A course for those concerned with artworking, manipulating and otherwise preparing images for output using the medium of print.
Amongst Photoshop's many other features are a wealth of tools and controls designed specifically for creating work which will ultimately be output for one of the print mediums. These might be fabric designs, customised separations for the silk-screening of garments or single item display materials. Most commonly however, the work will be intended for publication in the paper and board medium, using Pantone and process colour inks for volume editions.
The rules for preparing work that is destined for any print medium differ from instances where the final output is to be created with various light sources. In particular, Photoshop needs guidance from the user to properly convert the additive colours seen on screen to the subractive colour model employed by the printing process. It also needs to know how to calculate the extra black ink component used to enhance the subractive primaries of cyan, magenta and yellow and, as a result, how to best decrease the level of overlapping cyan, magenta and yellowtones which would have been imitating true back before the additional separation was made.
The conversion is further complicated by the surface attributes of the fabric (commonly paper or board) used in the print process, together with the age, condition and precision of the printing press itself.
In the world of mass market publishing the are two principal types of printing press in popular use. For large, high-speed jobs a web offset system is employed. This involves printing multiple pages that have been stripped together through a methodical process known as imposition. This usually involves dedicated software although the popular page layout programs increasingly provide optional æannexÆ facilities which replicate the procedure. The resulting spread of pages are then all printed side-by-side on large rolls of paper, firstly on the topside then on the lower, before being folded and trimmed to produce the final product.
For lesser print-runs or jobs requiring a bit more precision, a slightly more stable and less unwealdy process is used. Sheet fed presses operate slower as a rule but are able pull individual sheets of paper into marginally more accurate contact with the 'flat' printing plate. These sheets may also be quite large and have been prepared with the imposition process mentioned above. On the other hand they may simply be single or double-page spreads sourced directly from a page make-up program. They will still probably be moved on to a folding process and will certainly be trimmed.
In the case of colour printing, both types of press need to repeat the application of ink four times - once each for the standard process colours of cyan, magenta, yellow and black. Some kind of drying period or process must be accounted for between these applications of ink. Other jobs will require additional special colours and in this case it is likely that deliberate 'holes' will have been punched in the content of the four primary colour separations to create the 'gaps' that will be required for those extra colours. Photoshop will now probably be employed for this task which would traditionally have been rather labourious.
In other cases, jobs are printed using a single colour or a combination of less than four colours. These colours may be one of the four primary process colours, but frequently they are customised inks sourced from libraries like the Pantone sets. Sometimes even single colour or greyscale images can be enhanced by the availablity of two or more colours. One method is to 'tint' an image with a flat percentage of a second colour where the true image has already been printed first, probably with a darker colour ink. Another method is to print the same image multiple times in different colour inks, each time on top of itself but with subtle shifts in the tonal range of each colour to improve the dynamic range of final mix.
In Photoshop, the two examples above may both have been created using the multichannel mode where we can build custom colour separations rapidly - either for effects or direct artworking. In the latter case above, the duotone mode in Photoshop may be used to even better effect.
Multichannel mode is also used freqently for artworking individual silk-screens and other designs for garment or fabric printing. Duotone mode is often used for created or preparing images that are destined for two or three colour printing jobs. The expression duotone in Photoshop is generic and actually includes both duotones themselves and the monotone, tritone and quadtone variants. Duotone is, however, an anomaly where modes are concerned since we are unable to access the individual colour separations through Photoshop's channels palette. Instead we must create work using the principles employed earlier with the curves dialogue in image - adjust.
Photshop's seemingly endless facilities for photo-montage and compositing have already been explored, but the result is usually a complete canvas, with four distinct edges. A popular requirement in print production however, is the ability to export both complete files and component parts in a manner whereby they are 'enclosed' within a customised, finite 'shape' for inclusion in typographically-driven page presentations. In some cases Photoshop itself also provides all the facilities needed for this. In other cases we link with vector-based line illustration programs like Adobe Illustrator and Macromedia Freehand to obtain a combination of quality pixel information and the precise outline contours exhibited by shapes created with the Postscript page description language.
Photoshop makes extensive use of a technique known as 'anti-aliasing' wherby all edges in an image are digitally 'softened' to resemble the image received by the retina of the human eye. This offers us a degree of realism unobtainable using vector processing technologies like the Postscript model. It also eliminates the hard edges in images destined for video, multimedia or broadcast. The print medium, especially where paper and board are concerned, is liable to soften images further simply due to the way the final surface absorbs ink. Whilst it would be foolhardy to 'un-naturalise' an image by eliminating the anti-alaising, it is usually necessary to sharpen pictures rather more than we would for other media. This is a precaution against the aforementioned 'blotting paper' effect creating a significantly blurred reproduction.
Other images are too soft or unfocussed due to poor photography or the limitations of scanning devices. These symptoms are a particular problem with the otherwise excellant and multi-functional PhotoCD format devised by Kodak, which is largely based upon the light component and optimised for storing photography as recorded on 35mm transparency or film negative. Its attributes should match the standard of film grain, but the high-speed process does not attempt to correct or enhance the quality of photography, which itself may be good or poor.
Since the characteristics of the print medium are different from those of light, it is crucial to understand the process of 're-creating light as tone', since this is our ultimate goal. We must also use Photoshop to create the conditions where the light qualities of our computer monitor and the display of colour thereon, are as close as possible to the effect obtained when the work is finally printed. We must also consider that the devices used for creating 'dry proofs' in the electronic studio rarely share common attributes with traditional 'wet proofs' or, more importantly, the final printed job. What looks good on a laser writer for instance, will be a literal 'wash out' if printed on a glossy paper stock. On the other hand, an image prepared properly for that same stock will more than likely come out of the laser printer as a dark mess.
