Gradient midpoint interpolation when in LAB

[garrettm30]

There are two issues with interpolated midpoint colors in gradients:

1. If you have an object filled with a gradient in an RGB document, and then you change the document color space to LAB, then the gradient’s interpolated midpoint colors change drastically, even though the endpoints (stops) are the same.

2. If, while in the LAB color space, you edit the gradient by inserting a midpoint, the new stop that gets inserted does not match the color that was previously the interpolated midpoint. In fact, the new stop appears to be calculated as if it were in RGB.

Here is a video demonstration:

I would not be surprised if one of these two issues are “as designed,” but I doubt both are, as they are contradictory: when rendering interpolation of gradients, the interpolated color values are calculated based on the document color space, but when adding stops to a gradient, the new stops inserted are taken from the interpolated values based on RGB regardless of the document color space.

I was using Designer 1.9.3 on macOS 10.14.6 in the video. This bug report is a followup on a feature request discussion about how Affinity interpolates colors.

gradient_test.afdesign

[Sean P]

Hi Garretm30,

Point 1 is to be expected - as you're using a different colour mode, the blending between colours is going to be very different using Lab to RGB. This can also be seen in Photoshop as well when switching between RGB and Lab. Interestingly Photoshop always shows the preview of it rendered in RGB! - See below.

As for Point 2 - This does look wrong - I suspect it should take the midpoint of what you see in Lab, but instead it does appear to be doing it as though it was RGB as you say. I'll get this passed on and check with development.

[garrettm30]

22 minutes ago, Sean P said:

Point 1 is to be expected

[…]

As for Point 2 - This does look wrong

That makes sense. I thought it was either one or the other that was "as designed," but I was not 100% certain which it was.