Maxwell's House 5

Maxwell's House Episode 5

Maxwell's House 5 - Colour

Guests

There were no guests on this episode.

Topics

Maxwell's House News

Red Camera

Ray is skeptical that Epic and Scarlet are going to be DSLR killers due to price point and integration issues. These cameras would however be ideal in applications such as mounting a gyro stabilized camera in the belly of a Learjet and flying alongside a Boeing 777 for a promotion for Air Canada. Another application example Leo gives is a camera operator in Tasmania who does a lot of aerial and boating B-roll for the tourist board.

Ray mentions that the red camera is modular - the viewfinder can be placed in the cockpit and the camera can be operated by remote control. Ray does not foresee the typical still commercial photographer needing to purchase this camera right away.

Although the Red Camera will have mounts for both the Nikon and Canon lenses, Ray doubts it will couple to the auto-aperture and auto-focus of these cameras. Red does make a set of electronic lenses that are auto focus but not auto exposure.

Ray sees a move towards integrating still and video as evidenced by the Canon and Nikon line of cameras.

Adobe CS4 Design Premium

• Ray is very taken by how much the stitcher in Photoshop has been improved.
• Also, an incredible trick now available is auto-blend now allows you to stack a set of macro shots with different focal planes resulting in incredible depth of field in your macro work.

Dreamweaver

Ray is studying dreamweaver and thanks to Leo's teasing he is going to update his static website especially since he has a high page rank rating in Google. Ray promises the website will come back to life within a few weeks.

eSATA

Ray is quite surprised at the difference an eSATA plugin card for his MacBook Pro makes that he picked up from London Drugs. It plugs to the ExpressCard/34 slot to connect to an external Iomega Terabyte drive for backup. He typically gets write speeds of 14 MBytes per second and Read speeds 28 MBytes per second using FireWire 800. eSATA gets him 60 Mbytes per second both ways, more than four times the write rate.

Leo recommends BlacX eSATA adapters and docking stations similar to the desktop solutions he has adopted at the Twit Cottage. The only downside is Leo has not found the CardBus eSATA adapter to be hot swappable as advertised.

Maxwell's House Mail

Fluorescence

Relating to the imaginary magenta issue, what is fluorescent orange and fluorescent green? Can it be reproduced at a computer monitor? is it part of the electromagnetic spectrum? How are these colours made and what is the eye really seeing when we look at fluorescent paint colours? Ray took photos of fluorescent color patches and demonstrated that we can see fluorescent colors on a computer screen. A reflection Spectrophotometer puts out incident light and measures how much of that incident light gets reflected. A slide showing a scale of amount of light reflected with a max of 1.0 is demonstrated. Results of reflection of a white piece of paper (most of the light is reflected) are contrasted with results of a normal red pigment (absorbing energy in the green and blue spectrum, reflecting red spectrum). Spectrophotometer results of a red fluorescent sticker show it absorbs light in the blue and green spectrum and not only reflects light in the red spectrum but emits more light than is falling on it. It outputs more light than what fell on it. This seems to indicate a violation of the law of conservation of energy. This is obviously not the case. The explanation lies in the fact that the results of the reflection spectrophotometer indicate spectrum between 400 to 700 nanometers (*1 angstrom = 0.1 nanometers) which only looks at the visible part of the spectrum. Shorter wavelengths (ultraviolet) and longer wavelengths (infrared) exist. When light is shone on the red fluorescent sticker, ultraviolet energy which the eye can't see comes from a source (light bulb, fluorescent lamp, the sun, etc) hits the fluorescent dye pigments and cause the electrons and atoms to go to a higher energy level. The high energy levels falls down resulting the atoms of the dye to emit light in the visible spectrum. We get more red out of this than the red that was falling on it. This is fluorescence.

Phosphorescence

Whereas fluorescence is instantaneous - ultraviolet in gets visible light out - and as soon as the light source is removed fluorescence goes away, in phosphorescence the materials store the energy. When irradiated with ultraviolet light, electrons are excited and the material glows even after the light source is taken away. some fish are phosphorescent. The ultraviolet energy drives the electrons to a higher energy state in these particular materials. They then slowly fall back to lower energy states. When electrons go from high to low energy they emit photons, and depending upon what energy levels they are at it emits different wavelengths of light. Energy is being shifted from an invisible source (ultraviolet) down into the visible spectrum in fluorescence and phosphorescence.

