bioneural.net

Well, not for dummies—but for non-professional photographers, amateurs, or prosumers (as you like). This article represents my evolving understanding of the options you have during raw conversion with ACR 3.1, and during post-processing in Photoshop CS2 (following nearly 3 years of shooting JPEG, ignoring colour management, and tinkering with Photoshop 7). I like a "one product does it all" solution. I descibe a workflow that will hopefully make sense for most images—assuming that like me you want to work on one at a time, to edit once, and your interest is in capturing what's there and not "faking it". I'll make changes as I discover what does or doesn't work, and acquire new knowledge (maybe even skill?). Your feedback and suggestions would be much appreciated. This is a fairly long article, but there's a graphic summary at the end!

Capture and camera settings

I decided on a Nikon D70 set to raw + JPEG basic, a SanDisk 1GB Extreme III Compact Flash card, and an iPod photo 60GB with iPod Camera Connector for in-the-field storage (also works with my wife's Olympus C-70Z and my old Sony DSC-F717). As with my previous camera, the only "rule" I try to adhere to when shooting is this one:

...correct exposure means keeping the highlights as close as possible to blowing out, without actually doing so. Some photographers refer to this concept as "Expose to the Right" because you want to make sure that your highlights fall as close to the right side of the histogram as possible.

The problem with my last camera, a Sony DSC-F717, was that the electronic viewfinder was so poor I couldn't actually see the highlights (the D70 has an optical viewfinder).

Customisable image settings on the D70 are:

• Sharpening (Auto, -2 to +2, None);
• Tone compensation (Auto, -2 to +2, Custom curves);
• Colour mode (Ia is the default and mapped to the sRGB colour space; IIIa is also mapped to the sRGB colour space but is optimised for nature/ landscapes; II is mapped to the Adobe RGB colour space and is the best choice);
• Saturation (Normal, Moderate, Enhanced);
• Hue (+9° to -9°).

However, you "only" have to think about ISO. According to Schewe:

...when you're shooting raw, the only camera setting that will have a significant impact on the raw capture is the camera's ISO setting. The ISO setting changes the analog-to digital conversion process by amplifying the signal from the sensor. Virtually all other camera settings that you set on the camera will result only in an informational tag that conversion software may or may not use when processing the raw conversion.

That said, if you want to shoot raw + JPEG, all of the above settings will have an impact on the JPEG (but not raw) image. Since the iPod photo can store but not display raw images, it makes sense for me to store a small iPod-viewable JPEG along with each raw image. Whether you change any of the above settings thus depends on whether you plan to do anything with the JPEG images.

For additional thoughts on customising settings on the D70, see here.

Colour management basics

The concepts of gamut, colour space, device profiles, working space, image profiles, and colour management systems are discussed here.

Selecting images in Bridge

1. Copy from CompactFlash (or iPod photo) to a folder on the hard drive, named by date and topic e.g. "2005-09 Syria". This folder will serve as my primary raw archive and will be periodically backed up along with the rest of my data to an external hard drive (I use SilverKeeper). The raw files are much smaller than TIFFs and I can go back and "redevelop" them should I need to, since after saving to JPEG re-editing (and thus re-compressing) will cause cumulative damage.
2. Open Bridge to browse the above folder and use ratings to indicate images to be converted; apply the filter to show only those for conversion.
3. Optionally apply IPTC Core metadata. Choose File > File Info... and enter your data (e.g. name, website, etc.). Click the button at the top right to Save Metadata Template... Thereafter select your image(s) in Bridge and chose Tools > Append Metadata to select your template.

   

   Note: The D70 records whether the photo was taken in landscape or portrait mode, information which should see it display the right way up in Bridge automatically. If it doesn't you can rotate the image thumbnail to the correct orientation (left or right) at this point; it will then open in ACR in the correct orientation.

Raw conversion with ACR 3.1

1. Shadows: Ticking the checkbox will highlight in blue any areas of absolute black (where the image is too dark and details may be lost; the histogram will be clipped to the right).

   

   I was taught (I think in Martin Evening's book Adobe Photoshop 7.0 for Photographers, that it's actually a good thing to slightly clip the shadows and highlights for a technical reason I can't quite remember!

2. Highlights: Ticking the checkbox will highlight in red any areas of absolute white (where the image is too light and details may be lost; the histogram will be clipped to the left).
3. Space: ACR 3.1 gives you the choice of converting to Adobe RGB (1998), ColorMatch RGB, ProPhoto RGB (with the largest colour range or "gamut" of the four choices, resulting in less chance of clipping), or sRGB IEC61966-1. Watch what happens to the histogram as you change these. As an example, here is the histogram in ACR 3.1 with the colour space set to Adobe RGB. Notice the spike in the red channel to the right, which indicates highlight clipping (visible in the image behind as red blotches; the blue channel is also clipped to the left):

   

   Here is the same image with the colour space set to ProPhoto RGB; there are no blown highlights and the red and blue spikes have gone as there is no clipping due to the larger gamut:

   

   So, when converting 16-bit raw images I'll use ProPhoto RGB; when opening 8-bit JPEG images I'll stick to Adobe RGB (1998) to reduce the risk of colour banding.

