Image Processing for Hybrid Processes IV – Processing Raw Files


Raw files “suited for alternative processes” or “suited for Fine Art Photography” could be considered as equivalent definitions. Those qualifications would also imply the search for the maximum quality with no matter how difficult or time consuming the method employed is. The operations described in following paragraphs are the employed by the author whether the image will finish as a Palladium print or as a Photogravure on copperplate. In both cases, the goal is therefore a monochromatic image. As not always is already decided if the image will go to one or the other final art print type, the raw is always processed as a positive image. Lateral reversing and negative/positive form decisions are taken just before inkjet printing the respective negative or positive intended to a given alternative process.

Any raw file coming from a digital camera must be processed in a such way to preserve the maximum dynamic range and thus, tonal scale, the geometric properties of the subject and an image field evenly illuminated. Here we are considering that the exposure metering in the camera has been properly adjusted in order to capture the complete scene dynamic range; see the previous post Image Processing for Hybrid Processes III – Digitization with Camera to know the dynamic range capability of a given lens-camera combination. Processing raw files can be done with a lot of specialized software packs. The two common options are the proprietary software of the camera manufacturer or third part software as RPP-Raw Photo Process (low cost donation), GIMP (free access), CaptureOne, Adobe Camera Raw and many others.

After have been trying several programs for a period of time, I finally decided that Adobe Camera Raw (ACR) is the most complete option offered today, although this decision does not invalidate many other software of equal capabilities. It comes integrated with the Adobe Photoshop pack, that is in turn my best election for fine tuning the final image version. Details on that will be commented forward. This allows to use the same software to process the raw file and to fine tune the final image version. One recommended procedure is to perform as many processing as possible at the raw stage. The goal in this stage is to preserve the maximum information, not the more aesthetically pleasant image. As a general rule, remember that to apply contrast is always possible and easy to do. Conversely, reduce an excessive contrast can be possible but not always preserving textural information. Therefore, be cautious. An apparently flat image can be easy to fine tune than a contrasty one.

Figure 1. Adobe Camera Raw (ACR) user interface. The arrow marks the place to open the adjustments window commented in the text (click on the image for a larger version).

Find here a list of the main controls in the user interface of ACR and a brief description of its capabilities (CC2018):

