[JHV 3D] strange projection error when viewing from angle
Bug #994617 reported by
Keith Hughitt
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| JHelioviewer |
Confirmed
|
Low
|
Simon Spoerri | ||
Bug Description
Here is another error I noticed when zooming in/out when the image is viewed at an angle. See screenshots for before/after zooming.
Revision history for this message
| Keith Hughitt (keith-hughitt) wrote | #1 |
Revision history for this message
| Keith Hughitt (keith-hughitt) wrote | #2 |
Revision history for this message
| Simon Spoerri (simon-spoerri) wrote | #3 |
| Changed in jhelioviewer: | |
| assignee: | nobody → Daniel Mueller (dmueller-esa) |
| Changed in jhelioviewer: | |
| importance: | Undecided → Low |
| status: | New → Confirmed |
| assignee: | Daniel Mueller (dmueller-esa) → Simon Spoerri (simon-spoerri) |
To post a comment you must log in.
from my technical point of view this is not really an error but an accepted performance trade-off.
Here is why:
The ROI is determined by casting rays through the viewport and detecting the hitpoints on an image layer. The rays used for the hitpoint detection are cast through the corners and the middles of the edges. The hole in the image you're seeing can occur at the solar poles, when the pole lies between a corner and the middle of the bottom/top edge. This is because the maximal roi is mapped onto a mesh by getting the spherical coordinates for the maximally spanning roi area.
Reducing this effect could either mean casting more rays, which is an expensive thing to do, or by default increase the detected roi by some percentage, which is also somewhat subpar and neitzer solution does guarantee an optimal solution.
Furthermore, a better algorithm/model to transform the detected roi to the image mesh would take quite some time.
Daniel, could you please prioritize this bug?