The term linear conversion has become the norm for describing pure RAW conversion. We don’t see the need for this type of conversion in everyday photographic situations, but every RAW converter nevertheless has to be able to perform it internally.

Figure 5-4. Linearly converted image

A linear conversion converts the manufacturer/camera-specific RAW image data into RGB pixel data and alters the white balance and EV (Exposure Value) settings if necessary.

The result of a linear conversion is usually a very dark-looking image (Figure 5-4), but this is the way all(!) interpolated RAW images look when the Bayer pattern has just been demosaiced.

In order to make such an image usable (Figure 5-6), a tone curve like the one shown in Figure 5-5 is applied. This is not the type of curve you would apply under normal circumstances, but it is the right one to use here. Such a curve can be adjusted to suit individual cameras and situations, but the basic form remains the same.

Figure 5-5. Typical DSLR tone curve

Figure 5-6. Image after applying the camera tone curve

This is the single most important step in the RAW conversion process for cameras that work on the Bayer pattern principle.^[75] With only few exceptions, this means all digital cameras. The Foveon sensors used in Sigma SD series DSLRs capture three colors of image data per pixel. There are large numbers of demosaicing algorithms available for interpolating Bayer pattern image data, and they produce results of widely differing quality. Generally, algorithm quality – and consequently image quality – is constantly improving.

The balancing act between preserving image detail and reducing noise is a difficult one, and no single RAW converter has the ideal solution. We are prepared to let fine details go for the sake of producing clean, sharp images shot at moderate ISO speeds. Most RAW converters on today’s market are capable of producing acceptable results.

RAW converters generally include usable generic camera profiles for all supported camera models. Camera profiles (Creating Device Profiles) describe the individual color characteristics of a camera. The most precisely formulated camera profiles do not necessarily produce the best images, and the newest versions of ACR have several profiles for each supported camera model – including one that simulates each manufacturer’s own JPEG output.

A good tone curve will brighten shadow detail without sacrificing too much contrast or making shadow noise too obvious.

Some RAW converters allow you to add your own profiles or edit the ones they supply. Generating camera profiles requires additional software and a lot of know-how.

There are two basic schools of thought regarding the use of camera profiles:

We usually use the generic profiles supplied with our software, and we expect future versions of the programs we use to include profiles that are tailored to specific shooting situations. Generating custom profiles is complex, time-consuming, and usually not worth the effort, even for professional photographers. Adobe’s DNG Profile Editor [31] is designed to make generating and editing camera profiles for ACR and Lightroom simpler, and it is available for free download at the Adobe website.

Suitable white balance settings are critical to the RAW conversion process. There are four basic ways to set white balance:

Figure 5-9. Tone and color temperature curves

Figure 5-10. White balance preset offered by the ACR WB menu

Setting white balance is not a trivial task, and a good RAW converter will update all changes to the preview image in real time.

Correct exposure is vital in a digital image, so an essential part of a RAW converter’s toolset is a histogram display that can show under- and overexposure warnings in all three color channels. Overexposure is much harder to correct than underexposure, so burned-out highlight detail is usually irretrievably lost. Expose carefully, and avoid overexposure at all costs. All of the RAW converters we use have functional exposure correction tools.

Figure 5-11. Exposure sliders in ACR 6

You can see in Figure 5-11 that ACR’s exposure sliders are arranged in a logical sequence that we usually follow when adjusting our images. We hardly ever adjust contrast here. We use the Exposure and Recovery sliders to set the white point and Blacks (also available Lightroom) to set the black point for our image. Recovery and Fill Light sliders are generally a standard RAW converter feature, even if they have slightly different names from program to program. Recovery tools are designed to recover burned-out highlight detail as far as is possible. Fill Light brightens shadows selectively without lightening highlights and midtones. We generally reduce the Blacks value before making Fill Light adjustments.

One of the RAW format’s major advantages is that you can often rescue slightly under- or overexposed images using the Exposure, Brightness, and Recovery sliders. While Brightness slider settings affect the whole image, Recovery only darkens the highlights. Play with both sliders to see what results you get.

The degree to which you can recover overexposed image areas depends on your camera, how powerful your converter’s tools are and the nature of the overexposure itself. High-end cameras often have a whole f-stop exposure reserve, or even 1.5 stops if only a single color channel is overexposed.

