Range user here: This a great addition! Makes scanning way easier. I would like to see implemented of an option to colorized the point cloud based on its density as you’re scanning; to know if you have enough point cloud data as well as a hybrid feature to look at markers and features.
3 Likes
Hi @Jayson there is really not density while capturing cell frames so it is not possible to decrease or increase it while scanning , the pitch point cloud fusing density is between 0.1mm to 0.3mm and it is determined by the hardware itself and the distance you are scanning .
This means if you collected one frame or 10 frames of the same spot it will not influence the final fused point cloud quality , but actually decrease it as additional frames need to be manually removed before meshing .
For comparison an object needs minimal 12 frame cells to be completed in one 360 degrees rotation , anything else will be removed in the process .
In short how less frame cells is captured of the same area how better .
And the density (pitch point) is based on the distance between the object and the scanner so scanning with Range at the distance of 30 cm will produce much higher density than at 80 cm but can’t be greater than 0.1mm.
the color indicator is already there Too near-Excellent-Too Far.
So , for the best results focus on the proper distance , and not how much frame cells you captured of the area .
It actually would be great to see in real time a color indicator for overlapping of the frame cells , as many users tend to do that thinking they get better result after , but in reality , less is better .
In the future updates you will have an option to equalize the pitch point density of your already merged and meshed scans , reducing the size while automatic focusing on the details where it is mostly needed .
Please post your future suggestions and ideas to this thread : Leave your feedback & suggestions on the Revo Scan 5.0
It is recorded by the dev.team so , everything is possible in the future evolution of the software .
2 Likes
Thanks for the explanation. Question: When will full M1 support be a priority; (I use both PC and Mac)? Also, what are the main bottlenecks for getting faster frames. Is this depicted by shutter speed, say with the RGB camera as well as exposure of the depth camera, frame-grabbing and such, or more of a RAM bottleneck?
Sorry but I can’t answer you that question.
The main frame speed is around 10FPS in Feature Mode and 7 FPS in Marker mode , that’s normal operation , right now the frames are much higher for POP2 and MINI , and I don’t see any speed difference between faster or slower RAM .
WiFi or via USB .
However slow and interrupted WiFI signal can slow down the frame rate , also slow hard driver where the project is saved in real time can delay the processing time .
Also scanning objects without background or bottom will speed up the frame rate , how more of the objects is visible in the preview cam how slower can be the speed since a lot more of data is processed .
So if you scanning object, please avoid to scanning additional surfaces that are not important for your scan if the FPS is your priority, however I would not focus too much on the Frame rate , it becomes lately a little obsession by many and that is really not the most important thing here. If your object have enough features you can still get the same results with 6 or 18 frames per sec .
Pay attention that your system is not choppy while scanning , if the frames can’t be written down in real time they are placed temporary in the RAM before written down .
A lot of the processing is actually using. RAM in the process .
But there is sadly nothing much that you can do to speed up additionally the frame rate as the software is already set on optimal performance and the hardware reached it’s limits .
I am encountering a persistent problem with the latest update of the Windows and Mac operating systems, resulting in a maximum frame rate of 6.8 frames per second (fps) with occasional sporadic bursts. My previous performance was approximately 11-12 fps with equivalent scan procedures.
The utilization of USB 3.0 technology plays a pivotal role in ensuring rapid data transmission. I have conducted thorough examinations on this matter for four consecutive days. As someone with an extensive background in photogrammetry and structured light, I possess a comprehensive understanding of the technology’s intricacies. I have discovered that LED lighting does not influence the environment in any way. On the other hand, when scanning dark objects, particles on the surface become more prominent, which can enhance the IR sensor’s reflectivity. However, the clashing wavelengths of light may create issues. I have subjected this device to various tests to test its capacity fully. Unfortunately, I am experiencing issues with tracking accuracy. This device’s tracking algorithm needs further refinement, especially when scanning dark and intricate surfaces such as vehicles. Regrettably, despite the increased output speed, I am forced to revert to photogrammetry. Although the results are not necessarily superior, tracking concerns persist, requiring me to undo and repeat scans. Consequently, I am considering returning this product and upgrading to the EinStar for an additional $300, which offers better tracking capabilities and eliminates the need for beta testing with a “completed” product that costs $779.
