12. Optical Data Storage
Total Page:16
File Type:pdf, Size:1020Kb
12. Optical data storage 12.1 Theory of color RGB additive color model B B Blue Magenta (0.6, 0.2, 1) (0,0,1) (1,0,1) Gray shades Cyan (g,g,g) (0,1,1) White (1,1,1) (0.6, 1, 0.7) R R Black Red (0,0,0) (1,0,0) G Green (0,1,0) (1, 0.7, 0.1) Yellow (1,1,0) G The RGB color model converts the additive color mixing system into a digital system, in which any color is represented by a point in a color tridimensional space and thus by a three coordinates vector. The coordinates are composed of the respective proportions of the primary light colors (Red, Green, Blue). The RGB model is often used by software as it requires no conversion in order to display colors on a computer screen. 169 CMY subtractive color model Y Y Yellow Green (0, 0.3, 0.9) (0,0,1) (1,0,1) Gray shades Red (g,g,g) (0,1,1) Black (1,1,1) C (0.4, 0, 0.3) C White Cyan (0,0,0) (1,0,0) M Magenta (0,1,0) (0.4, 0.8, 0) Blue (1,1,0) M The CMY color model corresponds to the subtractive color mixing system. The coordinates of a point in this color space are composed of the respective proportions of the primary filter colors (Cyan, Magenta, Yellow). The CMY model is used whenever a colored document is printed. The conversion RGB and CMY models can be written as follows: For the orange point, for example: 170 HSV color model V Green Yellow 120° 60° (37°, 0.9, 1) White H 1.0 S Cyan Red 180° 0° V Blue Magenta 240° 300° Hue is given in polar coordinate. The Saturation is the radius from the V- H axis and describes the vividness of S Black the color. The Value along the V-axis 0.0 describes the brightness. Compared to the RGB and CMY, the HSV color model has the advantage that the colors correspond more closely to our subjective description. It is, therefore, easier to make a specific color selection. This color model uses three parameters: Hue, Saturation, and Value. By projecting the RGB unit cube along the diagonal from white (1,1,1) to black (0,0,0), we get the basic hexagon of the HSV pyramid. 171 LAB color model White +b* L=100 hue Green Red –a* +a* –a* +a* Black L=0 –b* LAB color model is commonly used in graphic arts and in the textile industry. Originally proposed by the "Commission Internationale de l'Eclairage" (CIE), it is also referred to CIELAB model. It uses L, a*, b* coordinates. L is the Lightness contrast, similar to the brightness Value of the HSV model, except that L values scales from 0 (black) to 100 (white). Hue and Saturation polar coordinates of the HSV model are here converted into cartesian coordinates a*, b*, which values vary from –100 to +100. The color space can then be represented by an ideal sphere. 172 Chromaticity diagram A convenient way of visually representing all L colors in a two-dimensional space is the use of y "observer" a cut view of the La*b* sphere along an observer plane plane described by L = a* = b*. The obtained x figure is called chromaticity diagram (CIE 1931). –a* This 2-D space has two axes x and y. b* –b* 0.9 0.8 spectral locus 0.7 a* color temperature 0.6 T [K] The curved line in this diagram shows where the colours of the spectrum lie 0.5 2,500 y 3,000 and is called spectral locus; the 0.4 4,000 wavelengths are indicated in nanometres 1,500 6,000 along the curve. The straight line on the 0.3 10,000 ∞ bottom connects the chromaticity co- 0.2 ordinates of extreme red and blue, 0.1 purple line called purple boundary. The area enclosed by the spectrum locus and the 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 purple boundary cover the domain of all x visible colours. 173 Practical color space and gamut The three primary colors are not available in practice as pure colors, Visible color gamut but are dirtied by a certain pro- "Pantone" gamut portion of the other primary colors. RGB color gamut The result is that it is not possible to CMY color gamut create pure white (in the additive color model) or pure black (in the subtractive model). The number of color shades that can be displayed or printed in practice with a three color system is limited to a color gamut smaller than the ideal color space seen by the human eye. Color gamut of a The use of additional primitive colors typical CRT TV screen allow in certain cases to extend the color gamut. Printing uses an additional blacK ink (CMYK quadri- chromy). Textile printing is even more demanding and often employs an additional orange dye and two more blue dyes and, thus, a total of 6 primitive colors (hexachromy). 174 Numerical color processing In theory it is possible to create every Number of bit / Number of graduations Number of shade using the RGB or CMY color primary color Red Green Blue colors models. How many colors can be re- 1 2 2 2 8 presented with these models depends 2 4 4 4 64 on the number of different gradua- 3 8 8 8 512 tions possible for each primary color. 4 16 16 16 4,096 5 32 32 32 32,768 To produce >16 Mio different color 6 64 64 64 262,144 shades, as in recent graphics cards, 7 128 128 128 2,097,152 256 graduations of each of the three 8 256 256 256 16,777,216 primary colors are needed. This means 3 x 8 bits = 24 bits are necessary for each color shade. The number of bits used to represent the color of a single pixel in an image is called the color depth. When the number of available graduations is not sufficient, as in most printers, a halftoning technique has to be used to simulate light color shades, obviously at the detriment of Examples of color halftoning with CMYK spatial resolution. (Cyan, Magenta, Yellow, blacK) separations. 175 !"#$%&'(#)%"*#%*+,-.$/-.(#$01*!203)"3 The Imaging Spectrometer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pectral imaging (/+.%/130D2)&$0+&#$'(%/-$($%&4 0.1')2"&&/&)1'3 A spectral camera is a combination of scanning4-5%&)4$-".# mirror !!! an imaging spectrograph and an area monochrome camera. It works as a line !"#$%&#"'( )*+%,%'- 6,1("'( scan camera, and provides full, !"#$%&'()* dispersing element detectors4$-".# !%-%.-/&# array +&,"-#./light contiguous spectral information for #()0-12 53$/*+%"3&1"+#(+*&)'&%$/&6+,"3 (/,).0%")- /./*/-%,& )'& +-& "*+#"-#& ,7/32 each pixel. Spectral resolution can range 3/..4 %()*/%/(4& & 5)*/& ,/-,)(,& 0,/ *0.%"7./&1/%/3%)(&+((+8,&%)&*/+2 from ca. 2 nm to 20 nm. Spectral ,0(/& $0-1(/1,& )'& -+(()9 imaging typically allows for recording 9+:/./-#%$&;!!! <&6+-1,4 !"#$%& pictures with the equivalent of 20 to Schematic diagram of an imaging spectro- 200 primary colors over the visible meter. Some sensors use multiple detector domain and can even extend to the near arrays to measure hundreds of narrow UV and mid-IR (hyperspectral imaging). wavelength bands. Since spectra obtained for an area on a picture can be matched with electronic spectra of chemical compounds (spectral fingerprints), spectral imaging allows for remote chemical analysis in a number of applications, such as for instance astro-nomy, surveillance, mining and geology, ecology, biology, and forensic science. RGB manitol glucose starch 176 12.2 High resolution spectroscopy Homogeneous vs. inhomogeneous line width –1 500 cm (12.5 nm) The spectrum of a typical organic molecule in a host matrix at T ≤ 4 K is 5·10–4 cm–1 (1.25·10–5 nm) schematized here on the left.