Martin Evening Adobe Photoshop CS4 for Photographers 258 Reduce Noise fi lter The other alternative to noise reduction in Camera Raw is to use the Reduce Noise fi lter, which is found in the Filter ➯ Noise menu in Photoshop. This fi lter uses a method of smart noise reduction that can remove noise from an image, but without destroying the edge detail in the picture. Overall, the Reduce Noise fi lter is a useful one shot fi lter that is better at reducing the noise in images that originated as digital captures, rather than reducing fi lm grain noise from scanned 35mm images. This fi lter is mainly designed to target the twin problems of digital luminance noise, which is like a very fi ne speckly grain, and color noise, Figure 4.12 Here is the Reduce Noise fi lter being used to help remove the noise from a digital capture which was shot at 1250 ISO, where the Reduce Noise fi lter helped get rid of most of the noise artifacts. 259 Sharpening and noise reduction Chapter 4 Figure 4.13 A close-up view of the Reduce Noise fi lter settings in Advanced mode. which is commonplace with digital captures shot at high ISO settings. The only problem with this fi lter is that it is quite memory-intensive, so be prepared for a wait while it performs its calculations. Although it can appear quite effective at removing heavy noise, if you do have to apply extreme settings you can sometimes end up with an enhanced noise pattern instead. In Basic mode you can simply adjust the strength of the noise reduction and then use the controls below that to modify the noise fi ltering; these should be adjusted in the order they are displayed. The Strength slider adjusts the amount of noise reduction that is applied, while the Preserve Details slider helps preserve the edge luminance information. The luminance noise reduction will appear strongest when you set Preserve Details to zero %, but as you increase Preserve Details more edge detail (and often more noise) will become visible. Below that is the Reduce Color Noise slider, which allows you to separately control the color noise suppression. After you have adjusted all of the above settings, it is very likely that the image will have suffered some loss in sharpness. The Sharpen Details slider allows you to dial back in some detail sharpness. But I would urge caution here, because adding too much sharpening can simply introduce more artifacts. Advanced mode noise reduction In Basic mode you are only able to adjust the Reduce Noise settings so that they affect the overall strength and preservation of image detail. When the Advanced mode button is checked you can apply the noise reduction adjustments on a per channel basis. This can be useful if you wish to apply differential noise reduction to individual channels. Whether you are treating a digital capture or scanned fi lm emulsion the Blue channel is usually the noisiest, so it can therefore be a good idea to apply more reduction to this channel and less to the Red and Green channels where the noise is usually not such a major problem. Color bleed caution There are times when you may want to crank up the Color Noise Reduction to 100% in order to remove as much of the noise artifacts as possible, but be aware that adding too much Color Noise Reduction can sometimes cause colors to bleed badly and cause too much softening of the image. Martin Evening Adobe Photoshop CS4 for Photographers 260 JPEG noise removal You can also use the Reduce Noise fi lter to smooth out JPEG artifacts. If you have a heavily compressed JPEG image, the Reduce Noise fi lter can certainly help improve the image smoothness. But I reckon you can also use the Reduce Noise fi lter in this mode to improve the appearance of GIF images as well. Of course you will need to convert the GIF image from Indexed Color to RGB mode fi rst. But once you have done this you can use the Reduce Noise fi lter adjustments to help get rid of the banding by taking the Preserve Details slider down to zero % and raising the Sharpen Details to a higher amount than you would be advised to use normally. Saving the Reduce Noise settings Favorite Reduce Noise settings can be saved by clicking on the Save Changes to Current Settings button. And Reduce Noise settings can be deleted by clicking on the trash icon next to it. 1 The Reduce Noise fi lter has a Remove JPEG Artifact option that can be useful if you wish to improve the appearance of an image that has suffered from over-heavy JPEG compression. But it can also help rescue a GIF image where a lot of the color levels information has been lost in the conversion to Indexed Color mode. 2 A GIF image will have to be converted to RGB mode fi rst. You can then apply the Reduce Noise fi lter. In this example I checked the Remove JPEG