use Canvas `.clip()` instead of compositing `globalCompositeOperation`?

[joedf]

Currently using globalCompositeOperation, maybe better (performance) and easier to use context2d.clip()?

more info here: https://developer.mozilla.org/en-US/docs/Web/API/Canvas_API/Tutorial/Compositing

[joedf]

Mainly, it would be use it in Utils.ComputeProbeValue_gs() and clip the canvas image instead of using ctx.globalCompositeOperation = 'destination-in' with a drawn ctx.ellipse(). However, we would lose the potential of using a gaussian intensity mask. That said, a non-uniform intensity distribution is not currently supported, see #26.

[joedf]

I note that I am still using getImageData for a rectangular area and not an ellipse. So the data may actually be inaccurate... Could be resolved / mitigated by using a mask color...
See get_avg_pixel_gs() usages such as ComputeProbeValue_gs().

get_avg_pixel_gs() only considers the red component, so perhaps I can simply fill the entire area with pure blue or green, assuming all grays will have the same value across all components.

Never mind, it is fine because I am already accounting for this by checking for "blanks" where pixels have 0 in red and 0 in alpha! Essentially, we exclude 0-0-0-0 rgba pixels from the calculation.