  • Open the raw file and begin by clicking on the link at bottom of the ACR user interface (Fig., 1)
    • Decide an output file resolution. For most photographic quality inkjet printers, 240ppi (default in Photoshop) or 288ppi is fine. Both figures are integer sub-multiples of 1440, that is the native resolution of those high end inkjet printers (1440dpi). Being the quotient an integer (1440/240 or 1440/288), the printer RIP (Raster Image Processor) does not perform any compression/interpolation process to adjust the number of ink drops available for each image file pixel. This is a safety precaution in order to avoid local loss in sharpness in the printed image. The use of special printing tools as Print Tool and/or Quad Tone RIP, can recommend a higher resolution like 360ppi, that in turn is also an integer sub-multiple of 1440. Nevertheless, I’m printing with Quad Tone RIP through Print Tool and image files at 288ppi. I haven’t noticed any improvement setting 360ppi. I use the printer to create negatives for Palladium or positives for Photogravure on Copperplate, both on Pictorico OHP transparent media. May be printing final inkjet positive prints on paper justify the recommended 360ppi.
    • Decide a bit depth. The two options are 8bit or 16bit, being the later (16bit) the only recommended for a suited fine image processing. Although in the computer display it may not be observed a noticeable difference, at the final printed image it is. Specially in the smooth shades of gray.
    • Decide also about a Colour Profile to be applied to the image file. For all B&W images, I use Gray Gamma 2.2.
    • Do not check any kind of output sharpening.
    • Do not apply resizing if not needed.
    • Check on Open in Photoshop as Smart Objects. This allows to come back to the ACR interface by double-clicking the Background layer of the file opened in Photoshop. In this way, you can fine tune the ACR settings even during the final image edition process.
  • Select the fourth tab from the left at the tabs window shown at right of the ACR user interface. Check to Convert to Grayscale and adjust to Default in Grayscale Mix. Then, adjust the cursors to achieve the look you want attending to the way each color translates into B&W. There is not a rule to apply, only a combination allowing to differentiate the color surfaces in the scene when they are translated to B&W. Is something like to use a coloured filter in front of the camera lens taking pictures on B&W film, but much more flexible. Do not worry about the look of the image in an aesthetic sense. Look for as much different tones of gray as you can. The local contrast will be further edited in Photoshop.
  • Select the first tab, Basic, from the left at the tabs window shown at right of the ACR user interface. This tab contain all adjustments about Exposure and Contrast. Those parameters might be adjusted as is derived from the camera OECF measurement. The goal is to achieve the maximum useful tonal range. At the upper right corner of the interface, there is the image histogram; it shows in real time the changes applied to the image. By clicking the triangles at left and right, the image shows the eventually under and/or overexposed pixels.
    • For most high end cameras, exposure can be adjusted on a range of ±1EV, depending on the ISO adjustment. Beyond this limits, noise or loose of information may occur. If there is a systemic need of modify the nominal exposure, this may indicate a bad camera exposure measurement method.
    • As the Contrast will be locally adjusted in Photoshop by the use of Adjustment Layers and Masks, it make sense to left this cursor at 0 setting.
    • The cursors Highlights and Whites allow to modify the position in the image histogram of the pixels with higher gray value. Conversely, Shadows and Blacks do the same with the pixels of low gray value. There are a lot of recipes about the use of those cursors. With my cameras, segmented area exposure mode and high or normal contrast scenes, Highlights and Whites remain in its default position or are moved to the left (negative values). Begin always by Highlights and continue with Whites if needed to recover some lost information. Be aware about that if pixels have been saturated by excessive exposure, lowering Highlights and/or Whites modifies the gray value but does not recover textural information. Recovering information from shadows can be done pushing the Shadows and/or Blacks cursors to the right (positive values). As earlier, begin always with Shadows and use Blacks only if needed. If shadows were excessively underexposed, pushing Shadows and/or Blacks to the right can show some noise, depending on ISO adjustment and camera sensor. In fact, an scene with a dynamic range not more extended than the camera dynamic range (OECF) and properly exposed, does not will show overwhelming problems. Accurate scene dynamic range metering is therefore a good and safe practice. Nowadays, with spot metering in all medium and high end camera bodies, there is no reason to fall into grave errors. High end DSLR and EVIL cameras equip sensors capable for a dynamic range of 10EV or f/Stops and even more in some cases. A film with a Dmax of 2.5OD can encompass not more of 8.3EV.
  • Noise can be minimized by the Noise Reduction control tools available in the third tab, Detail, of the ACR interface. In this tab there is also the Sharpening control. Sharpening is one of most complex and useful setting in any digital image processing and it might be applied in a specific way to each image. This will merit a post to discuss in depth. Therefore, if noise is not a problem because of the use of tripod and normal or low ISO settings, all cursors in this tab would be settled at 0 excepting Radius in Sharpening, that remain in its lowest value of 0.5. This 0.5 Radius do not apply any sharpening to the image as far as the Amount is multiplying by 0 the Radius effect.
  • About the second tab, Tone Curve, it do not makes sense to use it because Adjustment Layers and Masks in Photoshop offer better and locally applied control. Therefore, Parametric controls are settled to 0 and the Point control to Linear.
  • The fifth, seventh and eighth tabs, respectively Split Toning, fx (Effects) and Camera, are not useful with B&W images or are dedicated to aesthetic effects.
  • The sixth tab, Lens Corrections, allows to correct some optical aberrations coming from the camera lens design. Check to Profile tab and check again to Remove Chromatic Aberration and Enable Profile Corrections. The corresponding lens profile will be automatically applied if it is in the ACR database. Some lenses of third part manufacturers or older lenses do not provide to the camera interface its data and therefore, there are not the corresponding metadata in the raw file. In this case, some corrections as Transverse Chromatic Aberration fringes, Barrel or Pincushion Distortion and Vignetting can be applied with the controls in the Manual tab.
  • Geometric errors derived from the relative camera/subject positions, can be corrected with the tool Transform available on the top menu in the ACR interface. There are automatic and manual tools to modify the geometry of the image. As transform operations imply always downsampling and/or interpolation of pixels, it is recommended to perform that at raw level better than into Photoshop. Remember that with the file opened in Photoshop as an Smart Object, you can always come back to the ACR interface in order to apply a forgotten adjustment.
  • The ninth tab, Presets, allows the user to quickly apply a pack of previously saved settings. It is a good idea to prepare some presets packs dedicated to the most usual camera/lens combinations. Therefore, when processing a file or group of files taken with the given camera/lens combination, it can be done by simply apply a Preset pack. At the upper left corner of the ACR interface there are controls to apply the same presets to all the opened raw files (Filmstrip>Select All>Sync Settings). This method, not only is less time consuming, but also helps do not forget any setting. Each pack of settings specifically dedicated to a given camera/lens configuration or any other criteria you want to apply, can be saved with its own name. Remember than the more complete is the name of a preset, the more easy is to find and remember it in the future.
  • The tenth tab, Snapshots, allows to take snapshots of a partially processed raw file. If in doubt about the convenience or not to apply a given adjustment, take two snapshots with and without it. You can later recover any of the two versions avoiding the need to duplicate the raw file.