4. Depth: This was discussed here; set to 16-bits/ channel. Kelby suggests 8-bits/ channel, but once again the histogram reveals how this can damage the image data.

   

5. White Balance: May be "As Shot", "Auto", or chosen to match various lighting condition presets. Determines whether "white" in your image is a warm (yellow), neutral (you can also use the eyedropper tool to set this by clicking on a colourless area), or cool (blue) "temperature". The ability to change white balance at this point makes it less important to get it right in the field, so you can "set and forget" the D70 to Auto white balance. You can't change your mind once the image is open inside Photoshop. The D70 version 2.0 firmware update does not appear to encrypt white balance data.

   

6. Exposure: Used to set the brightness of the brightest part (the highlights), or "white point", of the image (sort of like the right-hand triangle in the PS Levels dialog—see image below—but at this stage we are working within a linear rather than a gamma-corrected space so have more room to move without loosing highlights, as I understand it). Move the slider too far to the left (up to 2 f-stops) or right (up to 4 f-stops) and you'll get clipping of shadow or highlight detail, respectively. Following the "expose to the right" rule that digital photographers often quote, it may be appropriate to move the slider to the right (where most of the tonal information is in the histogram) but stop before you see highlight clipping (see no. 2). If this makes the image look too bright, turn down the brightness setting (no. 8 ). Default slider setting is Auto or 0.00.
7. Shadows: This does the opposite to no. 6, setting the darkness of the darkest part, or "black point" of the image (corresponding to the left-hand triangle in the PS Levels dialog). Default slider setting is Auto or 5.
8. Brightness: Sets the gamma value of the image by adjusting the mid-tones, or "gray point" of the image to affect overall brightness (corresponding to the middle triangle in the PS Levels dialog; gamma values more than 1.00 increase brightness). Note that you can selectively adjust lighter or darker parts of the image using curves (see no. 16). Default slider setting is Auto or 50.

   

9. Contrast: Decrease or increase image contrast—but I'll uncheck Auto (the default slider setting is then +25) and use curves instead (see no. 16).
10. Saturation: Give the colours in the image more (or less) "punch". Default slider setting is 0; I tend to like +15 for landscapes as I prefer a rich "Velvia-like" look.
11. Sharpness: I'll sharpen in Photoshop itself instead, using the new Smart Sharpen filter after all other adjustments have been made. Go to Photoshop > Camera Raw Preferences... and choose Apply sharpening to: Preview images only. Default slider setting is 25.
12. Luminance Smoothing: For noise in images giving a grainy appearance due to differences in brightness between adjacent pixels. I'm ignoring it here; it can apparently soften your image (if you use it, apply it at 100% zoom to see the effect). I'll use the Reduce Noise filter in Photoshop itself instead. Default slider setting is 0.
13. Colour Noise Reduction: For noise in images manifest as coloured blotches. I'm ignoring it here; it can apparently soften your image (if you do use it, apply it at 100% zoom to see the effect). I'll use the Reduce Noise filter in Photoshop itself instead. Default slider setting is 25.
14. Chromatic Aberration: Aims to remove (or at least minimise) the purple fringe in the presence of strong backlighting (and other colour halos) you sometimes see. Default slider setting is 0.
15. Vignetting: Will lighten darkness at the corners of an image caused by the characteristics of the lens. Apparently this is an issue with the 18-70mm AF-S DX lens I ordered with my D70. Default slider setting is 0; moving it to the right will reduce vignetting. I went to a talk at the local Photographic Society on black and white digital techniques; the guest speaker routinely adds a vignette (push the slider the other way) because it draws you eye into the image:

    

16. Tone Curve: With no contrast adjustment the "curve" will be linear. The more the contrast, the more "S" shaped the curve will look. This works the same as Image > Adjustments > Curves... in Photoshop, and duplicates the functionality of uploading a custom curve using Nikon Capture. In the raw dialog you can use the Preview option to see the effect of manipulating the curve. For example, you might use a custom curve to bring out the blue and pink in an otherwise washed-out evening sky by pulling the curve below the linear marker to the right (you might then use the result as a duplicate layer in Photoshop for highlight recovery).

    Note: If you want to adjust curves using PSE3 (in 8-bit mode), see here.

    

17. Camera Profile: Adobe's colour profile for the D70 was introduced with Camera Raw 2.4, which is why ACR 2.4 is listed even when you are using the ACR 3.1 plug-in (confusing!). If you wanted to tweak Adobe's profile to taste, you could do so here using the tint, hue, and saturation sliders. If it matters to you, calibrating ACR in this way could give you results that are a closer match to the way images appear when converted with Nikon Capture. You can save a settings subset for application to future conversions if your preferences are consistent (e.g. as here), or ignore this tab completely.

Some of these settings can also be adjusted in Photoshop: why do them in Camera Raw? Fraser explains:

To make the best use of a raw image's available bits, use these controls to shape the image's overall tone and contrast. If you instead bring the image into Photoshop for correction, you'll lose a great deal of tonal information in the conversion from linear gamma to a gamma-corrected Photoshop working space, making your job more difficult or even impossible.

