New owner of a MetroX here, with a question about the resolution of the mesh produced by RevoScan5. I’m scanning a part of an acoustic guitar fretboard. The fretboard is a dull black with shiny metal frets. I’d like to squeeze out the highest resolution I can get out of the MetroX. I’ve scanned with parallel and crossed lines, with either Metallic Shiny, Black or General selected as the object type. I set the Target Point Distance as low as it will go (0.15mm) and capture around 10K frames, fuse them, then trim the scan down to a region roughy 50-60mm on a side.
When I create the mesh I choose the finest grid spacing (highest quality) allowed, which is 20µ. First, why does the software let you set the spacing that small when the point spacing is 150µ?
Second, even though I selected a 20µ spacing (figured I try it), the mesh ends up with vertices that are spaced more like 130-150µ. In the attached image, on the left, is a histogram of the vertex spacing (made in Meshlab). I guess it’s reasonable to get a ~150µ mesh spacing based on the lower limit of 150µ for the point cloud. I must be missing something re the offer of 20µ when meshing.
Apologies if this is a dumb question. I’m new to 3D scanners and am trying to figure out just how much detail I can capture.
For the best results use the Fusing settings as the GRID settings under meshing , setting the meshing at the highest resolution only induce artificial points and you don’t need it .
Set the GRID according to your fusing settings of 0.15mm and you get the perfect results according to the scan.
Thanks. Yes, that makes sense. But it doesn’t seem to be the way the software works. If I set the mesh resolution to 20µ I still get mesh vertices spaced at 130-150µ, which could be confusing to someone trying to get a high-res scan.
In contrast, I have a friend with a Raptor Pro. The lowest point spacing they allow is 100µ. After completing a scan, and fusing, there is an optimization step with an “ultra detail” option that allows a mesh spacing of 20µ. The result is a 20x increase in points in the cloud, and a mesh spacing of a few tens of µ. That doesn’t make a lot of sense because, as you noted, it just introduces “artificial points”, at a resolution beyond that in the underlying point cloud.
I suppose it’s good that MetroX doesn’t generate unreasonably fine meshes like that. But it would be nice if they’d be more clear about it and not let the user pick 20µ (which isn’t actually used), because that leads the user to think they have a really high-res scan, when in fact they don’t.
It is how the software works and how it was designed because it is based on my request and the reason it is there .
You can’t get higher resolution from a point cloud that have less points .
Having a cube with 8 points meshing at level 16 will not give you additional details only artificial points between that can induce bumps and noises since it is not a subdivision.
For best results true to the original scan and accuracy follow the tip I give you early .
Doing anything else will not create better mesh and that is a fact.
Anybody that have some experiences with 3D modeling know what GRID settings means .
Any software that pretending to create more than a point cloud offers is simple BS as that is impossible , it just induce artificial points between and nothing else , all software doing it and there are never a limitations regarding the settings .
It is like having a point cloud cube with 8 points and magically increase it quality by creating mesh with 16 or 32 points , it still don’t change nothing as the original cage will be always 8 points .
When it get about scanning , the most precious you getting is the main point cloud , meshing can destroy the accuracy and details so follow the software guide to get the best results ..
Follow the GRID settings and you will be just fine .
I think we’re saying the same thing. I’m just wondering why the MetroX software would let a user think they are getting a 20µ mesh when they are not. I agree there is no point in meshing finer than the spacing in the cloud. So why is 20µ even an option in the software when building the mesh?
But that brings up another point I’d like to get your opinion on - because I’m not exactly sure how the fusion process works. I could imagine that after taking several thousand frames there could end up being points that are a few tens of µ apart, even though the spacing in any one frame is set to 150µ. Assuming the accuracy of the scanner hardware good enough, one could presumably leave such closely spaced points in the fused result In that case it would make sense to have an option for a fine mesh, less than 150µ. An adaptive gridding algorithm would preserve the detail where it actually exists and increase the mesh spacing away from those closely spaced points. I wonder if that’s how the competitor software works, during their optimization step, where one can specify a fine mesh spacing. And is this the reason the MetroX software lets me pick 20µ when meshing?
I appreciate your comments. I’m just trying to understand how this process works.
nobody think that , if you have 0.15 mm point cloud you can’t really get better resolution mesh as that is technically impossible .
just because you have a slider option don’t means you should pull everything at highest settings .
The software don’t allow always the higher settings when meshing as it depends of the scanner volume , as very big objects will make your system run out of memory so the settings corresponds with scanned volume .
I don’t like to be limited as sometimes you can import a higher distance point cloud and mesh it at higher settings , limitations are not what we want with this version of RS for that reason it is the way it is .
Auto mode creates very fine point cloud and higher than the laser mode , sometimes people merge different scans so they need to be able to mesh it at the specific settings that are much below 0.15 mm point distance .
The accuracy is build into the hardware for that reason the laser mode despite having point cloud minimal setting at 0.15 mm still can preserve the finest object details of 0.03 mm or slightly below as long you follow the proper GRID settings .
That’s all I have for you on this topic and seriously there is nothing else to it .
Hi. I tried doing laser scanning with subsequent fusing at 0.05mm (on modified software version 5.6.7). It resulted in a lot of “noise”, and subsequent meshing with 0.03 did not change the quality. Although visually (compared to the typical value of 0.15), it seems that small details look “clearer”.
Hi @OBN_RacerMan the maximum fusing that delivering ok result is fusion not higher than 0.10 mm on modified version for Laser mode and nothing below that.
I was there before the product was even released to test out best options for quality , but since I am under NDA I can’t discuss stuff from pre production.
Meshing at higher quality than the GRID don’t really give anything extra because technically there is nothing that can improve it . It only makes the scan more dense but can’t create more details. So just illusion
You can’t modify the fusing settings of the software without adjusting also the scanning algorithms so the results will be meh .. on lower settings just for fusing .
This is exactly what I wanted to demonstrate to K3vinH - that decreasing the fusion value shows an apparent improvement, but in fact leads to a sharp increase in the noise level in the scans.