The left-hand histogram in Figure 5-12 shows a strongly overexposed image shot using the default camera settings. The right-hand histogram shows a better exposure curve that we produced by reducing the Exposure value by 0.41 and adding a Recovery value of 37 (in Lightroom). The image is rescued – if this had been a JPEG image, we wouldn’t have been able to alter the burned-out highlights.

Figure 5-12. The burned-out highlights in the left-hand histogram were eliminated using the Lightroom Recovery and Exposure sliders-

Most RAW converters include under- and overexposure warning functionality, allowing you to directly optimize exposure without risking additional color clipping. Activate the exposure warning and keep a careful eye on the image preview while you adjust your sliders. If you want to process an image later using Photoshop, be careful not to set restrictive black and white points – always leave some adjustment leeway at both ends of the histogram curve.

Figure 5-13. ACR overexposure () and underexposure () warnings

In order to judge exposure and white balance correctly, a RAW converter’s histogram needs to include individual red, green, and blue color channel displays that can be shown separately or simultaneously. The ACR histogram display includes two triangle icons at top left and top right that indicate when shadow and highlight clipping is present. The display also includes basic exposure data and the RGB value at the current cursor position. ACR and Lightroom allow you to set the image black and white points directly in the histogram window, should you feel this is necessary.

Figure 5-14. An ACR histogram. The non-black triangle at top left indicates that the image has clipped shadows.

Figure 5-15. RAW Developer’s histogram display, with the scaling slider on the right

RAW Developer [45] has particularly extensive built-in histogram functionality that includes a tool for adjusting the vertical scale of the histogram display (Figure 5-15).

Figure 5-16. RAW Developer has an additional wide screen histogram view.

Figure 5-17. Capture One can display the histogram in line form.

This makes it easier to identify smaller burned-out highlight details. The RD histogram can also be “stretched” in order to make judging the shadow and highlight ends of the curve easier (Figure 5-16).

The Capture One and RAW Developer histograms can be displayed as simple line curves, rather than as filled in curves, which can help you to judge finer exposure details.

RAW image noise should be reduced as soon as possible in the workflow (i.e., using your RAW converter) before it can be intensified by other correction steps. Although the quality of noise reduction offered by Photoshop plug-ins such as Noise Ninja [80] or Dfine [75] is better than that provided by most RAW converters, performing slight noise reduction early on in the workflow still helps to produce better overall results.

Some noise reduction algorithms are based on ISO and camera model EXIF data. Bibble has Noise Ninja’s excellent noise reduction functionality built in. The quality of the noise reduction tools built into RAW converters is improving rapidly.

Reducing image noise usually reduces overall image sharpness (to some extent), resulting in close interaction between sharpening and noise reduction processes.

Noise reduction and artifact removal are often lumped together as similar phenomena. The truth of the matter is:

It is often difficult (or even impossible) to remove moiré effects using a RAW converter, so we often resort to using Photoshop instead (Removing Moiré Effects). It is best to remove other lens-based artifacts, such as chromatic aberrations and vignetting, at the RAW conversion stage. ACR, Adobe Lightroom, Capture One, and Nikon Capture NX are all strong in this area. Lightroom and ACR even have separate Luminance Noise, Color Noise, Chromatic Aberration, and Defringe tools (Figure 5-18). The latter is used to remove blooming effects caused by sensor elements overloading at high-contrast edges.

You can also reduce moiré effects in ACR and Lightroom by using the Adjustment Brush with a negative Clarity value. Capture One has its own dedicated moiré reduction tool.

Figure 5-18. Lightroom 3 Noise Reduction and Chromatic Aberration correction tools

The better the interaction between the software and the user, the better the results a RAW converter will produce. All of the tools listed here have fast-reacting, interactive user interfaces. Non-modal processes (Non-modal Editing), as offered by Lightroom, ACR, and LightZone, are an advantage. Nikon Capture, for example, requires you to close one dialog before opening another. This is time-consuming and sometimes annoying, and makes this otherwise powerful tool less popular in some circles.