In the upper tools bar of the ACR interface, there are shown many other tools to locally apply changes in the image. All those processing operations can be applied in a better and more selective way into Photoshop by means of Adjustment Layers and Masks. Finally, some words about the image cropping. If the raw file is opened into Photoshop as a Smart Object, the Cropping Tool shows a Crop Preview but do not delete the surplus image area. Therefore, there is always possible to reconsider the image cropping during the edition process. Thus, it make no sense to crop the image in the ACR interface because, by now, it is an irreversible operation.

Finally, the processed raw file can be simply saved clicking at the Done button or opened in Photoshop by the Open Object button. In both cases, Photoshop creates a “sidecar” text file (.xmp) with all the adjustments applied to the raw file an the same file name. This file must be preserved or migrated beside the raw file in the same directory. Opening the raw file in the future, Photoshop will recall the adjustments from this .xmp file. Rename it accordingly with the changes applied to the raw file name. Now the raw file is ready to be edited in Photoshop in any desired aesthetic way.

5 Replies to “Image Processing for Hybrid Processes IV – Processing Raw Files”

  1. Muy interesante, como cabe esperar siempre de ti.
    Carles, te voy a hacer una pregunta que no tiene mucho que ver con tu artículo, pero creo que me puedes ayudar y te lo agradecería mucho.
    Comentas que para todas las imágenes en blanco y negro usas el perfil Gray Gamma 2.2. Me estaba yo preguntando, (de ahí volver a leerte), si no sería mejor para guardar fotografías en escala de grises y 16 bits de profundidad de color utilizar un perfil ICC más “estandarizado” y que no fuera dependiente de Adobe, pero no sé cuál podría ser. El único que he visto es el perfil que viene en el sistema del MAC “Generic Gray Gamma 2.2 Profile.icc”. En la página del ICC y de ECI no he encontrado una solución.

    Liked by 1 person

    1. Cuando todavía estaba en la UPC, recuerdo que hicimos un estudio de los diferentes perfiles de escala de grises disponibles en aquel momento. En el caso concreto de los Gray Gamma 2.2 y Generic Gray Gamma 2.2, la única diferencia era que el Generic aplicaba una curva de salida algo más clara en la zona de los valores de gris bajos (sombras alrededor de valor de gris 25-35). La modificación respecto al Gray Gamma 2.2 de siempre era muy sutil y difícil de ver en según que pantallas. Respecto lo de si es de Adobe, tampoco creo que fuese un problema en el caso de desaparición de la marca porque las imágenes siempre se podrían codificar con otro perfil en el futuro. Difícilmente nada durará eternamente. De todos modos, no soy ningún experto en este campo y quizá se me escapan cosas que pudieran ser importantes, sobre todo en el ámbito de los archivos con cantidades ingentes de imágenes.

      Liked by 1 person

Leave a comment