1. How do I calibrate more than one display?
SuperCal is not yet multiple-monitor aware, but it can be used in a multiple-monitor configuration. SuperCal currently operates on the non-mirrored, main display which is the display containing the menu bar. In order to calibrate a secondary display, use the Monitors control panel (under Mac OS 9) or Display panel in the System Preferences (under Mac OS X) to move the menu bar to the secondary display to make it the main display, then run SuperCal again. You may need to restart after changing the main display because not all systems update their configuration properly after re-arranging displays. For example, there appears to be a bug in Mac OS X 10.0 through 10.1 on a PowerBook using the external video that does not update the video configuration properly after a display arrangement change. In this case, a restart is required for the video hardware to configure itself properly.
In addition, you cannot calibrate with video mirroring turned on. However, you can calibrate and generate a profile for each display individually, then turn mirroring on and assign the each profile to its respective display using the Monitors control panel (under Mac OS 9) or Display panel in the System Preferences (under Mac OS X).
2. I don't see raster lines on my display when adjusting the brightness and/or contrast.
Older, lower resolution CRT displays used to have visible raster lines (or scan lines as they were also called) formed by the rows of illuminated pixels and the thin dark rows between them where the electron beam didn't illuminate the phosphor on the picture tube. As CRT displays have increased in resolution over time, the rows of pixels have gotten progressively smaller and denser, and the thin dark rows between them have nearly vanished because of the precise focus of the electron beams in today's CRTs. The imageable area now just appears to be a solid colored area. In addition, LCDs are a very different technology than CRTs and do not have raster lines.
In both of these cases, you simply need to judge the brightness difference between the solid imageable area and the unused border around the outside of the imageable area. On CRTs this is fairly easy because the border is usually large, but on LCDs, the border is typically very small which makes it more difficult.
3. I can't adjust the brightness high enough to see the raster area.
Some displays will not display raster lines or a non-black raster area with the contrast turned all of the way down and the brightness all of the way up. In this case, you will need to adjust both controls simultaneously using a reference image that has both black and white in it so that you can judge both black level and white level at the same time. The best image to use is one with a wide border of black around its perimeter and a sufficiently large patch of white in the middle. Currently, SuperCal does not have such a helper image (like the "Make White" and "Make Black" buttons), but you can use a desktop pattern or Adobe Photoshop to make such an image for adjustment. Remember, you don't need to make display adjustments while running SuperCal. SuperCal just provides the instructions and helper screens as a convenience.
Using the black and white image, you should set the brightness at maximum and begin adjusting the contrast upward until either white is at a comfortable level or black begins to become distinguishable from the unused surround. If you reach a point where white is bright enough and you still can't distinguish between the black in the image and the black in the over-scan area, then you're done. In this scenario, your display simply does not have enough range in the brightness control to overcome the ambient light level.
But, it is most likely that the black will become distinguishable first, so when it does, turn the brightness back down until the two become indistinguishable. This is the little game that you now have to play. Repeatedly, adjust contrast up and brightness down until white is a comfortable level and the black area and surrounding over-scan area are indistinguishable. This is the point where the black level is set properly for your ambient lighting and the contrast is set properly for a nice, comfortable viewing brightness.
4. SuperCal doesn't seem to be any better than Adobe Gamma. Why should I pay for this thing?
It all depends on what kind of display you have. If you have a really good CRT, you may see little difference between the results from SuperCal and the results from other visual calibrators. This is because other visual calibrators are designed to calibrate displays with response curves that approximate a power curve. The better your CRT, the closer its response will be to a smooth power curve and the better the calibration will be. Other displays like LCDs, projectors and less-than-perfect CRTs may have response curves with radically different shapes than a power curve, and as a result, the cannot be measured and calibrated with ordinary visual calibrators. This is where SuperCal far surpasses these other calibrators.
5. I see a color cast in certain gray values when using the created profile. What's causing this?
A color cast can occur in some gray values when a color channel has an irregularity in it that is not identified and/or measured properly during the measurement process. There are two ways that this can be introduced into the profile. First, it is possible to be too aggressive and measure too many points, possibly causing a wrinkle in a channel's curve which will cause a color cast in the final profile in the area of the wrinkle. Second, it is possible to measure too few points and miss an area that requires a unique adjustment to account for an irregularity in the display's response.
The first problem is easy to correct. When measuring, don't feel the need to measure every point that is available. If you have already measured a few well-spaced points and have the rough shape of the response curve already determined, don't feel like you need to measure every remaining point where the measurement pattern varies only slightly. Use good judgement and only measure points at well spaced intervals that truly require measurement. In addition, give your eyes time to adjust to the measurement patterns as you scan across the remaining points. The difference in contrast between the intensities used to make the patterns can require time for your eyes to adjust, so they will look different after a few seconds of visual adjustment than they do immediately upon seeing them. Plus, some points will use pattern intensities that are very similar and some points will use intensities that are very different, so they can be deceptively different at first sight (the small arrows below the graph illustrate which intensities are being used to make the pattern at the selected measurement point). As a result, stop and give your eyes a moment to adjust, then measure.
