On Tue, Mar 22, 2022 at 5:50 AM Martin Kreuzer <[email protected]> wrote: > Thanks ... And I did some tests this morning: > > Plotted the *bisector (3rd, 1st quadrant) and *sin using > > pd 'xrange _1 1' > pd 'yrange _1.1 1.1' > and > pd 'aspect xx' with these values: xx= 0.866 | 1.0 | 1.155 > > Results: > > aspect: 0.866 | 1.0 | 1.155 'xrange _1 1' > ------------------------------------------------------------- > vertical/horizontal unit length: 0.81 | 0.92 | 1.05 > height/width of grid frame: 0.87 | 1.0 | 1.15 > > 'aspect 1.0' does not produce an angle of 45 deg of tangent versus > horizontal at the root as expected.
Why would you expect this? My understanding is that aspect ratio was designed to compensate for pixels which were not square. If you are working on a machine with square pixels, the default aspect ratio (1) should be sufficient. In the general case (for example, xrange 0 1, yrange 0 1e6) plotting 0 0j1 1j1 1j0 0 is not going to give you something that looks like a square. > Try it with a different xrange and have a look how distorted it becomes: > > aspect: 0.866 | 1.0 | 1.155 'xrange _2 2' > ------------------------------------------------------------- > vertical/horizontal unit length: 1.57 | 1.83 | 2.13 > height/width of grid frame: 0.87 | 1.0 | 1.15 > > This clearly shows (at least to me) that the aspect value is > definitely governing the rectangular grid frame, and somehow (on the > way) influencing the units ratio. That makes sense. But I am not sure that I agree with your implication that we should view this as a problem. > I think that's a bug (as it should at least be the other way around). > Therefore I still plead for a separate pf option of e.g. 'compression > ratio' or 'units ratio' if possible. I am not sure if it's possible to implement such a thing -- it seems like the resulting system would be overdetermined. But I am also not clear on the purpose you had in mind for this mechanism. I think it has something to do with the shape over the drawing window? Thanks, -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
