Calibration - Speed vs. Accuracy
How fast can LightSpace CMS calibration be performed with accurate results?
LightSpace CMS includes two main profiling functions - full 3D Cube based profiling, and the far faster Quick Profiling method. Both methods can produce excellent results, but with a great differences in profiling speed.
So why the different methods?
Display Calibration Requirements
To accurately calibrate any display you first must profile (measure) the underlying properties of the display. It is this profiling that will define the final calibration accuracy, and the simple rule of thumb is that the more colour points profiles the greater the final calibration accuracy will be.
This is a simple fact that cannot be argued with - except by introducing the concept of 'acceptable' accuracy.
Calibration accuracy is basically the concept of what level of calibration is 'good enough' for any given display?
For displays that are intrinsically 'linear' in their response to signal input change (which means that for any given change in input signal the change in colour on the display is an equal level of change) it is possible to perform 'acceptably accurate' calibration with a small set of profile measurements.
On such 'linear responsive' displays the difference between large cube based profile calibration, and smaller Quick Profile based results, can be very small indeed.
3D Cube vs. Quick Profiling
The following shows the profiling points for a large 17^3 3D Cube profile set vs. a Quick Profile, both performed with an i1 Display Pro set to 0.75 Sec integration time, on a standard LCD display. The specific LCD display was chosen as it is an average example, with a less than perfectly linear response to input signal changes.
The i1 Display Pro probe was chosen as it is the cheapest probe of acceptable quality for accurate calibration, so is the most common probe presently in use.
17^3 Cube Profile
4913 Patch total, profiled in 1 Hour 20 Mins 4 Secs
81 patch total, profiled in 2 Mins 15 Secs
17^3 3D Cube based calibration gamut results
Quick Profile based calibration gamut results
17^3 3D Cube based calibration gamma results
Quick Profile based calibration gamma results
17^3 3D Cube based calibration Delta-E results
Quick Profile based calibration Delta-E results
As can be seen, the calibration results are near identical. Certainly within more than 'acceptable' tolerances, and with a calibration duration difference of 2 Min 15 Secs vs. 1 Hour 20 Mins 4 Secs.
In reality the small variations between the results from the two profiling methods can easily be put down to display and probe drift during the profiling and calibration procedure.
When to use Quick Profiling
Quick Profiling works best on displays with good a good linear response, and a good initial RGB Separation (a lack of Cross-Coupling errors within the display), combined with good RGB Balance. The following graphs show examples of 'good' vs. 'bad' RGB Separation.
Reasonable RGB Separation
Poor RGB Separation
If a display shows poor RGB Separation, has poor RGB Balance, and/or has any irregular non-linear response to the input signal, a full 3D Cube based profile will be required for an accurate calibration as the display will require full volumetric profiling to enable the inherent display errors to be 'undone'.
Note: Any display that is 'un-calibrated', set to its native/raw mode, should always show good RGB Separation. Any indication of poor RGB Separation indicates an underlying issue with the display's design, or a still active, and likely poor, internal colour management system. Post calibration it is possible for RGB Separation to show errors, as the calibration procedure may have to introduce cross-coupling to enable the best possible volumetric calibration - this is especially true if the display has a lower gamut that the target colour space.
The following left-hand graph shows an example of a display with a very non-linear response to input signal changes. Using a Primary and Secondary Quick Profile it can be seen the for red, Yellow, Cyan and Magenta the hue of the measured patches tracks along the edge of the gamut triangle, even though the hue for each input colour patches is identical.
The right-hand graph shows a response that is a linear desaturation of of the output colour as luma decreases. The measured patches maintain the same relative hue, and track towards the native backlight colour temperature of the display. Such a response will not prevent the use of a Quick Profile for LUT generation.