1. A concise explanation of the problem you're experiencing.
A .glb converted from a small .dae has duplicate parts, when loaded by ThreeJS GLTFLoader or
2. A minimal code example. If you've found a bug, this helps us reproduce and repair it.
I have a small collada model DMG_DMU50.dae (attached), and ran the following command to convert it to the GLTF binary format (latest version for x64):
collada2gltf-bin DMG_DMU50.dae DMG_DMU50.glb -b
But when I loaded this DMG_DMU50.glb into ThreeJS using the GLTFLoader, the model seemed to be doubled - mirrored itself (attached screenshot). Then I tried loading the same .glb file with https://blackthread.io/gltf-converter/, the same problem exists.
Another interesting observation is that, when I drag/drop the original DMG_DMU50.dae into your web application: http://52.4.31.236/convertmodel.html, the model is displayed correctly, i.e., no duplicate or mirrored parts. Further, if I load the output .glb from this web app with either GLTFLoader or blackthread converter, it's also working properly.So it looks like the DMG_DMU50.glb I generated using the command line is incorrect.
What is the right command (with options) should I use? Why is the .glb output from your web application correct, as it is using the same converter as I do?
Attached please find the screenshots, original .dae and the converted .glb from the command line. Thank you in advance for any help!
Thank you for your reply! I have posted the issue:
In the meanwhile, how can we know the exact version that the Cesium web app is using? I'd like to try that version. It's worth to check as it may help to determine if this is a regression.
So it is indeed a regression. The model converter we have up is supposed to go away/get moved soon (hence why it’s the only thing still active on that domain-less address).
When you said the model converter is going away/getting moved, it just means that it will be moved to another place, right? It’s not something that you will obsolete/deprecate soon, right?