Camera Raw

[Giovanni Frescura]

Hello
I discussed with a "guru" about the use of camera Raw.
I read that is better make little change in Camera Raw and all the other in Photoshop because the struments of pphotoshop are better.
He said me that it is wrong because all must be in Camera Raw..
What is right ?
Hehave Nikon and I canon so could b some difference because he use Nikon Capture ?
He used a lot upoint tecnique ..but I no founf it in cs&
Thanks for your help
Giovanni

[Graeme Sheppard]

Generally I think the best practice is to do what you can in Camera Raw first because it is working on the raw digital file. Here you can change white balance, exposure, brightness, contrast etc as much as you want. This should be a very quick process and will get you to point which is essentially an optimised 'straight-from-the-camera' type picture.

When you have the photo generally looking how you want it, move to Photoshop and make final adjustments - I guess you may want to fine tune levels or do local adjustments.

Personally I use Lightroom, though, but as no one else replied yet I thought I'd give my thoughts.

[Giovanni Frescura]

Thanks
My problem is that sometimes like use adjustament only to the subject (for example less noise, morelight...) and for mjake it is necessary use levels or make a selection: I think it isn't possible in Camera Raw.
Is it true ?
Ciao
Gio

[Roger Clark]

I do minimal tone corrections in ACR. I set the black point to zero, and I do aberration corrections and will correct exposure if underexposed and highlight recovery if overexposed. Beyond that I go to photoshop for fine corrections. I make corrections by selections on portions of the image (dodge and burn) rather than global tools in ACR. Beyond the simple (raw) corrections, it should matter little if in ACR or another image editor, as long as one works with at least 16-bit data. The only advantage to working in the raw converter is if it were operating in 32-bit floating point (at least), and I believe ACR is only 15 bit integers (yes I said 15-bit). Photoshop internally is 15-bit when working on a 16-bit image, as it uses signed integers.

Roger

[John Chardine]

Hi Gio- The later versions of Adobe Camera Raw have a very powerful Adjustment Brush, which allows local adjustment of many of the "Basic" sliders such as Exposure, Contrast, Highlight, Shadows etc., and sharpness and noise reduction. You can add many "control points" to the image each of which contain a different set of local adjustment values. You can change the softness and size of the brush, and flow/density as well. And all this is stored in the xml file so next time you open the raw image, the control points and settings are all there.

I think there are advantages to doing as many adjustments as possible in ACR because you are working on the RAW data.

[Roger Clark]

I think there are advantages to doing as many adjustments as possible in ACR because you are working on the RAW data.

Hi John,
All those adjustment tools are applied after the raw data has been demosaicked, so it is no longer raw data. At that point, it would be no different whether in ACR or another program, as long as one maintains 16 (actually 15)-bit processing. Digital camera output is nearing the point of needing more than 15-bit processing, which means our file sizes will double in the move to 32-bit.

Roger

[Dan Brown]

Hi Gio- The later versions of Adobe Camera Raw have a very powerful Adjustment Brush, which allows local adjustment of many of the "Basic" sliders such as Exposure, Contrast, Highlight, Shadows etc., and sharpness and noise reduction. You can add many "control points" to the image each of which contain a different set of local adjustment values. You can change the softness and size of the brush, and flow/density as well. And all this is stored in the xml file so next time you open the raw image, the control points and settings are all there.

I think there are advantages to doing as many adjustments as possible in ACR because you are working on the RAW data.

To add here, using ACR is non-destructive (as is Lightroom, DPP and NX2). A nice feature with ACR and PS is, if you do your adjusting in ACR and then open the file as a smart object, you can continue editing with PS and at any time, you can double click on the the BG layer smart object and you will be returned to ACR, where you can continue to edit there also. Then when you are done, open again into PS and the changes will be applied to all layers that are copies of the original SO layer. One thing to add, I have found that sometimes opening as SO crashes PS though?

[John Chardine]

Hi John,
All those adjustment tools are applied after the raw data has been demosaicked, so it is no longer raw data. At that point, it would be no different whether in ACR or another program, as long as one maintains 16 (actually 15)-bit processing. Digital camera output is nearing the point of needing more than 15-bit processing, which means our file sizes will double in the move to 32-bit.