Maxwell's House 5 Topics

CIELAB Colour Space

Most people are familiar with RGB (Red Green Blue). Another, YUV, is used in video and television. LAB is in photoshop. The color matching functions were discovered in the 1930s. A slide is shown of the model of human perception showing the response of the blue, green and red cones in the human eye. This was the first attempt to build a mathematical model or function that describes how the eye responds to electromagnetic radiation from 400 to 700 nanometers. If you illuminate any kind of dye and you put it through these matching functions you come out with three values X,Y&Z. These numbers are not linear with regard to perception. Hunter developed an equation where the X,Y&Z integrals were taken under these functions to create L,A,B.

Hunter's Lab Model Slide is shown depicting a vertical L axis (luminance) from zero to -l, absolute Black, and from zero to +l, absolute White. an A axis has from zero to -a, Green, and zero to +a, Magenta. a B axis from zero to -b, Blue and from zero to +100, Yellow.

To get red, +a and +b To get cyan, -a and -b

The bigger the value in a*,b* the more saturated the color is. The hue of the color is also represented. This is a model of human perception.

In photoshop, any time you change from one color space to another (from sRGB to Adobe RGB1998, or to CMYK SWOP coded for magazines, or sheetfed coated for set presses) it first translates the colour RGB values into LAB and from LAB translated back to the destination colour space required. This is done because we have to decide how the human being is going to respond to it.

Colour Profiles

The International Colour Consortium came up with a standard ICC profile. Consider a [1] calorimeter or a spectrophotometer that can measure the light that is being emitted from a monitor and it can convert it into the L,A,B values. In other words, it can measure how a human being would respond to that color and what color they would see using the LAB matching functions. To build a profile, we put out an R,G&B value to the monitor which is the input to the lookup table (slide shown). We measure the LAB value we get out and put in this lookup table these LAB values. We can translate RGB values for a specific monitor. (All monitors vary - with age, brightness setting, type of monitor, phosphorous used, LCD filters used - over time). To do serious colour work you have to profile your monitor from time to time. Each time you will put the calorimeter/spectrophotometer on your screen, the software will put out various RGB values and measure the LAB values that it gets back.

There are no correct colours, RGB does not define colour; it just stimulates your monitor to put out some colour and it will vary from monitor to monitor, from device to device.

There is no official RGB value for colours anywhere; you have to have an ICC profile to define what the colours mean. This is why when we exchange files we put tags in them like Adobe RGB(1998) or sRGB.

LAB is the Rosetta stone that we translate everything to. You translate everything to a common language. It's the Esperanto. We take this intermediate step because the human eye does not respond in a linear manner to spectral stimuli. We have to go through this model of human perception in order to predict what color the human is going to see when we put in certain RGB values.

An ICC profile for monitors, etc (scanners and cameras are the exception) always has two tables. Once you stimulate the monitor with RGB values and you get out LAB values, you build an inverse table put where if you put in LAB values you get out RGB values. They are called the AtoB table and the BtoA table by convention in the ICC.

The Standard Observer

The LAB values define the response of a standard observer to that color. The standard observer was defined in 1931 using seventeen all male college students. It is almost certain that one of them was red-green colour blind. This test has been repeated many times over involving men and women of all ages and it turns out they got lucky back in 1931. The standard observer is pretty much in the middle of the field purely by luck not by skill.

LAB can define Magenta. It can define colours that are outside the spectrum. (it would have to as it is reflecting what the human eye sees)

Pantone

For years Pantone company printed swatch books. The first swatch books had a Pantone and a number. To see a colour correctly you really needed to be in a 5000 degree Kelvin viewing booth in a printing shop. If you chose a colour from the swatch book, you could order an ink and the ink would come back being very close to the colour of the swatch. Later, they put out another swatch book that had CMYK - process colours for printing presses. What was wrong was that they never specified what colour space they were talking about - was it SWOP (Specifications Web Offset Publications) used for magazines in the US? was it US Sheetfed Coded (annual reports, commercial printing)? packaging standards (e.g. Kellogg's Cornflakes box, FOGRA, Toyo)?

Ray relates he once called Pantone and talked to one of their 'colour scientists' and asked if they made a separate swatch book for Europe and was astonished they said they used the same one as the U.S.

Ray recommends the X-Rite Colour Munki spectrophotometer if you are serious about colour and want to be ahead of the game. He also states that if you are a serious amateur photographer, the very first thing you should get is a calorimeter so that you can calibrate your monitor.

