Add CIELAB and RAL keys to support color matching filaments

[bani6809]

This is borne from my frustration trying to match filament colors for a project and going through 20 spools from 15 different manufacturers to find the closest match. It took weeks.

RGB is fine for UI simple representation, but fundamentally unsuitable for color matching. First of all, it does not define a color space. Second of all, there is horrible quantization with only 8-bit values, as anyone familiar with the blocky gradients in blue skies on DVD videos has seen. Newer video standards use a minimum of 10 bits to kill this.

CIELAB is the industry standard colorspace for color matching. The L A B values must be floating point. There also needs to be a key for the reference white point, as the color will appear different based on the illuminant. This should be a value in kelvin. It is critical to know what white point is used, a color under 6500 can appear significantly different from 3000, 5000 or 7500.

RAL classic color codes should also be supported. Several vendors already offer filaments matched to RAL colors. RAL codes range from 1000 to 9999.

Budget shops can just buy RAL fan decks and publish RAL classic tag values. This is way better than the nothing that we currently get. RAL does not require licensing to use or publish color values, unlike pantone. Therefore, we should avoid adding pantone tags.

[CZDanol]

Hello,
we actually had the RAL in the fields initially, but decided to remove it. We've concluded that the data quality would likely not be better than just using the RGB fields, at which point it would just be redundant data.

The first problem was - how would you even define the RAL field? There are basically three options:

1. You only fill in the field if the RAL color exactly matches the filament color
2. You fill the RAL field only if the filament was specifically intended by the manufacturer to match the RAL swatch (in that case, you cannot have for example both RAL and PANTONE fields filled in at the same time, as you can't really design a filament to exactly match two different colors).
3. You fill the RAL field with the swatch that matches the color the closest

For the first two options, it seems that the manufacturers are not doing a particularly good job. For example here:

At that point, it's imo better to not publish the "exact" colors rather than misleading people (and adding administration and data storage costs).

And as for the third option - matching the closest - at that point, you can do it algorithmically from other data, like CIELAB or RGB.

Now this is just our thinking and why we decided to exclude the RAL in the end. Feel free to share your opinions and arguments, it might have been flawed.

The CIELAB proposal is a different thing, and I will try to think about that one a bit more.