We will look at ways of addressing these situations and problems, looking at techniques and quick methods of working along the way. Before starting however, it is best to look back on basic calibration procedures and selection/masking techniques. These are covered previously in
Overview - File Menu and Design Techniques - Part One.
There are a number of additional calibration steps we may or may not wish to take if we have a dedicated output environment since, of necessity, those discussed previously are those for neutral or average conditions. Printing press operators may ultimately make additional and situation-specific adjustments themselves and it is equally important to be aware of the areas where it is best not to interfere unless specifically requested.
There are two ways in particular whereby we may adversly affect our relationship with the printing house. When we go file - save as (or file - save a copy when our master contains layers) and choose EPS we are faced with choices. If we are in CMYK mode, the option marked DCS will, if on by way of any method on offer, export the film separation data actually calculated by Photoshop as we originally entered CMYK mode. We, ourselves, will have needed this calculation in order to access the colour separation channels for artworking/masking and also to see an accurate visual of the printed result. The exact attributes of this calculation are not necessarily permanent and it may well be that the printer requires only a CMYK composite image - especially if the printing house generates final separations after the imposition process.
In both CMYK and RGB the EPS dialogue shows two checkboxes which if made active will export accurate screens and transfer functions. These are of vital import and exporting one or both may well prevent the printing house from being able to over-ride the settings at a later date. If we are required to export these attributes we should go to check them first.
Both are located in file - page set-up. The screens button will allow us to adjust the information for include accurate screens. In standard circumstances all inks are printed with the same line frequency but at different angles in order that the coloured dots do not collide and create strange patterns on top of each other. The same dot shape is commonly used for all inks. By far the simplest way to set this dialogue is to click on the familiar auto button where we enter the printing device's device resolution and the screen frequency required for the job. This will load defaults into each of the four colours' fields and should be correct. If custom variations are needed we adjust each separation's set of fields manually, then click save to store them for future use. The other option is to activate the checkbox marked use printer's default screens which will prevent us adjusting or setting anything. In this case the settings built-in to the end output device will be used to make the final calculation that converts the Photoshop pixels into the screen dots, but once again we may prevent anyone being able to edit the settings on route to that device.
Still in page set-up, the transfer button shows us the additional output-only curves that will be applied to each separation. These behave in a similar manner to our curves in the image - adjust menu but in this instance they represent a way of controlling ink levels to compensate for deficiencies in the press. It is extremely unwise to adjust these in any way unless using a proof already output from the printing press itself.
If we do have such a proof and a transmissive densitometer to measure the intensity of ink in the dots made previously (the screens button) we will be able to take an accurate readings of the dot gain (the spead of ink) at any known location on any colour separation. Thus, if for example the cyan ink was exhibiting a dot gain of 5% more than that assumed by our printing inks set-up at a position which represented 70% density, yet at the same time showed a 10% extra gain in positions of 30% density, we would create a curve here for cyan ink that would provide totally precise compensation. Using the above example we would take the all same box off, then choose the cyan button, entering 20 in the 30% field and 65 in the 70% field. The next proof run from the press would then print 70% at 70% areas and 30% at 30% areas.
Dot gain itself is not the only part of this equation from our printer's own point of view. Dots representing very light shades of ink may ultimately be too small to hold ink at all when transferred to plate. Equally, dots representing very dark shades may be so packed that their ink spreads out to form solid 'chunks' of that ink. Using transfer functions to make mathematical compensations will not cure this. Extra figures are often entered simply to boost the strength or eliminate altogether the lighter shades of certain colours. By the same token, dark shades are frequently made lighter than the values they should represent, simply to avoid the aforementioned 'spread' - not to mentioning the flooding of ink in the press itself.
This possibility is also something to consider as we move to file - preferences - separation set-up itself, As with printing inks set-up, this dialogue has a pronounced effect on what we see and work with after moving to CMYK mode, but it also contains specific attributes which we should check before exporting. Regardless of the inks and paper stock combination chosen earlier, there are two distinct methods by which Photoshop can display (and if necessary, export) separations.
GCR, the default, is Grey Component Replacement which essentially analysises the image to determine areas where the three primary colours overlap, averages out the percentage value common to each at the point of overlap and re-creates that percentage as a shade of grey on the black separation. The colours themselves in those same areas are reduced somewhat since the image would look overly heavy with the new black on top. Since RGB images have no black equivilent, the grey ramp as it is known is often given as a blending option where the RGB composite is unavailable. It is more or less the same information that becomes included in the Lightness channel of LAB colour mode, which is itself in turn the point where this dialogue is calculated. Since black is ultimately used to mask the reflective qualities of the paper stock, it is important to keep some percentage of real colour underneath it to enrich its presence.
GCR has a number of variables which allow us to control how much the black separation contributes to the density of the overall image. Depending on our choice, the primary colours are adjusted automatically accordingly to standard predefined tables. Choosing none will create a CMYK file with a black separation which is empty of content. What we see as black will actually be a 100% overprint of the primaries. This is of little use unless we wish to manually build our own black components, although a skilled operator could possibily do so to better effect than any other option.
Choosing medium will more often than not be the natural choice. Black is generated from overlapping primaries starting where the average tone is around 20% grey. The primaries are lightened in the darker areas of the image to a maximum of around 70%. The Light option starts the black separation in the region of 40% tones, whilst primary averages above 70/75% are flattened out to 70/75%. Heavy creates black from the tonal region of around 10% upwards, whilst the overlapping primaries are incrementally reduced to no more than a 30% tonal value a their most dense point. Maximum create a black separation which to all extents and purposes contains a full greyscale version of the entire image. Only very basic information remains on the separations for the primaries and if reconverting from a previous CMYK we may also find some totally new residue in previously empty areas.
Sometimes the GCR effect is not required. Newspaper stock has poor reflective qualities and ærichÆ black could look overly heavy. In this case, the second option, UCR (Undercolour Removal) is used. This essentially makes a black separation using primary colour overlaps again, but only after their combined value reaches a reasonable density. Black is therefore strong in dark areas. Light areas of the image are thus reproduced with a slight boost to the primaries and little if no black at all. A balance is maintained by removing excessive concentrations of the primaries under the newly created back separation. UCR is an absolute choice with no variables save our ability to dictate limits for the greatest amount of ink at any one point.
Both GCR and UCR have fields whereby we can dictate the upper limit for black ink and the upper limit for the combined inks resulting from this separation process. The limit for black is usually 100% but a printer should be given the opportunity to reduce this when that particular separation is overly contributing to a æfloodingÆ situation. It is not advised that production persons do this unless specifically told to do so by the printer. The same can be said for the total ink. The possible 400% would be saturating the paper or fabric and is unlikely to be used. A more common limit is 300% across all inks but there are many circumstances where the printer would reduce this further.
The GCR method offers one additional adjustment with represents an opposite to UCR. UCA is Undercolour Addition and allows us to dictate an upper limit for the dense areas of the primaries where they underprint behind the new black. Choosing any of the black generation options from the list (except none) will have reduced the amount of the primaries underprinting behind the new black. We can manually increase them again with this command and we will be enriching shadow areas as a result.
GCR also offers a custom setting on the black generation list. This will show the black curve for the setting we selected last before choosing custom. It allows us to modify the black separation only for the whole of that previously chosen set.
Many production environments do not operate within the confines of one printing press, let alone originate everything for the same paper/textile stock. For those that do it is worth spending the time and money involved to calibrate systems, monitors, page set-up and proofs. The setting of fixed file attributes and generating print separations earlier in the production process can save precious time where high-speed turnover is required. It does however make any variations to procedure harder to achieve. In particular, since quite complex operations are being introduced early in the production process, the amount of æback-trackingÆ required to cure later problems is considerable.
In the majority of production environments it is sufficient to work in known, but æneutralÆ territory - always leaving the gate open for professionals operating further down the line. We should refer back to the Overview - File - Preferences section for Printing Inks Set-Up, Monitor Set-Up and Gamma control. This will already have enabled us to customise Photoshop and our system for both our personal eyesight, the attributes of a particular monitor and the subsequent display of both destination inks and paper stock.
We continue by considering the timing of our conversion to CMYK mode, the possibility of making this a temporary conversion and the particular effects to be achieved with the facilities available when we get there.
There are good reasons to work in CMYK. We can visualise the behaviour of our printing inks and view an RGB screen optimised to create the best possible impression of process colour. We can also access our separations for corrections, masking and artworking. We can create extra æspotÆ colour separations and on rare occasions apply trapping. Yet there are equally good reasons why we should not be in CMYK. We will lose many of the more basic controls and, just by moving there, will have made permanent changes to our file. The timing for the change is therefore rather important.
Photoshop itself uses the LAB colour model for all calculations, even though we ourselves may be working in another model entirely. Whatever we do, the results are always displayed in RGB since these are the light sources built into the monitor unit and the only ones available. This is ideal for videographic artists since their own output employs more or less the same model, but not for those working in print. CMYK values are, to a large extent, a highly restricted sub-division of those seen in RGB. To this effect, what we see on screen in CMYK is the entire RGB range with unprintable colours replaced with the nearest equivilent printable mix. Most commonly, we lose saturation values above a certain point.
Unfortunately for us, a small percentage of the apparant colours perceived when we print in CMYK inks have no equivilent in RGB, nor in visible light itself. The human eye essentially acknowledges light, with what we perceive as colour (hue/saturation) being simply variations on the wavelengths of that light. The artificial simulation of colour we create when printing (or creating any artwork on solid material for that matter) is, in reality, achieved by masking-out with paints, inks or other pigments, the light we to not want to receive when real light is reflected from the paper stock or other material. Real light remains an important part of the equation and is most obvious when we compare the quality of printed materials in different lighting conditions. Daytime light and the artificial lighting used at night are prime examples.
Where ink or paint is seriously masking the reflectivity of our paper stock, as evidenced in dark and shadow areas of an image, the mask itself employs values of which we are aware but are not in themselves as component of visible light. Usually these values will be the dense solid areas on the black separation and may well have been created by us whilst setting the RGB to CMYK conversion above. In this case, the absence of any visible light component means the computer monitor will never be able to accurately simulate the final printing job, but, if we calibrated properly earlier, it should provide an adequate illusion for around 90% of the image range, If we bear in mind that the missing 10% or so will more often than not be a matter of slightly denser black and darker shadow areas, not too much can go wrong.
Tip: Those of us who may have worked in the traditional printing and publishing environment will be familiar with the process of viewing large format transparencies againgst a lightbox. There comes a point where everyone doing this starts to discard the æstained-glass windowÆ effect by making the mental, intuitive, jump to being able tovisualise the printed result. This awareness comes with experience - we need to witness the æbefore and afterÆ situation a good number of times to make the imaginary æleapÆ required. Yet, in many ways, viewing a computer monitor is exactly the same as viewing a transparency - complete with rear light source. Infact we have a new advantage - Photoshop can adjust our CMYK display to reduce the the viewed range far more than that of the lightbox.
Photoshop 3.0 introduced two items on the mode menu, which although not really æmodesÆ in themselves, are highly useful for print production purposes. These are known as Gamut Warning and CMYK Preview and appear at the bottom of the list. Choosing either one simply æturns onÆ the facility it represents whilst we actually remain in the mode previously selected.
Gamut Warning, as seen previously, shows any LAB or RGB colours that fall outside the CMYK specifications in a 50% flat grey tone that is relatively uncommon to most colour pictures and therefore highly visible on screen. We can, if needed, remix this colour ourselves by going to the file - preferences - gamut warning dialogue. This provides us with a quick gauge to the amount of colour that will be adjusted in the later process of converting to CMYK. If the æout-of-gamutÆ colours displayed are slight, it is possible to use the sponge tool in the toolbox (the last one at bottom right, option-click number three), set to de-saturate with a low pressure reading and a suitable brush, to gently bring the image into the process colour range. It will however be found to be time-consuming to do this with large areas - it is quicker to convert to CMYK and back again to obtain the same effect. There is no masking aspect to this facility either, so it is all too easy to overdo the strokes and thus remove perfectly acceptable hue/saturation values at the same time.
CMYK preview is a godsend to anyone working for print. Choosing this allows use to see the screen simulation of process colour using all the attributes we have set for a real conversion to CMYK. On the other hand, we will still be in RGB or LAB for all other working purposes. Since the eventual move to CMYK will 'force' everything into the printing range, there really is no particular need to avoid unprintable colours now that we can work without seeing them. Since many images for print are now being simultaneously used for multimedia projects which use RGB or Index Colour values, this also means that a master file for both output formats can be created at the same time.
The LAB colour model used by Photoshop was actually devised around 70 years ago - before the current technologies existed. In the digital age it is considered important because it is ædevice-independantÆ in that the values used to describe light, colour and form can be referenced in any environment, technology aside. This model, for example, has also been employed by Kodak in the creation and development of PhotoCD and the concept of ædigital negativesÆ using a customised varient known as YCC colour space. In particular, the LAB model embraces all the values of the other known models, including those listed as options in PhotoshopÆs own mode menu. It is, understandably, also the best possible place to work on images. Its problem, if it can be described as such, is that it contains a vast amount of information that goes beyond the limitations of human eyesight - let alone the previously described limitations of our computer monitors.
Of major consideration is the fact the an image in LAB colour mode contains three unique channels (or separations), one of which is 'light' itself. Converting an image to CMYK (and in many cases, Greyscale) is a matter of translating 'light' into 'tone'. Isolating the 'light' component is arguably the most crucial part of this transition since it is those values which will ultimately dictate how we built the 'masks' (inks) of tone that will obscure the purity of light that is later reflected from our paper stock. A characteristic of LAB colour is that the lightness channel will therefore contain all the information used to build the strongest 'masking' element of all - the black separation. That we have a way of accessing this information directly, especially prior to building or dictating values for our process separations, allows us to adjust the basic integrity of an image without compromising the pure values of colour contained within it.
An obvious example is the process of sharpening. A great deal of photography, both in its own right and as a result of its scanned condition, needs to be made 'crisper' than the 'reality' captured on film. The human eye perceives equally 'soft' images because detail is affected by the combined strength of all the wavelengths of light itself. When we 'see' images reproduced on paper, even very white and very reflective paper, that 'strength' of light is seriously muted by that fact that it is a reflection of the light source rather than the source itself. Since the light evident in the reflection is weaker, so too is the image created by the 'mask' of inks filtering that reflected light.
By introducing an additional black separation for process colour printing we have already enhanced the density of this 'mask' of inks. The resulting deeper shadows will therefore fool our eyes into believing the highlights are brighter than they really are. The end result is, ironically, a better impression of the reality. Unfortunately, it is rarely enough in its own right. The 'mask' of inks may need additional sharpening to exaggerate this 'tonal' difference between lighter and darker areas to better effect still.
Sharpening an entire image will not only affect these areas of highlight and shadow, but also any significant areas of the image where hue or saturation meet in transition. As a result, it is often standard practice to sharpen only the black separation in CMYK - leaving the primary colour inks alone to convey the subtleties of hue or saturation variations. The same rule can now be applied directly by using the lightness channel in LAB colour - albeit an almost impossible task in RGB mode.
Photoshop offers four built-in options for this in the filter - sharpen menu. The basic sharpen filter increases the tonal definition between all adjacent pixels in the image and, even when applied to a single channel, offers no real control over particular areas except by way of any selection mask we may have made. The sharpen more filter is simply a stronger version of the same effect. There is however a rather devious way in which these could be used using the relative tonal values of the pixels themselves to dictate the strength of the effect. The following is a variation on masking techniques.
For our demonstration we use a photograph containing a resonably large and distinct flower. It might be any object, but it is our intent to sharpen it in isolation from the rest of the image. Accordingly, we would use the magic wand tool and/or select - colour range to make a selection perimeter around its contour, before saving that selection to a channel. Whilst the selection is active, we also copy it to the clipboard. We then make two further new channels and paste the clipboard selection into position on each. On the second of the two channels we go image - map - invert to turn the selected area into negative.
The three newly numbered channels should now contain firstly a mask with a white 'hole' area, second a mask with a 'hole' occupied by a positive greyscale copy of the image and thirdly a mask containing a 'hole' occupied by a negative greyscale copy of that image component. When loaded into the image these masks will provide us with a variation on the 'relative' and 'absolute' settings used more commonly in image - adjust - selective colour alterations to ink levels.
