Bugreport plus feature requests

  1. Bug (almost): in Revopoint PC each time I open a Project, it drops back to default folder (Program Files\Revoscan 5) instead of remembering the last folder used. It does this way with both launching the Revoscan and after just pressing Home. Although when I use File->Import (or CTRL+I, handy BTW) it does remember the last used folder. It’s annoying, please fix it;
  2. Bug (for sure): when scanning with Android (Range was used) and after some scanning starting getting dual surfaces due tracking drift (“Revoscan” and “tracking drift” seems to be synonyms for a long time, poisoing the scanning process and hampering the abilty for Revopoint scannes to be a working tool instead of geek’s toy - mark my word), I paused the scanning and pressed Undo one or few times till those scrap scan parts disappear. Then just gone back to the menu and started a new scan. But when I have imported those project to PC and opened them, I’ve got all those scrap “overscans” back. Please fix what is to be read by Revoscan taking those Undo commands into consideration. Now scan data should be messed with to find how many frames I must delete to get rid of those crappy parts;
  3. Feature request based on the paragraph above: can you add a Play Bar like thing where you can check the frames scanned and be able to crop the end frames if necessary. That’s due to when you turn your scanning off, it is common to have some hand shake that add some crappy unaligned point clouds.
  4. In the future releases of Revoscan maybe you should truncate those empty data parts in the depth buffers? Instead of 640x400, seems it should be something like 518x400. It could save about 20% of disk space. Added a picture to make it more clear:
    You still can have compatibility with either checking the first frame size before opening a sequence or just checking project version in metadata;
  5. I know this feature request is some kind of too big one maybe, but it could save A LOT if implemented. Can you add a Retrack feature to Revoscan. When it will recalculate all the frames alignment from the beginning, but involving more complex math than in realtime scan. When it checks the frame not only being “connected” to the previous one, but to the entire point cloud field. Or something like this. Let it calculate complex 2K frames projects for a hour or two, but thus that imminent tracking drift could be greatly diminished. OK, we can’t have a complex realtime check fro what we scan, but maybe we can save the scans in the post process this way.


  1. One more bug: Lasso Selection tool doesn’t work after Aligning Meshes. It allow to select the points, but you can no more delete them untill you reload the software
1 Like

Just click the Merge tab and return back to edit tab to unfreeze it , not need to reload .
For some reason the main OpenGL window do not show the actual view when it happens.

I get it sometimes too , but not always .

It is of course a bug that needs to be fixed .

1 Like

One more feature request. It is relatively easy to implement, but it can seriousely benefit to the resulting quality by manually filtering out the “broken” frames. Sorry, no pictures since it would a bit hard to arrange.

Manual frame prefilter mode. This should work before fusing the points to the finalized point cloud. The purpose is to manually delete the frames or group of frames that do not fit to the point cloud due to misalignment. It would be easier to delete the frames than try to separate their influence to the point cloud when they are already fused.

When entering the frame filtering mode, we point the cursor at the point that belongs to the frame that “sticks out”. All the points belonging to this frame become highlighted, so you can see what you are going to exclude. Pressing the left mouse button you “delete” that frame hiding it both from being visible and from final calculation at the fusing stage.

By pressing, [ and ] or + and - you increase or decrease number of frames scooped by that function. For example, you point out to the point that belongs to 232 frame, but with that increased scoop you get highlighted and eventually deleted not only 232 frame, but also 231 and 233 frames. The “midframe” is highlighted with a bit different color to visually separate it from the “side frames”.

You can revert the process either by clicking global unhide (since it is easier to implement) or, that would be better, having mode that make all those hidden frames partially visible, so you can click them back to unhide.

Yes, you often can see meshing problems only after the meshing is done, but you can always go back to the beginning from the meshed point cloud. So you can have a workflow by quickly getting a draft model, examining it for the broken parts and then going back to the beginning, so you can filter out the frames that do not match with the point cloud.