The Innards of Photoshop

Slide demonstration displays how photoshop might receive a file with a note indicating that the data is in Adobe 98 RGB. The RGB values could be read from the file. However, in the file is an ICC profile that defines what Adobe 98 RGB is in LAB. The RGB Values go through the profile and output in LAB. Inside photoshop is now known what colours were intended to be sent. If RGB values were output from the file and sent to the monitor without knowing how the monitor is going to respond who knows what colour you will see. If the profile that came with the monitor was used it would be a 'best guess'. Next, the LAB goes to Custom Monitor Profile and is fed into the B to A table and outputs RGB values that are specific to your monitor to make it show the LAB values correctly. This means that what person A sees in the source monitor is the same thing person B sees in the destination monitor provided the monitors are calibrated properly.

In order to print the image and match what is seen on the monitor, The same LAB values comes into the printer profile. It outputs special RGB values just for the particular printer. The printer driver converts it into CMYK and prints the picture. Thus the picture on the printer matches what is seen on the monitor.

• Printer profile made using a spectrophotometer. You cannot use a Colorimeter on printers. You have to measure a whole bunch of colour swatches or use the standard printer profile. Unlike monitors, the profiles that come out with the current batch of high quality printers are quite good. They are not as good as a custom profile but are very good for the majority of serious amateur photography.

It is very seldom to work in LAB in photoshop. one trick is to sharpen the L channel without sharpening the a and b channel often times producing a better sharpening result that sharpening RGB. The reason is the eye is much more sensitive to luminance difference that it is color difference. You can smear the colours around a bit and they can be a little blurry but as long as the luminance information defining the sharp edges of the subject is there, you will be fine.

Under the hood Photoshop is always working in LAB. Any time it translates from one colour space to another or outputs colour to a monitor or to a printer, it needs an ICC profile to know how to translate the colour that you want for that particular device. RGB and CMYK are called device dependent colour spaces. LAB is device independent. (but Ray always laughingly says no it's not device independent because the device is the human eye - it is the model for the standard observer's human perception).

Adobe 98 & sRGB are standard colour spaces. There are no devices out there that produce colour exactly as described in these colour spaces. These colour spaces are good editing colour spaces; what makes them good for editing colour is that they are perceptually linear i.e. if you go up and make up a 5 unit change in the highlights (e.g. between 250 and 255 in a zero to 255 in RGB which is 8 bit), and you go down and make a 5 unit change in the shadows, perceptually you will see the same colour shift in those colour spaces. This is whether you are in the highlights, the shadows or the mid-tones. They are perceptually linear. The devices we usually talk to such as monitors, printers, etc are not anywhere near perceptually linear. So we edit and work in these. other examples of editing colour spaces are ektacolor, wide gamut space. There are actually synthetic colour spaces, idealized colour spaces that are designed specifically for Adobe,etc to work in when you are editing a colour; there is no device that exactly emulates those colour spaces. When your camera is saying that it is converting to those spaces it is an approximation. The RAW data coming out of the chip in your camera does not look anything like sRGB.

Books

If you are a graphics designer, a commercial photographer or anyone that works in this area you do need today to understand something basic about Colour Science. Ray recommends any books by Bruce Fraser and Jeff Schewe

• Real World Color Management by Bruce Fraser, Chris Murphy & Fred Bunting
• Real World Adobe Photoshop CS3 By David Blatner, Conrad Chavez, and Bruce Frasier

Notable Comments

• Leo mentions that 564Kbits/sec downstream is needed consistently for the Stickam live streams; 500Kbits for video and 64Kbits for audio
• Ray suggests a new format for the show: news to Maxwell's House, letters to Maxwell's House, Main Tech Stuff
• Leo mentions that the designer of the Adobe ConnectNow screen sharing technology used on the show, Giacomo Guilizzoni, was featured on Net@night episode 75
• GoToMeeting is proposed as an alternative to Adobe ConnectNow due to live connection problems during the show.
• Ray's printer Epson 4000
• Ray is an Electronics Engineer, but for the past 7 years has worked in Colour Science for Creo with colour scientists at Kodak, colour scientists at Imation, Dupont, Fuji. His company built special laser imaging engines for very high quality proofing. The printing industry is very demanding when it comes to colour matching. This is especially true for car brochures - they are the toughest customers.
• What used to be high end technical topics now have to be understood by 'mere mortals' who use digital cameras, printers and computers.

Production Notes

• Recorded Date: November 20, 2008
• Release Date: November 20, 2008
• Duration: 1:01:50
• Log line:
• Edited by:
• Notes: NA