I actually getting the frame rate much faster on Windows using the last update what is above 11FPS in High Accuracy mode compared to preview build that was much slower .
Since you understand structured light technology you should know that dark surfaces made from black plastic, PVC , silicone or paint do absorb infrared light and fir that reason it is difficult to capture the parts using POP2 or Mini that do not have enough power for that tyoe of surfaves . However Range can deal with some of them very well , the last update offers automatic depth camera exposure what is very helpful while scanning lighter and darker surfaces in the same session .
Algorithms have nothing to do with scanning different tones surfaces , if the surface can’t reflect back the pattern projected on the surface due to absorption of the infrared light , nothing will help here but increased exposure of Depth camera sensors .
Structured light scanners use object"s features to create tracking points , without it no one structured light scanner will scan your object no matter how much you going to spend on it , and if you believe that Einstar will magically do the job then you are here very wrong , you going to end with worse issues as you have at this moment. Trust me I used it and the constantly lose of tracking scanning simple full body was dreadful, not to mention the low quality and accuracy .
Einstar power is scanning blobs and cover them with good resolution of texture mapping , it is not scanner for doing any type of reverse engineering scans as the accuracy and precision is far from desired results , but you going to find it out by yourself , sorry for the spolier .
Scanning require patience and dedication , that why it is one of the most expensive services in 3D . You don’t find any scanner on a market no matter you spent $1K or $27K to do the job without any effort , preparation and proper planning , practice and little knowledge .
Photogrammetry is total different technology , far away from Structured Light technology and both require totally different approach .
After 25 years of experience in this field using most all available scanning technology I will tell you Revopoint got their device right , the new RS5 software need more polishing since it is a total new creation and nothing is set yet in a stone .
I suggest that you don’t give up on it too early , it is worthy your time to learn more about and find your own workflow in the process as many others before you that got frustrated .
Once you master it , you will never want to switch back . I know I don’t
And remember that all opinions are my own based on my own experiences as I am not paid here to say nice words about to just convince you otherwise.
I know you have your own experiences, but give it another try , at the end it will be worthy.
But as always whatever you decide it is up to you .
1 Like
I would rather the Revopoint Range lose tracking and stop and find tracking than go off to its own making a mess with the pre-existing scan (I understand I can undo it, but for the money, I should not have to, this should be detected and stopped… I understand it will stop when losing tracking but in some cases, it won’t and then keeps going off on a new tangent. I have seen the EinStar put out some very detailed/ usable scans, and I have done my research on that device. The build quality is also better than this thin piece of plastic (with no lens protection). I have used FlexScan, Metashape, and David-3d in the past, I get it.
Then why don’t I hear about these issues? Also, I am not looking for a “magical” device I am looking for a finished one (one made out of a rigid body would be nice). I understand there is an “art” to scanning, but that is not my issue.
They why are top professionals using the device, and why do they have a whole write-up on accuracy on their website with a good explanation? Are you saying a company that is making ~18-30k scanners doesn’t have the RND to produce a product with accuracy, I can’t believe that.
Can you prove this to me? I have seen video after video and images by renowned people using the scanner without the use of texture scanning with the same results. Here is one…
Revopoint Range looks about the same, but with issues.
And here as it lost tracking at the air filter box…
… But it has everything to do with how it picks up the scanned data from the sensor and interprets that data. I have noticed in preview mode it shows you more data than what is being captured/used when you press play, from the sensors so why not take advantage of that? Also, the auto exposure is nice but I have noticed it gets in the way of tracking as it jumps from one exposed setting to another. It seems you are trying to justify its shortcoming. If this is the new and awe-inspiring technology that you are making it out to be and it’s far better than the EinStar (a product by a company that seems to be in the hi-end of the market and understands a need to fill the low-end market, then I want my experience to depict that (A well built, working device that will last), and it’s not. The USB A connection is a connection known to fail that’s why they put a screw connection on it. A far better way is to use a high-quality aviation connection which the EinStar has. I want to love the Range its ticks a lot of boxes but reliability is important to me whether you are spending $300, $779, $1000.
