- Messages
- 11
- Country

There’s a lot of information on this topic in this forum and online, some of which is no longer up to date. For this reason, it’s difficult to get an overview and create the optimal conditions for the best possible results without going through trial-and-error cycles.
That’s why I’m asking those of you who already have considerable experience with MSFS 2020 to answer a few quick questions on this topic.
1) Difference between BGL and CGL
What is the fundamental difference between CGL and BGL in short?
Based on the information I’ve found, I was able to test the following methods for creating a scene:
a) BGL/modelLib with Google Earth Decoder
The data generated by the decoder was compiled into a readable package, but the placement was incorrect. Buildings were offset by 3–5 meters, and some surfaces were located below the mesh. (What is the source of the error here?)
b) CGL with QGIS and tiles2bing
The tiles generated by exporting from QGIS were converted using tiles2bing, compiled into a CGL file, and loaded into the sim without any problems.
I noticed that the BGL package apparently rendered trees and buildings from the map data (i.e., photogrammetry?) and the CGL package rendered only "flat" textures. Is photogrammetry even available for an “offline” scenery? If so, is it available for both BGL and CGL scenes, or only based on a BGL file?
2) Zoom levels and LODs
What should I keep in mind regarding zoom levels? For the test, I only exported and compiled tiles at zoom level 19. Is that generally sufficient or optimal, or should I also create tiles at lower zoom levels? How do the zoom levels affect the scenery?
3) Optimal Setup
Which method is recommended for achieving the highest possible scenery quality — BGL or CGL? What is the optimal setup for a CGL scenery and a BGL scenery? What limitations should be taken into account during compilation? For example, is it true that the compiler can process a maximum of 7.5 GB of map/image data?
That’s why I’m asking those of you who already have considerable experience with MSFS 2020 to answer a few quick questions on this topic.
1) Difference between BGL and CGL
What is the fundamental difference between CGL and BGL in short?
Based on the information I’ve found, I was able to test the following methods for creating a scene:
a) BGL/modelLib with Google Earth Decoder
The data generated by the decoder was compiled into a readable package, but the placement was incorrect. Buildings were offset by 3–5 meters, and some surfaces were located below the mesh. (What is the source of the error here?)
b) CGL with QGIS and tiles2bing
The tiles generated by exporting from QGIS were converted using tiles2bing, compiled into a CGL file, and loaded into the sim without any problems.
I noticed that the BGL package apparently rendered trees and buildings from the map data (i.e., photogrammetry?) and the CGL package rendered only "flat" textures. Is photogrammetry even available for an “offline” scenery? If so, is it available for both BGL and CGL scenes, or only based on a BGL file?
2) Zoom levels and LODs
What should I keep in mind regarding zoom levels? For the test, I only exported and compiled tiles at zoom level 19. Is that generally sufficient or optimal, or should I also create tiles at lower zoom levels? How do the zoom levels affect the scenery?
3) Optimal Setup
Which method is recommended for achieving the highest possible scenery quality — BGL or CGL? What is the optimal setup for a CGL scenery and a BGL scenery? What limitations should be taken into account during compilation? For example, is it true that the compiler can process a maximum of 7.5 GB of map/image data?


