Currently we modelize a circular spot. When the machine has 5 axes the post-processor may tilt the beam axis to compensate a conic beam (especially useful in oxy-fuel cutting and water jet cutting afaik).
Just the knowledge that the spot is oval is already useful: you could compensate cutting speed to correct for that and get a more even burn.
A simple trick to check the shape of the spot is to set the beam to low power and to purposefully increase the Z as far as it will go and project the beam on a piece of black anodized aluminum. You'll very clearly see the alignment of the individual diodes and as you lower the Z you see them converge on what eventually will become the focal point. This gives you very useful hints about the shape of the cone, on my machine the cone is definitely oval (if not outright linear!) in cross section, far longer in X than in Y. This results in an ~ 30% penalty in cutting speed depending on the direction of travel. To ensure full penetration I have to set the machine 30% slower, whereas if the software compensated I'd be able to run the same speed with much more consistent kerf width as a result and less time wasted.
Unfortunately, globally because LightBurn, the software that I use doesn't allow to compensate for this. But I could run my own G-Codes or maybe patch GRBL to demonstrate the effect.