Roger

OK Roger, so it sounds like in ACR at least, the RAW conversion is a two-step process: 1. Demosaic the RAW data, then 2. Apply all the ACR slider values to the demosaiced data before handing off to Photoshop. I guess I was under the impression that the RAW development took place once you hit the Open image button in ACR.

[Mike Milicia]

Hi John,
All those adjustment tools are applied after the raw data has been demosaicked, so it is no longer raw data.

Hi Roger,

I have never been able to find any documentation as to the exact order of things in the Camera RAW processing pipeline or even a differentiation as to what is done before of after demosaicing. Everything I have read indicates that this is Adobe proprietary and has never been disclosed. Can you provide a pointer to any resources that talk about this?

I did find this comment from Jeff Schewe on another forum which seems to imply there are definite advantages to doing things in Camera RAW :

"In the case of Camera Raw/Lightroom Thomas and the Camera Raw team simply won't say EXACTLY how the processing pipeline works, why should they? It's proprietary information and disclosing it may help a competitor.

Regardless of what is being done pre or post demosaicing, everything in CR/LR _IS_ being done in linear gamma except the final output to a color space. As a result, the processing pipeline is optimal and the functions superior in the pipeline to opening the image in Photoshop for post color encoded processing."

[Graeme Sheppard]

The bottom line for me is that my image manipulation skills are good, but not brilliant and I change my mind about things. Anything I can do in a non-destructive, non-linear fashion makes my life a happier, simpler place.
So I use LR as it seems to do its magic pre-demosaicing, post-demosaicing and even post-jpeging, choosing the optimal point to perform each process on the image even if I apply things in the wrong order.

The last post made me think though, in a workflow using Camera Raw and PS, surely any adjustments that are best made before the white balance/colour space is finalised should be in camera raw. Other things could be done in either really to the photographer's choice. Optimal or not, I doubt there would be any perceivable difference to the image created.

[Roger Clark]

Regarding the order of raw processing, a look at the DCRAW source code (available online) will tell a lot. Except for Bayer demosaicing, my scientific workflow on calibrating spacecraft sensors is similar (I don't have any Bayer sensors on spacecraft). I added in a line to say when the gamma correction gets applied to the dcraw source code. I used the raw fiile from this month's raw processing exercise.

dcraw -v -T -6 -C 0.999 1.001 -n 3 roger.clark.night.scene.C45I1088.CR2
Loading Canon EOS-1D Mark IV image from roger.clark.night.scene.C45I1088.CR2 ...
Wavelet denoising...
Scaling with darkness 2049, saturation 15280, and
multipliers 2.306266 1.000000 1.406737 1.000000
Correcting chromatic aberration...
AHD interpolation...
Applying gamma curve...
Converting to sRGB colorspace...
Writing data to roger.clark.night.scene.C45I1088.tiff

What this means, and filling in between the lines:

read the raw image
separate the red, green, and blue channels into three separate images.
run a wavelet noise suppression algorithm on the red, green and blue images.
Subtract the dark offset (the 2049 value) from the red, green and blue images.
scale the red and blue channels (this is the color balance).
chromatic aberration: this is scaling the size (number of pixels) in the red and blue channels.
AHD interpolation: this is the Bayer demosaicing.
apply the tone curve (the gamma curve)
convert to sRGB color space (default--I could have specified Adobe RGB).
Write result.

In photoshop, exposure would be applied to the scale step after dark subtraction. Color balance is applied to the scaling step. Highlight recovery would be applied before the scaling step. Other steps, like contrast, brightness, hue, saturation, vibrance should be applied after the tone curve.

Roger

[gareth thomson]

Following up on what Dan said earlier - ACR is non-destructive, and if you set it up to open the file in Photoshop as a smart object, then from w/in Photoshop you can always re-edit your original ACR settings - and that's a heck of nice feature. My box has never crashed when processing smart object files, but it does get a bit slower.