All of the RAW converters we describe have full color management support. This includes:

Figure 5-19. RAW Developer includes a range of standard profiles and can convert data to all color spaces installed on its host system.

Nearly all RAW converters include browser functionality for viewing RAW image files before they are converted. Photoshop, Bridge, and ACR form a complete RAW processing toolset (The Influence of Individual Tools on the Workflow) in which ACR generates the preview images that are displayed in Bridge. If necessary, Bridge can be set up to display only RAW images. RAW browsers are also useful for transferring selected conversion settings from RAW-edited image files to other, similarly exposed images.

Figure 5-20. A loupe helps you view individual image details even in the smallest preview window.

RAW Developer, Apple Aperture, Lightroom, and Capture NX allow you to view and compare images side by side, and others are sure to add similar functionality soon. A loupe function is nowadays a standard feature, helping you to view image details without changing magnification of the entire preview. Bridge introduced a loupe function in version CS5 (Figure 5-20), and Apple Aperture has the best implementation of this feature that we have seen so far.

Typical Preview thumbnails are not large enough to allow you to judge the quality of an image. You need a larger preview window in order to correctly judge image sharpness, contrast, saturation, and color. ACR/Bridge, Capture One, RD, and several other programs provide preview images that are large enough so you can make an initial judgement regarding the quality of your images. Preview images should have at least the same resolution as your monitor.

A real-time preview of image corrections saves time during the workflow, but large preview images do use more disk space. You can set your program preferences to delete preview images after a certain period of time, but saving preview images permanently saves time and improves overall performance. Preview images are usually saved in a compressed JPEG format. Deleted preview images have to be regenerated from RAW image files, which takes time and requires a lot of computing power for large collections of images.

Quality RAW converters allow you to save your conversion or adjustment settings in order to apply them to other images.

Figure 5-21. Lightroom allows you to apply image adjustments and metadata to other image files.

The programs listed in this book all allow you to cut and paste selected settings from one image to another (with varying degrees of simplicity). All-in-one tools also allow you to apply RAW image settings to images saved in other formats. Some programs include functionality for cutting and pasting metadata and keywords, which saves time when you are cataloging your stock of images.

Selective corrections – such as creative sharpening or brightening of shadow areas – can be applied using very little disk space in a RAW converter. Removing blemishes caused by dust particles on the image sensor by copying neighboring pixels is a selective correction that we apply regularly. Because these types of corrections are applied non-destructively using parameters that are saved separately from the image data, they are easy to save and apply to other images, too. (If there are dust particles on your image sensor, they will probably be visible in all the images from a particular shoot.) You can then fine-tune each image separately if the transferred settings don’t suit your other images.

Aperture, Lightroom, ACR, LightZone, Nikon Capture NX, and most other up-to-date RAW editors allow you to simply retouch small blemishes using brush tools or soft-edged, lasso-style selections.

These types of correction require a lot of computing power in order to refresh image previews in real time. If you can’t immediately see the effects of your corrections, you will tend to over-correct, which is frustrating and time-consuming. So, be patient when performing image corrections. Photoshop Layers techniques (Chapter 7) are one way of working around such limitations. As computing power increases, future RAW converters are sure to include increasingly complex correction tools.

Many jobs produce sequences of images that require the same white balance and exposure settings (and similar corrections). In such cases, batch conversion can save time and effort. The very simplest approach is to convert all images in a shoot to TIFFs or JPEGs using your RAW converter’s default settings. A more advanced approach is to correct a prototype image of a sequence, apply its settings to all the other images, and finally batch-convert them.

Batch processes usually run in the background, but nevertheless can use up a lot of computing power. If you are using a laptop or other less powerful computer, it is better to run batch processes once you have finished making other, processor-hungry corrections. RAW Developer has specific settings for controlling batch processes, but these are not yet available in ACR. We often need to produce multiple versions of an image in different formats or resolutions during conversion – functionality that is supported by Capture One, Bibble, and Capture NX as part of the batch conversion settings.

All-in-one tools generally leave RAW images in RAW format, and applying settings from other images only involves a set of previously saved parameters – a process that doesn’t cost too much processing power. Image export is a CPU-hungry process and should therefore be set up either as a background process or as a set of parallel processes. For this reason, most all-in-one programs farm out batch processes to their export modules.

