I try adding gmt events -Rd -JG${MOVIE_COL0}/${MOVIE_COL1}/15c "temp_timefile.txt" -Ar100c > temp_line_points.txt in main.sh but at the end of the movie, the start of the lines goes north (to the Bering Strait).
I got your first movie (Venice) to work well without any problems (I think, you should check). Here are the important parts:
The 10.66k is not good enough, but that is probably just a tiny minor part. The other part is that while the default is great circle distances, the default ellipsoid is WGS-84. So there is possibly some issues related to geocentric vs geodetic latitudes. In the end, I used gmt set PROJ_ELLIPSOID Sphere in both scripts and used -I10.645980883k.
Probably much more important: I experimented with a much higher dpi in -Ar. I think twice the dpi is OK (2160) but I also ran with 4320. This yields a much finer sampling of the circles and since we are going from continuous lines to discrete circles the jittering is probably due to picking the next nearest circle. Please give this a try. If you agree this works well and removes the jittering then I think I will declare a but and add that factor or 2 (or 4) into events so that we can still do -Ar1080 but get the finer resampling.
Please try this and give feedback and if you agree it is working well then I will do a PR on events before we move on to other cases.
No, you still need to run events -Ar1080 inside your main script. I meant we do not need to experiment with higher values in that call as the problem was in sample1d.
Yes probably OK since your frames are all very similar in local map scale so running -Ardpi once may be OK. But in general the recommendation I will add to the events docs is to say that if your -R -J are changing with the frames then you may need to rerun that command.
Only now realizing you are using -Ar in the plotting call. The point of the -Ardpi is to convert the line to a dense set of points that you can plot as circles with -Sc and then control via -M -E. Plotting with -Ar cannot allow any change to the line tip.
Just to be clear.
I made these two animations:
I create these two lines and in the left (right) video I plot the first (second) line. The movie looks fine when I plot the line created without events -Ar100c.
I need to work up some better documentation for dealing with lines in events and play a bit myself too. I think your case works because events will cut the line exactly at the timepoints, insert the intersection point as the last point, and plot the line. So it does not need -Ardpi at all then, of course.