Artifact box. To remove the color banding, the Preserve Details had to be set to 0%, and to make the image sharp again I increased the Sharpen Details to 70%. Figure 4.14 Accessing Reduce Noise settings. J Y J i t R o t o fi t S a Fi gu re 4.14 Ac ce ss in g Redu ce Noi se s et ti ng s 261 S o far I have shown you just how much can be done when editing a photo in Camera Raw, before you bring it into Photoshop. Some of the techniques described in this chapter will appear to overlap with Camera Raw, but image adjustments such as Levels and Curves play an important role in everyday Photoshop work, plus this chapter also explains how to work with images that have never been near Camera Raw, such as images that are supplied directly as TIFFs or JPEGs. I’ll start off by outlining a few of the fundamental principles of pixel image editing such as bit depth and the relationship between resolution and image size. After that we’ll look at the main image editing adjustments and how they can be used to fi ne-tune the tones and colors in a photograph. Chapter 5 Image Editing Essentials Martin Evening Adobe Photoshop CS4 for Photographers 262 Pixels versus vectors Digital images are constructed of pixels and as such are resolution-dependent. You can scale a pixel image up in size but, as you do so, the fi nite information in a pixel image can only be stretched so far before the underlying pixel structure becomes apparent. Objects created in programs like Adobe Illustrator are defi ned mathematically so if you draw a rectangle, the proportions of the rectangle edges, the relative placement on the page and fi ll color can all be described using mathematical expressions. An object defi ned using vectors can be output at any resolution; it does not matter if the image is displayed on a computer screen or as a huge poster, it will always be rendered with the same amount of detail (see Figure 5.1). Figure 5.1 Digital images are made up of a mosaic of pixels. This means that a pixel-based digital image will always have a fi xed resolution and is said to be ‘resolution-dependent’. If you enlarge such an image beyond the size at which it is meant to be printed, the pixel structure will soon become apparent, as can be seen here in the left-hand close-up view. But suppose the picture shown above originated not as a photograph, but was drawn as an illustration using a program like Adobe Illustrator. If the picture is drawn using vector paths, the image will be resolution- independent. The mathematical numbers used to describe the path outlines shown in the bottom right example can then be scaled to reproduce at any size: from a postage stamp to a billboard poster. As you can see in the comparison shown here, the pixel image starts to break up as soon as it is magnifi ed, whereas the outlines in the vector-drawn image will reproduce perfectly smoothly. ‘Stalkers’ by The Wrong Size. Photograph: © Eric Richmond. Pixel-based image Vector-based image 263 Image editing essentials Chapter 5 Confusing terminology You can see from this description where the term ‘lines per inch’ originated. In today’s digital world of imagesetters, the defi nition is somewhat archaic, but is nonetheless commonly used. You may hear people refer to the halftone output as ‘dpi’ instead of ‘lpi’, as in the number of ‘halftone’ dots per inch, and the imagesetter resolution referred to as having so many ‘spi’ or ‘spots per inch’. Whatever the terminology I think we can all logically agree on the correct use of the term ‘pixels per inch’, but I am afraid there is no clear defi nitive answer to the mixed use of the terms ‘dpi’, ‘lpi’ and ‘spi’. It is an example of how the two separate disciplines of traditional repro and those who developed the digital technology chose to apply different meanings to these same terms. Photoshop as a vector program Photoshop is mainly regarded as a pixel-based graphics program, but it has the capability to be a combined pixel and vector editor because it does also contain a number of vector-based features that can be used to generate things, such as custom shapes and layer clipping paths. This raises some interesting possibilities because you can create various graphical elements like type, shape layers and layer clipping paths in Photoshop and these are all resolution-independent. These ‘vector’ elements can be scaled up in size in Photoshop without any loss of detail, just as with an Illustrator graphic. Terminology Before proceeding further let me help clarify a few of the confusing terms used and their correct usage when describing resolution. ppi: pixels per inch This describes the digital, pixel resolution of an image. But you will notice the term ‘dpi’ is often inappropriately used to describe the digital resolution