Shifted image extents in 2.0.0.b3 #6773
Assignees
Milestone
Comments
mdboom
added a commit
to mdboom/matplotlib
that referenced
this issue
Jul 17, 2016
...on non-rectilinear axes
|
Good eye! There's a very subtle adjustment in the image code to account for the fact that the edges of normal rectilinear axes are pixel-snapped -- the images are correspondingly pixel snapped as well. For non-rectilinear axes, as here, that adjustment isn't necessary. #6776 should fix this. |
QuLogic
added a commit
to QuLogic/matplotlib
that referenced
this issue
Nov 16, 2016
I don't understand why there's a + 1 here, but it rounds up to the next-next pixel instead of the next pixel. This makes images extend past the end of the frame. Fixes matplotlib#6773. Fixes matplotlib#7350.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Comparing images in 1.5.2 and 2.0.0b3 shows a reasonably large change. Originally, I thought the sampling was just different and it was intended, but closer inspection seems to indicate that it might actually be a small shift in the extent.
I don't have a simple example, but here are two from Cartopy to illustrate what I mean:
If you switch between them, you can see that 2.0.0b3 seems higher, though I think it's actually the right edge that's a bit further out.
A better example is:
It's clear on this one that the image is just a bit wider in 2.0.0b3 vs 1.5.2.
The code for these two plots is global_map.py and test_regional_projected.
It looks like one edge is off-by-one or something, but it's also possible that the transforms were just corrected in some way.
The text was updated successfully, but these errors were encountered: