Advanced fusing results always produce much better meshes for me than standard fusing and manual clean up.
However, advanced fusing automatically limits the fuse distance I can go to, depending on the size of the scan.
This wouldn’t be a problem if it didn’t affect the details of the mesh… but using the exact same scan settings (far distance, high accuracy, feature tracking), the smaller scans look much more detailed and much better, not because they have richer data, but just because the advanced fuse can go to a lower point distance with a smaller model.
Meanwhile, I tried to get around this by processing the larger scans with standard fusing and manual cleanup, but despite using overlap detection, isolation, and manually selecting big chunks of mis-aligned frames, unfortunately I just can’t get a standard fuse manual clean up process anywhere near the advanced fuse.
So much data and detail is being left on the table, I can literally see it when I compare small vs. large scans on the same scan settings - So my question is: How do I get around this? Scan in separate scans, import them separately, advanced fuse them all, and then combine them? That could work but what if I’ve already scanned a full object as one scan and can’t scan the object or person again? Scanning in sections also adds time to the scan, which doesn’t matter for an inanimate object, but for humans, it increases errors as they shift about - so if I could figure out a workflow to do a single scan and then either
a)process with advance fusion in sections made after
b)process it with standard fusion and get results closer to advance fusion, that would work.
You can’t get big scans to have the same details as a small scan because you will run out of all memories available on PC system , the mesh would be so big that you can’t even move it in your screen in the viewport .
Advanced mode do not limiting anything , it has advanced algorithms and it produces best results possible .
Just ignore the settings as it will give you nothing better if that was set in lower settings .
Standard settings are for those that want to capture the raw data with noises and all the bad data and have advanced skills to clean it all manually in other software , and I never use it despite having that skills , just waste of time doing it .
Just use the advanced mode as it produces the best results and the accuracy is as it should. Especially for fiber details.
Also remember that the distance you scanning at are responsible for the details , how closer the scanner is to the object how better it captures the details , all scanners with Structure light works this way .
So actually you’re right, and this is exactly why I’m asking: I prefer the noise and details because I re-sculpt manually in Z. For me I do notice a big difference in fuse point, it’s always better to lower it from the recommendation, much much better. So this is why I’m trying to figure out how to best use standard fusing, and so far getting pretty bad results that require much more work and time to generate but at least preserve some of the noise and details. I would definitely prefer to take the advanced fusion past the bottom limit (not all the way as low as standard can go, and I understand your point), but definitely past the bottom limit always suggested, because their limit always gets a smooth result at the cost of blurring everything. My dream is to get POP3 level scanning of entire body but with the distance and speed of MIRACO or RANGE so that I can scan more dynamic poses without having to repose or model them as much, etc.
And by the way your full body scan using POP3 was awesome, it showed just how much could be done. In my case I’m forced to scan T-poses, really not even A-poses, so I’m still thinking about solutions.
The noises are not nesesery, it is not any data , I do scan humans and work in Zbrush , you really don’t need that all junk data , there is nothing more than you will get using Advanced mode , just not suggest yourself by the proposed settings , just slide the slider to get the best accuracy in advanced mode .
And always manual adjusting . And skip the quick edit processing.
You have 2 choices , you getting the easy way or you fight with data that don’t give you anything extra but extra work.
You saw my showcases , I never use Standards mode for fusing.
That’s rather impossible , unless you scan that person partially using Miraco Near mode and then stitch the parts together to blow off your computer RAM . You know even Zbrush has own limits regarding polygons count.
I tried do that before ending with head of 179 K if polygon that Zbrush or other software don’t wanted to import after waiting for 1 hours and that’s just head. Now imagine the rest of the body stitched together. Just unrealistic expectation.
That’s were I reached my dream .. or rather the end of it
If scanned partially humans was possible like for example statues , then you still could do partial sections like 30 millions per section as different sub tools , then can project the data to subdivided dummies. Trust me I tried everything and that’s was the option .
But I had to decimate the scans in Revo Scan to 30 millions of polygons per section and lost little details doing it , however it’s a lot of work and retouching after.
Yes but I used the Body mode , just the face I think was separate. On top the software reduced the meshing due to volume to optimal so never at full details according the point cloud.
On this case you need to export your point cloud after full cleaning to iCloud Compare and the mesh it unlimitedly at the proper settings to get the max from the point cloud yet not induce artificial points, as it can take over the fine details and destroy it if the meshing settings are too high .
Lots of practice to find your own workflow the way you like . Because as usual other person workflow is not always going with your flow or way of thinking.
I still think the limit is too high, especially given there’s no way to override it. Take that RevoscanX computer profile thingy and apply different limits based on that or something.
Developers worried that scans will be too noisy or software crashes, but people complain whether it’s noisy or soft. Can’t win that game with an arbitrary setting limitation.
We don’t talk here about RevoscanX
And it’s limitation are lifted producing so big scans in sizes that are no more usable outside RevoscanX so what’s the point of having it ?
There are no scanner or software limitation, the question is what software allows you to import full body scan meshed at the polygons count of 1 bilion of polygons ending editable in real time . Graphic cards do have also limitations.
I don’t know any and Zbrush in this case is not either , as it allows for best performance only 30 millions of polygons , where RevoscanX produced 179 millions of polygons only for a face . That’s where the adventure ends.
Regarding Revoscan regular version the meshing of level 8 for both Range series and MIRACO is optimal vs the GRID settings of cloud point , sliding the slider to maximum will not create higher resolution as the point cloud itself . Sometimes there can be still room for get out more of that point cloud but not as much any more.
Example , you have a point cloud of a cube , 8 points , meshing at 100 times the points value will not produce better cube as the original 8 points . It will induce artificial points that do not add any details to the model and only pollute it with noises and uneven surface. Just the opposite.
Apologies, I thought we were discussing Advanced fusion slider limitations not meshing. In which case my reference to RevoscanX was that it performed a sort of performance scan of your PC which could be leveraged in Revoscan to allow a relaxation of some of the arbitrary limits.
In either case, I don’t usually agree with arbitrary limitations like these as they assume the developers know everyone’s use cases and what’s best for them.
The limitation in RevoScanX in advanced mode for example for the Laser mode is because that’s the best the algorithms can produce at this moment, I personally tested the over limits and all results on higher settings , the best results was picked for the public version . There was not reason to have it higher as it affected too much the details with artificial noises. So a real algorithms limitation here and not done on purpose .
The advanced fusing mode settings do not affect as much the accuracy , as that is already pre set with the hardware and can’t be changed as it happening while capture. The slider is more for the resolution. Just because it says it is 0.15 don’t means the accuracy is 0.15, just the resolution.
RevoScanX is much more limit free than any other version before , for that reason it is also a separate software with higher system requirements than traditional Revo Scan .
The latest Revo Scan have some limitation lifted to higher settings as well in the last version like for example Mini 2 scans can be meshed at level 8 now as well.
Personally I prefer to not have any limitations and have a free will to choice the best settings for myself as needed. But then you have users that run PC with 8 or 16 GB of RAM complaining it crashes their PC every time they trying to mesh a Scan at highest level .
Answer is simple " get more RAM " but that is not the answer they want to hear. Bitter sweet situation .
