Loading of lower zoom detail-level tiles in basemap

[jjkohler]

Currently ET loads the more appropriate detail base layer: https://earthtime.org/explore#v=40.44103,-79.95157,11.145,latLng&t=3.3&ps=50&l=bdrk,bdrk_detailed,acs2016_all_income&bt=19840101&et=20191231&startDwell=0&endDwell=0

While Refactor shows the less detailed tiles

https://staging.earthtime.org/explore#v=40.44845,-79.95432,10.254,latLng&t=1.31&ps=50&l=bdrk_detailed,acs2016_all_income&bt=19840101&et=20191231&startDwell=0&endDwell=0

This does not seem fully explained by the differences in projection or waypoints. Even when zoomed to nearly exactly equivalent zoom levels, the underlying base map detail level is different.

[rsargent]

With just bdrk_detailed, comparing with golden.earthtime.org: https://golden.earthtime.org/explore#v=40.39317,-79.88281,10.2,latLng&l=bdrk_detailed https://staging.earthtime.org/explore#v=40.39317,-79.88281,10.2,latLng&l=bdrk_detailed

Some zooms+window sizes look identical, some are coarser with staging.

[rsargent]

Fixed in commit acf40839b7f4451752aac5cab3cb38a299ffa6da

[jjkohler]

golden https://golden.earthtime.org/explore#waypoints=1heLmeuPp7j4itr0cK8H4chugOpp7cU8p_VEB9CFfPlY.0&v=-10.99506,-61.19203,8.525,latLng&t=3.24&ps=100&l=blsat&bt=19840101&et=20161231&startDwell=0&endDwell=0

Staging https://staging.earthtime.org/explore#waypoints=1heLmeuPp7j4itr0cK8H4chugOpp7cU8p_VEB9CFfPlY.0&v=-11.04999,-61.18178,8.345,latLng&t=0.08&ps=90&l=blsat,wdpa&bt=19840101&et=20191231&startDwell=0&endDwell=0 Land Sat imagery still loading in at decreased resolution.

[rsargent]

Paul suggests check on level threshold for time machine layers

[pdille]

I took a peek and this is all coming back to me now.

First off, we are looking at two different datasets in our Landsat comparisions:

Golden is using 2016 data: https://storage.googleapis.com/earthengine-timelapse/herwig/earthtime_annual_1984_2016/v14/1068x600/1/0/0.mp4

Staging is using 2019 data: https://storage.googleapis.com/timelapse-test/tmp/herwig/export/cmu_20200106_resampled/1068x600/1/0/0.mp4

I'm recalling again Chris H. saying that he previously (back in 2016) had full control over ffmpeg settings and was able to apply various sharpening filters. (this is why 2016 set looks almost too sharp) This is no longer the case now at Google and they are forced to use company wide settings and can only modify things like fps and dimensions.

[pdille]

I would suggest breaking off the Landsat specific discussion from the initial issue raised here. They are different things.

[jjkohler]

Current issue is not the same as what was originally opened and has been opened separately in #159