of images as well. This is an incorrect use of the term ‘dpi’ because input devices like scanners and cameras don’t produce dots, they produce pixels. Only printers can produce dots! However, it’s become quite common now for scanner manufacturers and other software programs to use the term ‘dpi’ when what they really mean is ‘ppi’. Unfortunately this has only added to the confusion, because you often hear people describing the resolution of an image as having so many ‘dpi’. But if you look carefully, Photoshop always refers to the input resolution as being in pixels per inch or pixels per centimeter. So if you have an image that has been captured on a digital camera scanned from a photograph, or displayed in Photoshop, it is always made up of pixels and the pixel resolution (ppi) is the number of pixels per inch in the input digital image. lpi: lines per inch This is the number of halftone lines or ‘cells’ in an inch (also described as the screen ruling). The origins of this term go back way before the days of digital desktop publishing. To produce a halftone plate, the fi lm exposure was made through a fi nely etched criss-cross screen of evenly spaced lines on a glass plate. When a continuous tone photographic image was exposed this way dark areas formed heavy halftone dots and the light areas formed smaller dots, which when viewed from a normal distance gave the impression of a continuous tone image on the page. The line screen resolution (lpi) is therefore the frequency of halftone dots or cells per inch. Martin Evening Adobe Photoshop CS4 for Photographers 264 dpi: dots per inch This refers to the resolution of printing devices. An output device such as an imagesetter is able to produce tiny 100% black dots at a specifi ed resolution. Let’s say we have an imagesetter capable of printing at a resolution of 2450 dots per inch and the printer wished to use a screen ruling of 150 lines per inch. If you divide the dpi of 2450 by the lpi of 150, you get a fi gure of 16. Within a matrix of 16 × 16 printer dots, an imagesetter can generate a halftone dot varying in size from 0 to 255, which is 256 print dots. It is this variation in halftone cell size (constructed from the combined smaller dots) which gives the impression of tonal shading when viewed from a distance. Desktop printer resolution In the case of desktop inkjet printers the term ‘dpi’ is used to describe the resolution of the printer head, and the dpi output of a typical inkjet can range from 360 to 2880 dpi. Although this is a correct usage of dpi, in this context the dpi means something else yet again. Most inkjet printers lay down a scattered pattern of tiny dots of ink that accumulate to give the impression of different shades of tone, depending on either the number of dots, the varied size of the dots, or both. The principle is roughly similar to the halftone process, but not quite the same. But as you might expect, if you select a fi ner print resolution such as 1440 or 2880 dpi, you should expect to see smoother print outputs when these are viewed close up. While a correlation can be made between the pixel size of an image and the ‘dpi’ setting for the printer, it is important to realize that the number of pixels per inch is not the same as the number of dots per inch created by the printer. When you send a Photoshop image to an inkjet printer, the pixel image data is processed by the print driver and converted into a form that the printer uses to map the individual ink dots that make the printed image. The ‘dpi’ used by the printer refers to the fi neness of the dots. Therefore a print resolution of 360 dpi can be used for speedy, low quality printing, while a dpi resolution of 2880 can be used to produce higher quality print outputs. 265 Image editing essentials Chapter 5 Megapixels to megabytes If you multiply the ‘megapixel’ size by three you will get a rough idea of the megabyte size of the RGB image output. In other words, a 12 megapixel camera can produce a 36 MB RGB, 8-bit per channel image. Quoting megabyte sizes is a less reliable method of describing things because document fi le sizes can also be affected by the number of layers and alpha channels present and whether the fi le has been compressed or not. Nevertheless, referring to image sizes in megabytes has become a convenient shorthand when describing a standard uncompressed, 8-bit per channel fl attened TIFF image. Choosing the right pixel resolution for print There have been theories about choosing the appropriate pixel (ppi) resolution to match the dpi resolution of the printer. For example, it has in the past been suggested that the optimal pixel resolution should ideally be the printer dpi divisible by a whole number. Therefore, if you intended printing at 2880 dpi, the following pixel resolutions could be used: 144, 160, 180, 240, 288, 320, 360. More recently, this theory has been displaced as it has been shown that there isn’t really a need to make the pixel resolution match any particular formula in relation to the dpi setting used on the printer. Image resolution What are the minimum number of pixels required to print at a particular size? Plus, what is the relationship between the pixel dimensions and image resolution? These questions crop up time and time again. Digital cameras are usually classed according to the number of pixels they can capture. If a CCD chip contains 3000 × 4500 pixel elements, it can be said to capture a total of 13.5 million pixels, and therefore be described as a 13.5 megapixel camera. When we talk about the resolution of an image we are principally referring to the number of pixels that are contained in the picture. Basically, every digital image contains a fi nite number of pixels and the more pixels you have, the greater potential there is to capture more detail. The pixel dimensions of an image are an absolute value. Therefore, a 2400 × 1800 pixel image will contain 4.32 megapixels and this is an absolute measurement of how much information is contained in the image. But a digital image of this size could be printed at 12” × 9” at 200 pixels per inch, or it could be printed at 8” × 6” at 300 pixels per inch. So if you want to know how big an image can be printed, you simply divide the number of pixels along either dimension of the picture by the pixel resolution you wish to print at (see Figure 5.2). This can be expressed clearly in the following formula: the number of pixels = physical dimension × (ppi) resolution. In other words, there Martin Evening Adobe Photoshop CS4 for Photographers 266 is a reciprocal relationship between pixel size, the physical dimensions and resolution. If you quote the resolution of an image as being so many pixels by so many pixels, there can be no ambiguity about what you mean. Figure 5.2 In this diagram you can see how a digital image comprised of a fi xed number of pixels can have its output resolution interpreted in different ways. For illustration purposes let’s assume that the image is 40 pixels wide. The fi le can be printed small at a resolution of 40 pixels per cm, or printed big (and more pixelated) at a resolution of 10 pixels per cm. Repro considerations The structure of the fi nal CMYK print output bears no relationship to the pixel structure of a digital image, since a pixel in a digital image does not equal a cell of halftone dots on the page. To explain this, if we analyze a CMYK cell or rosette, each color plate prints the screen of dots at a slightly different angle, typically: Yellow at 0 or 90 degrees, Black: 45 degrees, Cyan: 105 degrees and Magenta: 75 degrees. If the Black screen is at a 45 degree angle (which is normally the case), the (narrowest) horizontal width of the black dot is 1.41 (square root of 2) 267 Image editing essentials Chapter 5 1 The halftone screen shown here is angled at zero degrees. If the pixel resolution were calculated at x2 the line screen resolution, the RIP would use four pixels to calculate each halftone dot. 2 To reproduce a CMYK print output four plates are used, of which only the yellow plate is actually angled at zero degrees. The black plate is normally angled at 45 degrees and the cyan and magenta plates at less sharp angles. Overlay the same pixel resolution of x2 the line screen and you will notice that there is no direct relationship between the pixel and line screen resolutions. 3 There is no single empirical formula that can be used to determine the ideal ‘half toning factor’. Should it be x2 or x1.5? The black plate is the widest at 45 degrees and the black plate information is usually more prominent than the three color plates. If a half toning factor of x1.41 (the square root of 2) were used, the pixel resolution will be more synchronized with this angled halftone screen. There is no right or wrong half toning factor – the RIP will process pixel data at any resolution. If there are too few pixels, print quality will be poor. But having more than the optimum number does not necessarily equate to better output, it just means more pixels. 4 Each halftone dot is rendered by a PostScript RIP from the pixel data and output to a device called an imagesetter. The halftone dot illustrated here is plotted using a 16 x 16 dot matrix. This matrix can therefore reproduce a total of 256 shades of gray. The dpi resolution of the imagesetter, divided by 16, will equal the line screen resolution. 2400 dpi divided by 16 = 150 lpi screen resolution. The relationship between ppi and lpi [...]... x 15 3000 x 2000 6 17 .5 23 .5 15 x 10 37 .5 x 25 10 x 6 .5 25 x 17 350 0 x 250 0 8. 75 25 33 .5 17 .5 x 12 .5 44 x 31 11 .5 x 8 .5 29 x 21 4000 x 2 850 11.4 32 .5 43 .5 20 x 14 50 x 36 13 .5 x 9 .5 33 .5 x 24 450 0 x 3200 14.4 41 54 .5 22 .5 x 16 56 x 40 15 x 10 .5 37 .5 x 27 50 00 x 4000 20 57 76 25 x 20 62 .5 x 50 16 .5 x 13 .5 42 x 33 .5 Figure 5. 6 The above table shows a comparison of pixel resolution, megapixels, megabyte... that for really extreme image size changes they still prefer the 10% step method Chapter 5 Image editing essentials Inches Centimeters Inches Centimeters Pixel size Megapixels MB (RGB) MB (CMYK) 200 ppi 80 ppc 300 ppi 120 ppc 1600 x 1200 2 6 7 .5 8x6 20 x 15 5 .5 x 4 13 .5 x 10 2400 x 1800 4.3 12 .5 16 .5 12 x 9 30 x 22 .5 8x6 20 x 15 3000 x 2000 6 17 .5 23 .5 15 x 10 37 .5 x 25 10 x 6 .5 25 x 17 350 0 x 250 0... ppi 6 .5 26.1 A4 Magazine double page 133 lpi 200 ppi 8 29.6 266 ppi 13 52 .2 A4 Magazine single page 150 lpi 2 25 ppi 4.7 18.7 300 ppi 8.3 33.2 A4 Magazine double page 150 lpi 2 25 ppi 9.4 37.4 300 ppi 17 66.4 Figure 5. 7 Here is a rough guide to the sort of file sizes required to reproduce either a mono or CMYK file for printed use The table contains file size information for output at multiples of x1 .5 the... or Adobe RGB (see the following section on RGB edit spaces) 281 Martin Evening Adobe Photoshop CS4 for Photographers Figure 5. 11 The Color Settings dialog Top view Side view The RGB edit space and color gamut No Photoshop book for photographers would be complete without a discussion about which RGB working space to choose from the RGB Working Spaces menu in the Edit ➯ Color Settings dialog (Figure 5. 11)... undo a series of adjustment panel edits Figure 5 18 This shows th two clipping layer gure 5. 18 shows the ws w modes for adding a new adjustment layer The one on the left is for applying an adjustment to all layers below the adjustment layer The one on the right is for clipping adjustments to the layer below only 287 Martin Evening Adobe Photoshop CS4 for Photographers Auto image adjustments The Auto... resolutions, both in inches and in centimeters Output use Screen ruling x1 .5 Output resolution MB Grayscale MB CMYK x2 Output resolution MB Grayscale MB CMYK A3 Newspaper single page 85 lpi 130 ppi 3 12 .5 170 ppi 5. 5 21 .5 A3 Newspaper single page 120 lpi 180 ppi 6 24 240 ppi 10 .5 42 .5 A4 Magazine mono single page 120 lpi 180 ppi 3 NA 240 ppi 5. 3 NA A4 Magazine mono double page 120 lpi 180 ppi 6 NA 240 ppi 10.6... settings for that particular layer Figure 5. 15 This shows the Adjustment layer workflow Go to the Adjustments panel and choose an adjustment type (in this example Levels) and make the adjustment, which will then be added as a new adjustment layer You can also mouse down on the Add Adjustment layer button in the Layers panel (circled) to add a new adjustment layer 2 85 Martin Evening Adobe Photoshop CS4 for Photographers. .. accurate representation of the image levels it is advisable to force Photoshop to update the histogram view, by clicking on the Refresh button at the top of the panel 2 75 Martin Evening Adobe Photoshop CS4 for Photographers Interpreting an image A digital image is nothing more than a lot of numbers and it is how those numbers are interpreted in Photoshop that creates the image you see on the display We... the new pixel values, by referencing the surrounding pixels 271 Martin Evening Adobe Photoshop CS4 for Photographers Planning ahead Once an image has been scanned at a particular resolution and manipulated, there is no going back A digital file prepared for advertising usage may never be used to produce anything bigger than a 35 MB CMYK separation, but you never know It is safer to err on the side of caution... compressed 277 Martin Evening Adobe Photoshop CS4 for Photographers Understanding bit depth To understand what the bit depth numbers mean it is best to begin with a grayscale image where there is just luminosity A 1-bit or bitmapped image contains black or white pixels only A 2-bit image contains 4 levels (22), 3-bit 8 levels (23) and so on, up to 8-bit (28) with 256 levels of gray Therefore, a 24-bit RGB color . 21 4000 x 2 850 11.4 32 .5 43 .5 20 x 14 50 x 36 13 .5 x 9 .5 33 .5 x 24 450 0 x 3200 14.4 41 54 .5 22 .5 x 16 56 x 40 15 x 10 .5 37 .5 x 27 50 00 x 4000 20 57 76 25 x 20 62 .5 x 50 16 .5 x 13 .5 42 x 33 .5 Output. 5. 5 x 4 13 .5 x 10 2400 x 1800 4.3 12 .5 16 .5 12 x 9 30 x 22 .5 8 x 6 20 x 15 3000 x 2000 6 17 .5 23 .5 15 x 10 37 .5 x 25 10 x 6 .5 25 x 17 350 0 x 250 0 8. 75 25 33 .5 17 .5 x 12 .5 44 x 31 11 .5 x 8 .5 29. Figure 5. 2). This can be expressed clearly in the following formula: the number of pixels = physical dimension × (ppi) resolution. In other words, there Martin Evening Adobe Photoshop CS4 for Photographers 266 is