Not a lot of issues in this video…
https://www.youtube.com/watch?v=cyvYKPtOTa4&t=150s
They may be your opinion but I have seen your name in their videos and you reply to these messages quite often and quickly. I can’t trust this statement (would you?) that this is entirely true.
I had an Einstar and couldn’t get anything even remotely usable out of it. I’m quite familiar with Einscan software due to owning an Einscan-SE and having access to an HX as well as others.
It may have been my lack of PC specs though. Einstar requires an 11th Gen i7 as a minimum. I have an 8th Gen i9 so technically not good enough. I wasn’t willing to drop a grand on a new PC for it so I moved it on.
The other main limitation of Einstar is that is needs a fairly beefy 12V PSU to run. This completely precludes any kind of mobile scanning unless you plan to carry an inverter around with you.
1 Like
Hi, Rilot
I think this may depend on your use case.
I probably will stick with the Range in the hope that the Software gets better, which I see they are putting money where their mouth is. The scans are sharper on the Revopoint I will agree with that. I just am banking on a company that is new, and that is scary. So I hope they make this something dependable and I hope the hardware gets some love on the rigidity factor. If they made this built better; with an anodized aluminum body I would spend $300 more on it, It just feels like plastic.
This defiantly is an issue, I already have 3 computers I don’t need more. That’s one reason I like the Revopoints on-board processing (it needs a fan). I would love to see the same display feature that the EinStar software has with Green and Red gradation with point cloud “accuracy!?”
Hi @Jayson
I am not going to debate or argue about what is better for you or not . It is your own decision , since you already don’t trust my words it is just pointless.
I am community volunteer not a sale person and my time is very valuable and limited . I am here to give a hand to people that want to be helped .
To give you a feedback on your scans , most of the parts did not scanned due to “shadow” ( other parts blocking access where the pattern can’t reach the area ) other issue will be oily surfaces or objects that are too small in volume or too deep .
Normally you should scan partial scans to reach other areas and merge them later .
Also the scanning frame rate will slow down if most of the FOV is scanned , it is a lot of data to be processed .
I will tag in @Johnathan , he is specialized in scanning automobile parts and know better about scanning of this kind of object from own experience in this field .
I will also move this topic into separate thread as it is not right place for doing it , also discussions about other company scanners is against the forum TOS so please keep it to minimum nesesery.
Have a good day
I am not trying to reach those areas I fully understand the “shadow effect” that the projected IR projectors project, it is no different than Strucuered-light.
Then make/buy a processor that is capable of doing more processing… People will pay for a well-rounded good product. I understand that merging is a feature used for these proposes, but this has nothing to do with what the issues regarding tracking not stopping when it over/undercorrects.
No need, I understand that you need to have a matte finish to get better scans. I also understand that you can use markers and a trick is to use markers on top of blue tape for easy removal. I also understand that markers will give you the best results regarding tracking, even if you are using feature tracking. This stuff is not rocket science. You are projecting a light pattern with a grid of lasers, and white or IR light strips, to create contrast for the points to be easily captured.
Now that’s transparency!
Well, we should, that’s how you get a better product is debating. If we don’t talk about these things, things won’t change. Hell, in the app it asks for feedback so it seems important, don’t you think? You also didn’t reply to any of the issues I was talking about, why?
@Jayson You can post your feedbacks and suggestions in the proper thread .
I don’t have the time on debates especially if you throwing in another devices that you don’t even used as your argument regarding what is better just to trigger the reaction .
This is a Revopoint community for helping users in their progress while using Revopoint products, but since you know everything already and don’t need any help
I don’t see anything else here that I can do for you .
For any other technical issues please contact Customer Support at customer@revopoint3d.com
What are you talking about? These devices are clearly in the same field (a low-cost scanner) These are simple questions. Just say you don’t know or have the answers to, instead of demonizing my questions because they don’t fit your narrative, that will suffice.
~fin