The second problem is more difficult to identify and account for. Sometimes, an irregularity in a display's response will lie at a point where you do not yet have an available measurement point. As a result, you may never see the problem until you make a profile and notice the color cast in the resulting grays. If this happens, try and identify the approximate gray value where you notice the color cast, then focus on this area in a subsequent calibration. The graph on the measurement step can help you identify the area since its horizontal axis corresponds to the numerical values you see and use in image editors - black or "0" being to the left and white or "255" being to the right. Remember, you will get an available measurement point at the midpoint between two already measured points, so if you don't have a point available where you need it, you may need to make a measurement above or below that point to make one available. Again, be careful not to make too many measurements, or you could introduce the first problem discussed above - more chances for wrinkles and errors that way.
6. I see posterization in gradients when using the created profile? What's causing this?
Part of the calibration process consists of generating a corrected look-up table for the video card. A look up table basically maps an input value (like the RGB pixel color values that you see in a Photoshop color inspector) to an output value that is actually sent to the display to be viewed by your eyes. This look-up table has a finite number of values, and when you generate a profile, you effectively reduce the number of usable values when some of the values are re-mapped to correct the response of your display. In a perfect world, the display would have hardware adjustments to adjust it to your liking and the look-up table in the video card would be a linear table (slot 0 = 0, slot 1 = 1,... slot 255 = 255), where each RGB color channel has full use of all 256 values, resulting in perfect 24bit color output (256*256*256 = 16.7 million colors). However, when a corrected table is generated, some neighboring input values might contain the same output values, which reduces the overall total number of output value combinations. As a result, compression occurs in the table which causes posterization on-screen. Its the exact same visual effect as taking a 24-bit color image and reducing it to 16-bit color. Some colors will no longer be available and will be changed to the closest color that is available. But not to worry, display calibration doesn't change your files. It only changes the way that they are presented on-screen.
Unfortunately, this is an unavoidable side effect of calibration when dealing with video cards with 8-bit digital-to-analog converters (if you're interested in learning more, read up on video card design). This is why it is very important to adjust your display properly using the hardware settings on the display before calibrating the display. By doing so, you will maximize the dynamic range of your display and minimize the destructive effects that result from a calibrated look-up table.
7. Why can't I edit and tweak the curves?
In a future build, you will be able to edit the profile after the display has been measured. This will allow you to correct any mistakes that were made during the measurement process or tweak them to suit your subjective preferences. Currently, your only options are to use "Undo" during the measurement process, or to re-do the full response measurement. It may seem tedious, but once you've done one full measurement, you should have the hang of it and be able to do subsequent error-free measurements in only a few minutes.
Because the patterns generated for a particular point during the measurement process are dependent on measurements of previous points, it is simply not possible to go back and tweak a particular measurement point because its effect will then ripple through any subsequent points (kind of like the time paradox of going back and influencing your own birth). For this reason, editing will be limited to the gamma table that is generated from the measurements, not editing of the measurements themselves.
8. When I move my head, the screen really changes. How am I supposed to do a good job when this happens?
You really need to assume your normal viewing position and keep your head stationary throughout the measurement process. If you are seeing changes when you move your head up and down or side to side, you most likely have a PowerBook with an LCD display. PowerBook LCDs have very limited viewing angles, and no software or hardware calibrator can overcome this limitation. You will get the best results if you perform a very careful calibration from your typical viewing position. Try making a simple chin rest with a stack of books to make it easier.
9. Why does my display flicker sometimes when I start SuperCal?
This is an unfortunate visual blemish that occurs when SuperCal checks the video hardware to see what capabilities it has. SuperCal works by sending a table of numbers to the video card to adjust what colors the video card sends to the display. The odd display flicker happens when SuperCal sends a pattern of numbers to the card, then asks for them back to see if the card could remember them all. In some cases, when a table of numbers is sent to the video card, ColorSync intercepts the numbers and turns them into an equation (for some odd reason). If this equation is returned when SuperCal asks for the numbers, it doesn't give SuperCal any information it can use. So, SuperCal sends a pattern of numbers that ColorSync cannot turn into an equation. The visual blemish is simply a very quick change of the red, blue and green colors on your display.
10. Why do the patterns look funny in the measurement screens on my LCD?
This should only happen on LCD displays that are connected to the computer through an analog video connection. What is happening is that the analog video signal is being converted back to a digital signal by the display, but it is not being converted properly and produces visual artifacts on the display where "high frequency" images are being displayed, like single-pixel checkerboard patterns. You can usually correct the problem by adjusting what is typically called the "phase" setting in the hardware controls of your LCD display. For adjustment purposes, you can use the patterns in the measurement step or you can produce a single-pixel black and white checkerboard pattern in an image editing program like Photoshop and use it instead. Simply display the pattern and bring up the on-screen hardware controls and adjust the "phase" setting until the image is clear and stable.