We pretend for minute that we simply want a 50% red fill for this flower or other object. We would obviously mix our colour and use either fill or the paintbucket set to 50% foreground colour. Since the fill is to be contained within our flower/object we will load the selection mask for the perimeter zone. Depending on which of these three mask variants we load, the effect of the fill will differ. Loading the first channel will allow us to fill the entire flower with an 'absolute' 50% tint of our chosen colour. When we load the second channel the tonal range of the positive image copy is reflected in our mask. Dark areas of the flower will dark tones on the mask, whilst light areas will be off-white but almost a full 'hole'. The resulting fill will therefore be something less than 50% in any light areas that are less than pure white, whereas the darker areas will only receive the most marginal of tints. By loading the third channel where the mask contents are negative we get the opposite effect. Here the dark areas get more of the tint and the light areas get less. This latter example is akin to printing a greyscale copy of the flower in 50% transparent red ink on top of the real colour image.
Channels loaded as masks control the behaviour of anything being applied through them. In the case of the above example this was simply the paint's opacity. If, however, we apply an effect rather than paint, then it is the strength of the effect that is dictated by the mask. In the case of our flower/object, we can now apply the sharpen and sharpen more filters in such a way that the strength of the sharpening is dictated by the greyscale tonal values of the flower/object itself. With some time and the skill of experience, we can extend this further. For example we might load the first made channel onto the second and make additional adjustments to the flower area on the mask itself by using the image - adjust - brightness/contrast or image - adjust - curves controls. We could also load the negative mask onto the positive one and make similar adjustments or, indeed, load a mask onto itself to do the same thing. The permutations are endless.
The filter - sharpen - sharpen edges option offers somewhat more definition in its own right. This finds areas of significant pixel value shifts and considers them to be edges. It then sharpens by improving the definition of the adjacent pixels along those edges. Once again. the black separation or lightness channel offer distinct tonal reference for this exercise.
The last of this group is the unsharp mask filter and it is by far the most common choice for improving the clarity of images for print, Since it also works on 'edges', but with the additional facility to 'spread' those edges, this particular filter can effectively bring badly focussed photography back into focus - well beyond the requirements of basic print-specific sharpening. It is controlled with a dialogue box and a sample area display which we can choose by clicking somewhere on the image and enlarge or reduce with the +/- icons. If the box marked preview is checked, the results will be seen in the image simultaneously.
The radius slider relates to the aforementioned 'spread' factor since we use it to decide how many pixels around the edges we want to affect. The amount slider, which ranges between 1% and 500%, controls the strength of the sharpening effect we want to apply. The threshold slider is used to control the sensitivity of the edge-detection facility. A setting of zero will read all areas where the 256 greyscale tonal values change. That's a lot of edges and a lot of overall sharpening. Increasing this value will lessen the areas involved in the sharpening since it will discard a greater number of tonal 'jumps' from the operation. At its maximum, 255, it will sharpen only areas where black and white join. This will be fairly easy to see on a channel. On a colour composite we are bext able to gauge the effect by considering the image itself as a greyscale.
Most sharpening operations are actually achieving their effect by increasing the pixel contrast in selected areas. For this reason, general sharpening can often be obtained by using image - adjust - brightness/contrast to slightly boost the contrast only on the lightness or black channels, possibly within masked areas.
A variation on the same technique is to select the entire composite image, copy it to the clipboard and then paste it onto a new channel. We then improve the contrast of that channel (or use other map adjustments like threshold and posterise) until the distinctions between the areas want to affect are well defined. This channel, when re-loaded onto the composite image or the black/lightness channel, allows us to adjust contrast within the relative areas created on the adjusted channel, with the emphasis on the lighter ones. To switch the emphasis to darker areas, we would simply load an inverted variant of the same adjusted channel.
This technique can also be employed to create masks for colour corrections. Many colours become muted when converted to CMYK and some redefining of balance is needed in order to draw the viewer's attention back to areas which were more prominent in the original image. Isolating those areas on a channel mask created from the original itself also provides a good visual reference as to where those same values might also exist and be affected in less important areas of the image. Whilst this technique will often work prior to converting to CMYK, especially if using CMYK Preview, it should be remembered that full conversion to separations will probably be necessary if we want to access the detailthe modified colours themselves.
Compositing of images and major colour adjustments to the resulting canvas are best done in RGB (or possibly LAB colour) and recent additions to the image - adjust menu include selective colour which allows us to modify the work using hypothetical CMYK inks. This facility has been discussed earlier but there will be limitations to the overall adjustment of an image when the exact nature of the black separation has not yet been calculated. Many colour enhancements for print purposes are best done simply by going to the individual CMYK separations and modifying them after they have been produced.
The Preferences - Printing Inks Set-up in Photoshop does apply a real change to the file at the point of CMYK conversion and provides us with an accurate on-screen visual with which to work. This conversion also, as we have seen, makes serious changes to the file in the way that the separations themselves are generated. We must this bear in mind if we even remotely anticipate the need to return back to another colour mode. Re-conversion to RGB for example will, unless we later create new colours, always contain the limited CYMK mixtures but as re-constituted RGB colours. If, as will be required, we need to apply different inks in order to run a matching-compatable proof on a laser printer, the chances are that crucial information will have been lost as primary CMY values were reduced to accomodate our black generation. We may also need extra dot gain for the proofing device, but only the difference between the expected dot gain of that device and the amount we have applied already. This is fine for the final job, less so for the laser printer which is an altogether different beast.
Laser printers exhibit a dot gain (or spread of ink) that exceeds that of the worst newspaper stocks. The 'inks' are actually magnetic powders which are melted and sealed onto the paper surface using heat. The end result may look quite slick, but the nature of the process is such that the integrity of any image will be ruined. It may even be largely unrecogisable unless the final job was set up for newspaper inks - the nearest relation in the chain. This is an example, but many desktop and studio printers in use today have little in common with the final output method. To obtain a reasonably accurate 'dry' proof we need to print a separately configured version. This may be a copy or it may be a temporary change to the real file. Either way, it must be made or done before converting the master file to CMYK inks.
Photoshop's Preferences - Printing Inks Set-up dialogue changes 'on-the-fly' - that is to say, we can re-apply inks at any time. Nor are the settings in any way really 'locked' to the file. The dialogue always shows the last setting made on screen and will display all CMYK images according to those settings. Whenever we print a proof from Photoshop it will use the last setting as a guide but the instructions sent to the device will principally be those contained in file - page set-up. These, as we saw earlier, may have been customised with attributes for exporting the final job to another destination program. They will almost certainly be inappropriate for the proof, but are nonetheless ælockedÆ to the file. In order to actually print a proof we need a working copy in which we can enter different commands for the device running that same proof. We can make this best with image - duplicate.
The hardest decision is what to do next. If we have already generated CMYK separations for this file at any stage we are stuck with those changes whatever mode we are currently in. For the proof to appear 'correct' we must not only enter attributes for the proofing device, but also add extra compensation for the differences in the 'ink' behaviour between the final output and the current proofing device. If, for example, we added a serious volume of density to the black component, we need to reduce it again here and then increase it again or decrease it further according to the requirements of the proofing device. This will mean re-configuring the information in page set-up - transfer (button). In most cases reducing the amount of black using a curve representing the opposite of that applied during the separation set-up will suffice for practical purposes.
It is worth a further mention of the Kodak colour management system which is built into Photoshop and an increasing number of other software products. Device independent, this system can build and record a colour profile for an image at the input stage such as scanning, then convert the same image to reflect the attributes of a known printing device. The attributes of this output are recorded as a device profile and can be kept in the KPCMS folder in our system folder. When opening a PhotoCD image for example, we can usually choose either LAB colour or RGB as our destination in Photoshop. If we have been given or created additional printing device profiles, this operation can be extended to opening direct in CMYK, in which case the necessary information for our printer will be applied automatically before we see the result on screen. Although some kind of sharpening operation may be necessary, this system speeds up conversion and processing time considerably.
Having used Photoshop to prepare or adjust an image we are commonly faced with the task of exporting the image to another destination. Except in the case of minor editorial uses, the PICT format is to be avoided since there can be some loss of quality. The preferred choices are normally EPS and TIFF. The latter provides the best quality since it retains all pixel information, but the EPS standard allows for the inclusion of the precision pre-press data mentioned earlier. It is not uncommon to find documents that contain a mix of EPS and TIFF images, but we should bear in mind that the more complexity of information in a file the harder it is to process.
There are two instances in particular that have some bearing on our decision for the export format. The first concerns 'clipping paths' whereby we can convert the rectanglar shape of the Photoshop canvas into a precision shaped file. For this we must use the EPS standard or work with a line illustration program to merge a TIFF image with such a shape. The second instance is where we may wish to create one or more additional 'spot' colours to print in combination with process inks. This involves creating extra channels which can only be exported after being removed from the composite image. Since these are often areas of 'flat' colour that are also contained in 'shaped' segments inside the main image, a combination of TIFF and EPS may be desired to acheive the best results.
The process of making a 'clipping path' which will ultimately enclose our image in a defined shape is simply a matter of building a selction mask over the relevant area and converting it to a pen path in the paths palette. How complex that path may be is another matter. As an example we must imagine an image containing something like dounuts, since viewed from above these would each contain both an outer contour and an inner contour. Since we can only export one individual 'clipping path' we must take steps to ensure this complexity is all represented.
Photoshop's paths palette replicates the drawing procedures utilised in Postscript line art programs like Illustrator and Freehand. By converting a selection shape into a path we can edit it using the Postscript simulation before re-coverting it into a true Photoshop selection. We can also use the paths palette as a 'shapes' library and export any or all of these shapes to the Illustrator or Freehand programs. Any path, although it may sound like a singular entity, can in fact hold a number of different shapes as part itself. This could, for example, be three dounut outlines. When any physical outline contains another outline shape which is wholly enclosed within it, Photoshop now automatically applies the Postscript attribute known as an 'even/odd fill rule' by which those inner shapes are considered to be travelling in a reverse direction, thereby 'knocking a hole' in the greater outline shape.
To export three isolated dounuts we must create a single path in the paths palette which threfore contains three outline shapes, each with another 'inner' shape contained inside it. The simplest way to achieve this is to work with Photoshop's selection tools to pick up the background areas of the image behind the dounuts. We then add the inner 'holes' of the dounuts to this selection before going to select - inverse to effectively switch those areas themselves into the masked ones. In difficult cases, we may need to artwork this on a channel to build a precision mask. This is discussed in the design techniques section. We then click on the dotted circle icon at the bottom of our paths palette to convert all the contours into a single path, then double click on the named 'work path' to give it an acceptable title. We can also modify this further by adjusting the points and handles which will then be available on the path itself.
Exporting this file using the EPS format will allow us to choose the entire path we just titled as the outline shape for the image. In this case there will be three distinct areas, each with another 'knockout' area contained within it. There will however be two drawbacks to this basic technique. Firstly, the contours themselves will dictate the 'hard edges' common to the postscript language. Secondly, the pixel contents of those shapes will also be described as smaller ojects with coloured fills - as is the norm with Postscript - which may lose some marginal image quality.
A slightly more long-winded technique may in some cases compensate for the latter problem. This involves exporting the file as a rectangular TIFF image, whilst using a program like Illustrator or Freehand to apply the path shapes already made. To do this we go file - export - paths to illstrator and chhose the titled path as before. We also save a TIFF file. Crossing to Illustrator or Freehand, we open the Illustrator format file just created with the path in it. We than place the TIFF image and position it aligned on top of the path. Lastly we cut the TIFF to the clipboard, highlight the path in the document and use the paste inside command to put the TIFF image within the shapes. We can now save this entire file as an EPS with the include TIFF images function set in the preferences. This will arrive in a page layout program as if we had created a clipping path, but with a possible improvement in image quality.
