/ic/ - Artwork/Critique

>>6652844
It seems like some people find rotating an ellipse easier to understand than rotating a rectangle, and then they use the ellipse to find the rectangle. The ellipse is tangent to the top and sides at their midpoints.

To me, this is so backwards that it sounds like insanity, but I guess it's just a different approach. Personally I do the literal opposite.

>>6281608
If accuracy is what you really are after, the circles that define the inner curves of each arch should be inscribed in squares.

First, over those squares, you ought to add lines in order to indicate the vertical, diagonal, and horizontal axes, four in total (pic related).

Also, to guarantee consistency, the points at which the diagonal axes intersect with the circumference should follow their own vanishing lines (pic related).

This method also works with the bases of cylinders.

>>6281608
If accuracy is what you really are after, the circles that define the inner curves of each arch should be inscribed in squares.

Then, over those squares, you ought to add lines in order to indicate the vertical, diagonal, and horizontal axes, four in total (pic related).

Also, to guarantee consistency, the points at which the diagonal axes intersect with the circumference should follow their own vanishing lines (pic related).

This method also works with the bases of cylinders.