You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been going over the code and why the mouse focus loading isn't working properly
It might be good to clarify what the best tile loading order is and why, so I will try to lay out what I think it is here for discussion.
My strategy previously was loading the lowest resolution first is the best way to get something on the screen immediately. There is not much point loading lower resolution tiles than the z level where where 1 tile could cover the whole area (but wont always so it's 1-4 tiles). But loading more zoom levels out is not that expensive either.
First load the visible area from low res to high and form the mouse cursor out, so:
The tiles at the zoom level that could cover the whole view area (1-4 tiles)
The tile under the mouse cursor at the next zoom level out
The ring of tiles around that tile
the ring that is at the edge of the visible area
... Repeat for zoom levels and rings up to the current zoom level
The tile under the cursor at the current zoom level
Concentric rings around the mouse position the current zoom until the visible area is covered
After loading visible tiles, load a halo around the visible area so that panning and zooming
doesn't cause flashing white artifacts. The order doesn't matter as much but probably low res first is best here too
The outer halo (not shown in the plot) of the furthest out zoom in the range
... Repeat for outer halos to the current zoom level
The text was updated successfully, but these errors were encountered:
This makes sense, your resolution will never be big enough that this is truly harmful. But IMO we could probably drop a lot of the intermediate zoom levels, maybe make a hard limit of 3 intermediate zoom levels for truly large plots.
I'm not sure I understand this reasoning. It doesn't save much to cut intermediate layers, as the main layer is always 3x more tiles than all the other layers combined.
And the effect of the area you are looking at getting better resolution as quickly as possible will make it seem faster to load the intermediate levels than to jump straight to the desired level its more about perception than download counts. Skipping levels means you really do need to wait to see much detail in the whole image. Google earth loads all the levels.
But these are details. The main problem is the current code is doing the opposite of this. Background tiles are loaded last rather than first, and the rings around the mouse do not work at all.
I've been going over the code and why the mouse focus loading isn't working properly
It might be good to clarify what the best tile loading order is and why, so I will try to lay out what I think it is here for discussion.
My strategy previously was loading the lowest resolution first is the best way to get something on the screen immediately. There is not much point loading lower resolution tiles than the z level where where 1 tile could cover the whole area (but wont always so it's 1-4 tiles). But loading more zoom levels out is not that expensive either.
First load the visible area from low res to high and form the mouse cursor out, so:
... Repeat for zoom levels and rings up to the current zoom level
After loading visible tiles, load a halo around the visible area so that panning and zooming
doesn't cause flashing white artifacts. The order doesn't matter as much but probably low res first is best here too
... Repeat for outer halos to the current zoom level
The text was updated successfully, but these errors were encountered: