Its live but it doesnāt work with pbr at the current moment i think, thatās what i meant by when will it be released
Oh I see, I wasnāt able to see what you were replying to if you did. PBR is one of the features we are working on adding to the importer. I donāt have a time estimate at the moment unfortunately.
I know the plan is to merge importers, but I also know there is separate work going on in the avatar importer (eg layered clothing). Have they reached a point where these projects will all go into the new importer? Just wondering where I should be watching for fixes. Thanks!
oh my bad, thank you for clarifying!
Yes, we have a team working on bringing avatar capabilities of this new importer up to par with the existing avatar importer.
That importer has issues too. I was wondering more if we could expect any further improvements to that importer, or if that has been tabled for this project.
This is the particular bug Iām watching:
Iām not on that team, but work is still ongoing for layered clothing and bugs should get fixed for that and a few other features on the old importer.
Filing a bug report as a separate topic for your issue should get the right eyes on it.
Finally, another triangle limit increase. Will there ever be another triangle limit increase in the future?
I honestly think that if they go big then itād be too much for the servers because of how many calculations are needed
Agreed. Also given how many tricks there are to make low poly look high poly, this is more about the learning curve than system constraints now.
I stumbled upon this importer, because I have an issue (probably with UV maps) and textures are not rendered correctly for certain meshes. Iāve seen that people had similar issues before and this was supposed to help, but the new importer didnāt solve it for me. Maybe itās also a problem with the importer, but Iām not very experienced in this field. I recommend you see my thread:
Heya!
When will the first week update come out?
Is it delayed or did something happen behind the scenes?
Hi all, the upload throttling feature should be live, I was able to import a scene with 729 individual meshes. It was quite slow (as expected due to the back-end throttling), but it worked. Please let me know if you run into any issues importing larges scenes.
Note: We will be tuning some settings to see if we can make it a bit faster without annoying our cloud friends too much!
I liked the progress bar! Maybe it was there before, but I happened to need it today.
Redwood scale? Probably. Doesnāt take much to have a forest.
I wonder if we can import the mesh offline faster since it doesnāt need to throttle with the cloud. Once we publish the game, every mesh is uploaded at once?
Itās suppose to do that, so I guess it does. But what do you mean by doesnāt throttle with the cloud? Anything published to Roblox, or saved to Roblox, is uploaded to the Roblox cloud so that you can re-open the experience or place later.
I am not entirely sure how the mesh importer API works, but I am assuming that the mesh gets imported to the client and uploaded to the cloud; therefore, there is some throttle in importing meshes, because they take a while to upload. However, if you remove the uploading per mesh part and only import them to the client, all the mesh would only be uploaded when you save to Roblox or publish to Roblox (to the cloud) resulting in a faster mesh import time. This is just a hypothesis. I had the same issues when doing massive terrain parts conversion, and found out it was faster when it was done offline and uploaded to Roblox once all ready.
This is especially annoying because models canāt be deleted. As a workaround, Iāve resorted to giving the file a dirty name, which causes the entire upload to get blocked by the filter.
Unrelated: I donāt know if this was mentioned already, but the File Dimensions field doesnāt seem to be calculated correctly. For example, if I try importing the default Rthro Normal rig, the height should be nearly 6 studs, but it is displayed as only 2.08. There is no scale unit that corresponds to the expected value. Compared to the actual model, the dimensions (in red) looks like this:
Each component appears to be getting overwritten from the size of some arbitrary part. e.g. X appears to be taken from the lower torso, Y from the upper torso, and Z from one of the feet.
Is it possible to not generate unique meshId for linked mesh or meshpart with its base prefab the same?
Having 300 unique meshId would cost dramatic longer loading times. This is a big issue on bad network players and streaming in/out.
Does all those cubes share the same MeshId?