I now tried to find a simple test setup to illustrate the issue. I’m aware that reflective scanning spray will greatly help here, but my goal is to show that this is a useful feature request, meaning that pausing the turntable while capturing the frame would improve quality.
Here’s my test set-up. The Revoscan MINI 2 is pointing down on the large turntable with the turtle on it. RevoScan 5 settings are high quality, feature tracking, dark object mode.
To make sure I don’t accidentally measure any tracking issues, I then captured a single frame while the turntable is standing still. As expected, the result is good:
I then turned on the turntable, waited for the turtle to be in roughly the same position, and captured a single frame while the turntable continues to rotate:
The result is that the point cloud is super noisy and to me it looks smeared in the direction that the turntable was moving in.
My conclusion based on this test is that if the turntable rotates while the camera is capturing, that degrades the quality of the resulting point cloud. And that means if I can automate the process of rotate, pause, capture frame, resume, rotate, pause, capture, … then it will produce better results than if I let the turntable continuously rotate and use the continuous capture mode.
The Single shot function is to capture steady objects not in motion , so of course you will get bad results . The turntable originally never was intended to use with this function as it just arrived shortly long after release if Dual Axis Turntable , so yes we need that new function to be covered.
Use Lazy Susan, the best cheapest alternative until Dual Axis Turntable have official single step function available , or use the Android app as I mentioned before .
I use a cheap black lazy Susan, not all objects can be scanned automatic due to lack of features or complex dimensional geometry , especially if you deal with smaller objects where you need to keep excellent distance, and it performance isgreat .
I would certaily welcome the ability to synchronize a turntable start/stop and single shots. My experience is that single shots are more accurate. From signal processing and physics point of view it would seem obvious that you will get better results if the frames taken are from an immovable object. Single shots are slower to take, but might win overall as the post-processing is faster, with less to fix manually. As a bonus, you get more detils in the scan.
So, I firmly believe in single shots for now, but I am open to be proven wrong. We have now seen examples of excellent results with single shots, so I would think that what remains to be seen is the objective comparison of single-shots vs continuous mode. For example, do we have a similar scan of a lego block, but done with the continuous mode?