Photoshop: safety in duplicates

I first create a duplicate layer by dragging the Background layer onto the "Create a new layer" icon at the bottom of the Layers palette. By default the duplicate is called "Background copy":

This means I can now apply filters etc. to an exact copy of the converted raw image, and still be able to go back to this point without having to convert from raw all over again.

Note: This isn't the same as an adjustment layer (see below).

Incidentally, you can also use duplicate layers to adjust for under- or over-exposure. Set the blend mode ("Normal" by default) of the duplicate layer to "Screen" to lighten, or "Multiply" to darken. If the effect is too strong, change Opacity to a suitable percentage.

Photoshop: the lens correction filter

Lens correction: Straighten with the Straighten Tool; remove distortion (barrel or pin cushion) with the slider of the same name; perspective correction with the Transform sliders.

Photoshop: right tools for the job

Crop: After clicking OK to close the Lens correction dialog, use the Crop Tool to trim the image.

Healing: Spot Healing Brush (to clear specks of dirt or other distractions (e.g. far-off birds can look like dirt).

Red eye removal: Remove red eyes caused by flash light bouncing off the retina (my wife wants to know if it can also fix closed eyes!).

Photoshop: easy colour correction

Colour correction is the bit I find most difficult (maybe because I'm red-green colourblind?). That's why I like Kelby's take on Photoshop's Auto Colour correction. First there is some one-off setting up to do:

1. Open the Levels dialog then click the 'Options" button;
2. Select "Find Dark and Light Colours" as the Algorithm;
3. Check "Snap Neural Midtones";
4. Click on "Shadows" and enter R:20, G:20, and B:20, then click "OK";
5. Click on "Midtones" and enter R:133, G:133, and B:133, then click "OK";
6. Click on "Highlights" and enter R:244, G:244, and B:244, then click "OK";
7. Check "Save as defaults" then "OK".

Done. To use these settings in future, choose Image > Adjustments > Auto Colour.

Photoshop: tweaks on adjustment layers

An adjustment layer is like a transparency on which you make your edits; they overlie your original but leave the original pixels underneath untouched (i.e. edits are non-destructive). You can't, however, apply a filter to an adjustment layer (there's nothing there to be filtered!). You instead use an adjustment layer to make adjustments that you find, naturally enough, under the Image > Adjustments menu. If you look at the options under the Layers > New Adjustment Layer menu, you'll see they correspond to options in the Image > Adjustments menu.

With the Background copy (duplicate layer) selected in the Layers palette, I can choose for example Layers > New Adjustment Layer > Curves... which will open up the Curves dialog (the same one I get by choosing Image > Adjustments > Curves...). The purpose of each adjustment layer is clear via automatic naming, and because individual adjustments can be made on individual layers you can subsequently delete one without undoing the others:

Adjustments you might consider doing are:

• Levels if further tweak to overall brightness required (middle triangle);
• Levels if a (further) colour cast correction is required (middle eyedropper tool);
• Curves if a further tweak to contrast is desired;
• Hue/ saturation adjustment.

Photoshop Actions (optional): An additional adjustment layer can also be useful for applying and reversing custom Photoshop Actions e.g. Thom Hogan's "Digital Velvia" action (it comes with his eBook on the D70).

Photoshop: final steps

Noise reduction: Luminance smoothing and colour noise reduction. As a suggestion, start with Strength: 7, Preserve details: 45%, Reduce Colour Noise: 65%, Sharpen Details: 0%.

Sharpening: Filter > Sharpen > Smart Sharpen... will define edges and deal with motion or lens blur. As a suggestion, start with Amount: 56%, Radius: 1 pixel, Remove: Lens Blur, and check "More Accurate".

Photoshop: output

The type of file will depend on the success and purpose of the image.

Great shots

If it's a winner, then I'll save it with the maximum of information and quality possible. This entails saving in Photoshop file format (less universal but resulting in a smaller file size than TIFF). For example, at 300 dpi, 16 bits/channel, with a ProPhoto image profile a file that weighs in as an 80 MB TIFF might be 58 MB in .psd format.

OK shots

Most of the time I merely view my images on my own screen. Occasionally, when a print is required, I'll use an online service that accepts JPEG uploads like PhotoBox, Colourmailer, or Kodak via iPhoto. As mentioned above, I can always "redevelop" raw originals if the need ever arises. Maximum-quality (level 12) JPEG at 300 dpi converted to the Adobe RGB colour space (JPEG is only 8-bits/ channel and can't contain the wealth of colour data used in the ProPhoto working space) seems to offer me the best compromise between fidelity, speed of on-screen display, storage requirement, suitability for colour-managed print processing and uploading to iStockPhoto, and as a starting point to re-purpose for e-mailing or web display. More on re-purposing here.

The workflow in summary

The stuff in brackets I'm calling optional.

Credit: Thanks to José Salcedo (website) for his valuable advice on the need for re-purposing and rationale behind some of the above settings.