There is also a technique by which we may in some cases be able to use one of the above methods and still retain Photoshop's soft edges. In an example we use Quark Xpress as the destination program and go the the page where our image is destined to reside. If the requirement is for text to 'runaround' the image shape against a coloured page background, we then choose file - save page as EPS then return to Photoshop. In Photoshop we file - open this EPS as a canvas and then montage the dounut cut-outs onto that surface. We still need to employ some outlining technique for the file, but the outlines themselves can be expanded slightly to include soft edges, since they already contain colour which matches the destination.
Note: The above 'colour pinching' is arguably not needed if we take readings of the colour values from the Quark page and then mix them in Photoshop to use as a background. In most cases the printed result will be fine, even through the on-screen visual provided by Quark may demonstrate differences in the colour match of this background mix. The technique is rather a safety precaution and making colour palettes as EPS files in Quark is recommended for frequently used documents.
Projection: Quark will also soon be launching an extension program for Xpress called Xposure. This will be an image-editing and montage facility which employs mathematically defined command lists to instruct 'Photoshop-style' changes to any image. This technology is similar to that used by the X-Res and Live Picture programs and will be discussed shortly. Real-time previews will be available on-screen and the application should come with an Xtension module for use in Xpress itself. This will probably eliminate the need for the techniques covered above.
Although the overall sharpening of images is a particular requirement for print, the 'soft edges' mentioned above are best obtained by use of Photoshop's blurring facilities - the opposite effect. This should be applied only around the contours of a component and by no more than a few pixels. The 'feather' effect discussed in the overview and montage sections will not always be appropriate since the 'vignette' effect obtained is actually a 'fade-out' of the image over a certain area - even if that area is quite small. In the case of a 'shaped' file, the path created for export may also be used to soften those same edges.
To demonstrate this we zoom into some detail and gauge the number of edge pixels we want to affect. We then go to the brushes palette and double click on an empty position to create a new brush. The diameter should be the amount of pixels just estimated. It should be totally hard and round, with spacing around 20%, We then choose this brush for the blur tool, set to normal and around 30% pressure. Returning to our paths palette, we drag the path name onto the icon at the palette's bottom which displays a bold circle outline. This paints, or in this case applies an effect to, the image using the path as its guide. We see the result and undo.
Whilst the above may be sufficient for our needs, it is often slightly more effective if applied to the colour channel where most detail is contain. We look at each colour channel to gauge this, then repeat the blur stroke on that channel only. Then we compare the result in the composite image. If the blur is too slight, we can re-apply the effect using variations on the pressure setting. An very obvious blur should be avoided since the path itself will be making its 'cut' halfway across the diameter of the effect.
Tip: Regardless of the colour mode we are working in, the colour channels making up our document often exhibit differences in tonal values that are less than clear in the composite. When using the magic wand tool or the colour range facility to obtain our selection masks, it can save time to pick up the ranges on the channel where they are clearest and then flip back to the composite channel with the selection still active.
There are a number of third party plug-ins and applications that can be used in conjunction with Photoshop in production for print. If the adjustments and conversions we are applying to an image are going to be repeated, in is a good idea to create command keys that will provide quick access to dialogue box and also to save settings in those boxes that we may wish to reload later. For the processing of 'batch' jobs, it may be well worth investigating various 'macro' plug-ins that will record and repeat entire tasks automatically, including those that are pre-programmed to sharpen, adjust and convert images with the separations for particular paper stocks in mind.
There are two programs which use similar technologies to aid the task of montaging large files. X-Res and Live Picture both create 'substitute' files from large pixel canvases, allowing us to combine these images using mathemtical definitions which describe the information in precise 'object-orientated' form. All the decisions we take and the changes to the works are recorded as a series of delayed commands which are later condensed to the minimum needed to execute the task before being applied to the real pixel images later. The memory requirements when working this way are a fraction of those required in normal bitmap/pixel editing such as with Photoshop. Quark's new Xposure program employs similar concepts for working on invidual images. Due to the absence of pixel information and subsequent detail, these programs enhance productivity by providing a complement to Photoshop itself, rather than any real substitution.
Colour and other adjustments are often made to entire images and the facilities offered by the programs above will speed this process. On other occasions, we need to build a mask to make changes only to selected parts of the image. Whilst this too can be done by the above over object-defined areas, we sometimes need to use exact pixel information to create masks which apply 'relative' value changes rather than simple 'absolute' or 'blanket' changes. As mentioned earlier, we might for example want to mask a colour image of a flower with a greyscale mask which includes a copy of the flower's tonal range itself.
When we need to make repeated adjustments using this 'relative' information it usually requires saving a greyscale mask on a channel. This also increases our file size by up to one third. If the area to be affected is a small part of the overall image, bear in mind that, when selected, we can define that area as a brush - either with a solid ('absolute') fill or with a positive/negative picture of that part of the image as its ('relative') fill.
There will be instances when we need to use Photoshop to create additional separations for our process colour file. This might, for example, be an extra Pantone colour to be used for a single colour graphic or some text in a selected part of an otherwise full colour image. We will now look at the method of creating this.
We will assume we are superimposing a simple logo - a crescent moon shape with a cut set into it. We can create this using the eclipse selection tool to carve a rough circle, then hold the command key and carve another rough circle overlapping it. We then hold command again with the lasso and create the 'cut' in the resulting crescent, before saving the selection to a channel. For best results we should do this after moving to CMYK mode.