A powerful print module is an essential part of any all-in-one workflow tool. Whether you use your RAW converter for making anything more complex than a test print depends on the way you work. Most high-end prints that you want to optimize and add a frame or a signature will require you to use Photoshop or some other high-end image processor. With the exception of Lightroom and Aperture, we only make prints in Photoshop or a specialized print application or RIP (RIP, see Specialty Printing Software and RIPs).

We consider the degree to which we can integrate a RAW converter into our workflow critical to its value. This view is, of course, subjective, but we consider Capture One Pro (with its fast, real time preview and background processing support) to be the best currently available converter. ACR’s integration with Photoshop and Bridge makes it a powerful tool too. RAW Developer is our current favorite among Mac-based converters.

Most of today’s RAW converters and browser tools with RAW support (such as Photo Mechanic, Breeze, Extensis Portfolio, iMatch, and Phase One’s Expression Media DAM systems) allow you to preview RAW images before you convert them. All-in-one tools offer the best integration of image browsers and databases.

We would like to see better image management integration for RAW converters and functionality that logically groups image versions that only differ from each other in respect to their current adjustment state. (Deleting an image version does not and must not delete the associated original!)

Since early 2006, nearly all new notebook and desktop computers are equipped with multi-core processors (and sometimes even with multiple processors). Multi-core processors counteract the endless spiral of ever-increasing processor speeds and the associated increased heat generation. Multi-core processors are an advantage for performing CPU-hungry image processing tasks. Provided that your RAW converter and your image processing software are capable of taking full advantage of the new technology, multi-core processors refresh preview images faster and allow you to process images while others are still being imported in the background.

Multi-core support also makes other background processes (such as RAW conversion itself) faster without limiting overall processing power. Multi-core processors make image correction processes perceptibly faster. Capture NX and Canon Digital Photo Professional do not yet support multi-core processors and are relatively slow as a result. But Lightroom and Aperture still have room to improve in this respect – here, too, we hope that future releases will address these issues.

The auto functions built into many of today’s RAW converters have advantages and disadvantages, but we generally avoid using them.

The main advantage is that auto settings quickly produce generally usable images that you can either apply directly or as a basis for further corrections.

The disadvantage is that the algorithms that drive auto functions cannot be adjusted and sometimes produce results that don’t suit the image purpose at all. For this reason, we usually deactivate ACR’s auto correct functionality (Figure 5-28).

Figure 5-22. ACR includes a number of Auto corrections.

Figure 5-26 illustrates the basic corrections, steps, and settings involved in the RAW conversion process. The sequence is not set in stone and can be varied to suit the image(s) being processed. The demosaicing step is only included for completeness’ sake, as it happens automatically and doesn’t involve any user intervention.

The “correct” sequence for straightening and correcting white balance or exposure is widely debated. We prefer to straighten and crop our images first, as this often eliminates image areas that no longer need to be corrected. If a strong exposure correction is required, we often perform a preliminary correction in order to gain a better overall impression of our image before we adjust white balance.

Most RAW converters perform automatic compensatory sharpening in order to give the user an idea of the “real” look of an image. If your converter doesn’t sharpen automatically, you may have to bring the sharpening step forward in the conversion sequence.

The sequence illustrated in Figure 5-26 doesn’t completely account for the selective corrections that are becoming increasingly available in RAW converters. For example, if we are processing an image that includes an overexposed sky, we use ACR or Lightroom to add a graduated filter relatively early in the workflow. We then perform all our other adjustments on the corrected version of the image.

We only perform the final sharpening, scaling, and conversion/export steps when we are sure that we don’t want to make any further adjustments (e.g., in Photoshop or LightZone). If you use an all-in-one program, these steps usually don’t apply, as you simply leave your image files and their accompanying corrections and settings in their native RAW format and only convert (render) images on demand, e.g., to work on them in Photoshop, to pass them on to another application, to print them outside the all-in-one application or to pass them on to another user or customer.

Figure 5-26. Steps involved in optimizing and converting RAW image files

Non-modal application architecture (as explained in Non-modal Editing) helps to keep the workflow smooth if we need to switch between multiple image versions or states during complex editing processes.