11. The Black Level and Response Measurements don't work. I see a gray pattern and nothing changes when I move the sliders.
This was an unfortunate problem that occurred on certain video cards under Mac OS 9, but it has been fixed as of version 1.1.2 (we hope). We discovered that under Mac OS 9.x, some systems incorrectly reported that the video hardware had a 2 byte-per-entry gamma table instead of a 1 byte-per-entry gamma table. To our knowledge, this problem has not occurred on any systems under Mac OS X. ColorSync display profiles are identical between OS 9 and OS X, so if something like this should ever occur again, you can boot into the second system (which hopefully won't exhibit the same problem), build a profile, then copy it back over to the first system. Again, this should be fixed as of version 1.1.2.
12. I see visual artifacts around the buttons and sliders in the measurement screens.
Native Mac OS controls are used on measurement screens and as a result, annoying color fringes can occur around their edges during measurements. This is purely a cosmetic issue that results from altering specific gamma table values during measurement. We will probably create custom controls in a future build to eliminate the visual flaw.
13. How do I undo the changes that SuperCal made to my monitor settings and set it back to default?
Just open the Displays preferences pane under OS X (or Monitors control panel under OS 9) and select the Color tab, then choose a different display profile. SuperCal doesn't actually make any permanent changes to any existing files or settings. It simply creates a new profile in your ColorSync profiles folder, then tells the system to use this profile for the particular display. The effect is exactly the same as you copying a profile to the ColorSync profiles folder, then manually selecting the new profile that shows up in the list in the Color tab. In fact, selecting profiles from the list in the Color tab is how you choose profiles for different gamma settings and switch between them to preview artwork on your display under different gamma conditions.
14. Do I need to make any changes to any preference pane or control panel settings before using SuperCal?
No, you don't need to make any changes to the monitor profile settings in the Display preferences pane on Mac OS X or the Monitor control panel on Mac OS 9. Regardless of which display profile is selected, SuperCal temporarily adjusts the video card during measurements to the proper state for measuring. You usually won't see the change happen because SuperCal uses a full-screen, black background during measurements and this hides the effect of the change.
However, If your monitor only has software controls for brightness and contrast like on iMacs and some Apple LCDs, you will of course need to adjust those settings when you get to the control adjustment step.
15. I use my display under different lighting conditions, with different brightness settings and with different color temperature settings. Is a single profile valid for all of these conditions?
The answer is simply no. Each of the mentioned conditions affects the response of the display and what you see from it. If you want accurate display calibration, you should create a new profile under and for each viewing condition. While this is not practical in the everyday world, it is doable in a graphics environment where you may proof images under different conditions as part of your daily routine. If you work with a CRT and use 6500 degree white temp settings for proofing images for print, but switch to 9300 degree for web design, simply build a profile for each setting and switch between them as you need to. For color critical work, you should probably be working in a controlled environment, so brightness and contrast adjustments should not be a regular ordeal. If they are, then you should probably address your ambient working conditions as best you can because it is not possible to solve the problem of varying conditions with a single profile.
16. Why does the default Apple calibrator ask me to specify the color temperature but SuperCal doesn't?
Unless SuperCal knows the current color temperature of the display, it is not possible to accurately adjust the color temperature to another known value. In the case of Apple LCDs and other displays that don't have a hardware color temperature adjustment, it is possible to do a reasonably accurate adjustment because the display will always be in the known, default state to begin with. Apple makes use of this fact and provides the control in the default calibrator to adjust the profile to a color temperature of the user's choice.
However, if a display has hardware color temperature controls, the calibrator should not be used to adjust the color temperature. The calibrator should be left at the "native" setting and the temperature should be set with the display's hardware controls. The reason is that the calibrator has no way of knowing what setting the display is currently at. It assumes that it is at the native settings, but if it isn't, the control in the calibrator will only make matters much worse and put the display into an unknown state. This is the main reason that SuperCal does not currently have a color temperature setting - that there are many displays other than Apple's and we cannot know for sure what state these displays are in. In addition, using the software control has a degrading effect on the color depth of the display, distorting the look-up table pushed into the video card, while using the hardware control will have no such effect.
Another factor is that the ambient light temperature affects what you see from the display because of a principle called chromatic adaptation. Your eyes constantly adjust to the temperature of the surrounding light, which will make relatively small sources of light (like the display screen) look a different color than they really are. This is why computer displays tend to look very different in an office under fluorescent lighting than they do in a home under incandescent lighting. The longer you focus on a particular area, though, the more your eyes adapt to the color of that area. But the key is that your eyes are constantly adjusting, and unless the ambient light is at the same color temperature as the display is set to, you will really not see the color temperature that you believe the display is set to.
The end result is that SuperCal will likely also provide a specific color temperature adjustment in a future release once we craft a suitable interface for it. For now, the subjective white balance adjustment allows the user sufficient adjustment to accomodate a variety of ambient lighting conditions. The only downside is that it results in a state that doesn't have a descriptive color temperature number attached to it.