We load this selection into the image and delete to white. We take a look at each channel and set that a 'hole' is clear in each. Next we move to the channel containing the selection itself and choose image - map - invert which, by making a negative image of our mask stencil, actually gives us a positive separation for the spot colour. Next we double-click on the channel's name and remix it with the Pantone required, setting the opacity to 100%. This is now an aligned separation in its own right and we can see a complete visual by returning to the composite channel and switching our new separation on using the 'eye' column.
Photoshop offers a trapping facility in CMYK mode only. This will spread colour where needed to avoid visible 'cracks' where separations join as absolutes and with no overlapping information. It uses predefined tables and our only contribution to the process is determining the amount of spread. This exercise is rarely needed for photographic images or other artwork employing 'natural' media. Since it works by calculating across the process colours, nor can it be used for this new separation. Instead we must use the magic wand tool to select the solid parts of the new separation and then use select - modify - expand to spread the solid colour outwards by the required amount. Since this only expands the selection outline, we must also refill the area with black. We then have a manually applied form of trapping.
This separation can be printed direct from Photoshop simply be selecting it from the channels list before going to file - print. We can then print the other separations by going to the composite channel and doing the same. Unfortunately, export is more difficult. The composite image can be exported in its present state as a TIFF file, but this will not include our extra separation. We must export that separately.
To do so, we must separate the new separation from the main image. We go image - duplicate to get an exact copy of the whole file, then choose split channels in the channels palette sub-menu of that copy. This will dismantle the whole document to give us several one channel files. The most recently made Pantone opne will be topmost. We can now save this file independantly although we should make a clipping path around the solid colour areas and save in EPS form to export both the Pantone reference and the shape for the overlay.
A similar technique is used to create silk-screen designs for printing directly from Photoshop, as if preparing hand separations. We can also employ it for creating two colour (or more) documents for print. We make a new greyscale canvas, then click the new channel icon to get an empty channel. We then change to multichannel mode and remix the colour for each channel as a 100% Pantone (or other). By painting or pasting greyscale components on these two channels we can now built up a two colour image.
Photoshop does not, at present, offer any way of combining the duotone mode with any two-colour document we may have created, although they can be exported separately and combined on a page in a destination program. In graphical terms only, both modes can be combined in any other standard colour mode as montage components. Officially, that is.
There may be a way around this in some situations. Duotone, as we know, is created using output curves to represent the variations in the physical separations. Those separations cannot be accessed by us directly. Since duotones can only be made from greyscale originals, we can to some extent work with multiple copies of the greyscale itself to create in advance what will ultimately be our duotone 'separations'. The technique is similar to the above. We duplicate copies of the greyscale image onto separate hand mixed channels. We then go to each channel in turn and adjust its appearance using the image - adjust - curves dialogue. We will see the effect as it happens if we have the 'eyes' of our other colour channels switched on. When we see the desired result, we save the curve for that effect and it will later be loadable in the dialogue box for the duotone curve for that ink.
There is of course nothing to stop us doing this for selected areas of an image only. If we create a multichannel document for two colours as we did previously, we could paste a greyscale component into each of its channels and, provided they are properly aligned, could then adjust the curve for that part of the composite only - on each if needed. The result is identical to the effect obtained in duotone mode itself - but with two actual separations where we can combine those curves with our own custom areas of solid. An example might be solid blue text superimposed on a blue/brown duotone.
Photoshop's duotone mode also includes tritones and quadtones for three or four colour printing. A hand-crafted quadtone could equally be created with the four process colours. In a similar fashion, the above technique can be used to re-create the four process colours in multichannel mode, with the effect obtained then pasted into a CMYK mode document. Once again, care should be taken with our alignment and some monitoring of the channels themselves will be in order to check that our superimposed components are not in conflict with the amount of black generated at our CMYK conversion point.
Many effects and enhancements are available by employing these methods to different parts of a CMYK image. Some, like removing grain components separations, can be used to create science-fiction style landscapes. Others, like slight input curve adjustments to different areas of an image, can be used to boost the feeling of distance between foreground and background components. It is ultimately up to the user's own judgement how far an image can be acceptably be manipulated in the interests of enhancement for print.
We conclude this section with a word about screen-free digital printing. This process may ultimately replace the conventional screening technologies since, once established, it should be more cost effective and will also result in the need for slightly less high resolution file sizes.
Stochastic printing, as it is known, effectively generates separations for process colour by direct pixel to microdot translation. Rather than the conventional screening approach, which applies a line frequency uniformly over each separation, the new method creates clusters of dots according to the density of pixel values in the electronic separation. Where four pixels are required to calculate screen angles and dot shape using conventional screens, it is possible that only a single pixel will be required to represent the equivilent microdot.
The advantages are obvious, although one possible drawback is that the process is unsuitable for large areas of flat colour. Looked at closely, flat areas gain a degree of texture - although arguably this is an enhancement for many images.
For the present, a reminder to follow the rules for resolution. The figure used in our printer's line screen frequency should be doubled to estimate resolution for an actual size image. Ie: 133 lpi indicates a same size image resolution of 266dpi. It is advisable to add 20/30dpi to this figure if we are exporting files and are likely to engage in marginal resizing at the destination.
The rule of thumb is that once made, we cannot size upwards in Photoshop - by dimensions or resolution. Increasing resolution will create artificial duplicate pixels and may seriously blur or otherwise corrupt the image. Increasing dimensions will effectively do the same by inventing new pixels to occupy the greater space. If in doubt, we go for the largest possible size and resolution - sampling down later is no problem. There are a number of techniques covered elsewhere for disguising low resolution images, but these are just that - they are not a substitute for the real thing.
And in so far as our printer's concerned - keep talking!