytraining.com.

[RonH]

Hi Ron,

Sometimes I use it on landscape images, but on portraits it's generally a train wreck. IMO it's probably just "a lazy man's content/creative sharpening"

Thanks for your input Colin. Lazy man's way and also maybe the way of the L plater ... I fall into both camps so might be smart to persist auto learning this before I move onto the manual gearbox