Tests, advices and tips

Hey Guys!

I will tell you some of my experiences, tests, and tricks that I am applying for scanning.I took two weeks conducting scans of different types of models of different sizes, and I will share my experiences.

To begin, I use the tripod for free scanning and standard mode scanning with turntable. Is very flexible, and doing a few tests, you can scan objects of respectable sizes, using the automatic mode.

The light is very important. If the room has plenty of natural light, the calibrations will not work, or will scans misaligned.

One of my first models was a dragon with a broken claw.

Attachment 6689

The figure is complex and has dark colored varieties. I added three standard scans at different angles. To scan areas in shadow, I made 3dprintings pieces that help me support models at different angles (wedges and supports) .The result is surprising and very fast compared with other scanners (Matter & Form). Add other examples of scans, such as the figure of a virgin, which exceeds scan size "standard", but to use the tripod, I got a range that was able to catch all the model, and it was automatic . The figure was matte white. The result is spectacular.
Attachment 6690

After these examples, I scanned a lion marble protruding from the turntable, but add 4 scans, not only with different angles, but also with protruding parts of the turntable drive. The result was excellent, but with some noise caused by the material, so I had to do some retouching in Zbrush.
Attachment 6691

The hardest to scan parts are metals and dark colors. Even materials that do not seem problematic, as had the lion, can cause noise and loss of details.For avoid this, users are using talc to solve the problems, but in metals talc is not permanent, it is not fixed. To create a coat of matte white, I tested the liquid vinyl, and it works great! The liquid vinyl is used to paint plastic parts for cars, or thumbnails, and the biggest advantage is that ... removes easily with your fingers. For this to be done, you have to give at least two or three coats of paint as if it only gives the layer is very thin and breaks when removed, and is much more work. The layers are very thin, and no detail is lost. I performed two tests, one with the lion..and the result is very soft, and recovers details in the mane.
Attachment 6692

Finally, I attached a base of a collectible figure, which I painted with vinyl and is relatively large ... the same size as the turntable. Only I applied a layer of vinyl only, so separate it is becoming more difficult ... with a thicker layer would have been better to remove.
Regardless, this layer is sufficient to achieve high detail in scanning, detecting very fine grooves on the base. Besides ... this scan was performed in free mode. Only 3 scans were sufficient to achieve the end result, then retouched in Zbrush.

Attachment 6693



I hope my advice and my experience will serve to enhance your scans, and encourage Shinig3d to improve the software, as there is much room for improvement, as the Cancel button, point cloud export ... and .. ¿Colour? : D: D


Keep in touch!

Nice pictures and tips, thanks! I hope to test mu scanner soon, did anybody hear if batch 3 is underway?
I see you use Zbrush, I now use MeshMixer and can recommend it to people to use for editing mesh files.

good tips.

Thank you so much for your tips! We will work hard.

Quote:

---

Originally Posted by ariakas

Hey Guys!

The light is very important. If the room has plenty of natural light, the calibrations will not work, or will scans misaligned.


Keep in touch!

---

This was a major breakthrough for me! I was about to pitch the whole thing in the garbage, scanning was so frustrating... Turned off my lights in the office and closed the blinds, now all of a sudden scans come out actually aligned and the software does not crash (as often).

Still trying to figure out how to add additional scans on the turn table. If I take my model and continue the scanning with another round but taking the figurine (for instance) and laying it flat, it will come out merged. if only the software would remove the scanning table from the mesh, I think it would actually resolve. Any ideas there and how did you add more scans to your project?

Quote:

---

Originally Posted by RonnieW

... I was about to pitch the whole thing in the garbage, scanning was so frustrating...

---

Wow! You sure do get discouraged quickly!

If you get a few minutes, you should read through all the threads on this forum, I think you will find that this scanner is very capable of doing more than the Shining3D says, not less. (Which is surprising in the 3D scanning world)

Free scan mode is where it is at with this scanner, with a properly setup "lazy susan" bearing you can turn the model yourself at any angle you want and take as many scans as you want, as long as that winds up being less than 200, unless you have an amazing computer.

I cannot imagine a model that would be best suited to the automatic turn table. Maybe a baseball? The turntable is great for calibration, then I put it away.

The is the back of a transmission I scanned a couple days ago, by tweaking my approach I was able to scan the entire transmission in a single session. Now if only we could save as we were scanning... :rolleyes:

Attachment 6929

Dave @ Nerv

hey guys,please remeber the scanner is optical device, so the light will obviously affect the scanning.and there is a preview window on the left under the turntable mode and calibrating mode.with tripod mode, there will be two there.you should look at it before you do anything, please.thank you.

Quote:

---

Originally Posted by Alfred_technical_support

hey guys,please remeber the scanner is optical device, so the light will obviously affect the scanning.and there is a preview window on the left under the turntable mode and calibrating mode.with tripod mode, there will be two there.you should look at it before you do anything, please.thank you.

---

..and maybe to add to this advice, if you see red areas on your object it means it's too bright or overexposed. Use a lower setting in the software or reduce the light.

P.S: great scan Dave ! :cool:

That is a great scan Dave, I was going to drive a lazy susan using the existing turntable, by adapting it. Now after seeing this scan I probably won't bother.
That's exactly what I want to do, and all that remains is to get it into CAD.
What size was that scan Dave?...I just did a small torsion bar tube for one of my engines, and it ended up being 66MB as a STL file. :eek:

Thanks for the feedback guys.

That scan, 1.2 GB. A properly setup quadric edge collapse decimation algorithm can knock it down to 1/10 with no visible difference.

This is another scan, the entire transmission in one session: I've seen lower quality scans of transmissions with scanners worth 50x as much as this scanner!

It does take a very specific set of tricks to scan large items like this without any alignment problems and no crashing. Getting to this point has taken a ton of work.

Any surface deviation is actual casting flaws, not scanning data issues.

If you have tried scanning fins with the Einscan-S, you can appreciate this scan that much more :)

Attachment 6957

Wow, this is really great ! :cool:

"quadric edge collapse decimation algorithm"

Man, I need to place that into a conversation once, LOL

Which software you used ?

Another great scan, Dave ... :;)

Quote:

---

If you have tried scanning fins with the Einscan-S,...

---

I've just tried that from the turntable, and results are mixed at the moment, it was Ok until it meshed it into a cube like form!
I've also attached the tripod and trying that from the turntable after calibration. Just playing at the moment.
So these are magnets and matt grey balls on the end of the studs, over holes etc?
Impressed
Mike :)

OK, I know what the "meshed into a cube like form" is, it's the limits of the turntable model :mad:
I'm with you Hugues "quadric edge collapse decimation algorithm" that's a combo and a half, you'd be into total geek mode at a Meshlab party!
LOL

Hey Mike,

I suggest stop using the turntable, I'm sure it has a use (calibration), but not for any serious scanning.

Hughes, the decimation was in Meshlab. SolidWorks has the same method build into it's Scanto3D add-in, but Meshlab is faster. Having said that, SolidWorks can break your mesh up by curvature, and then you can run that algorithm afterward in an effort to isolate and preserve any really sensitive meshes you want untouched (like holes).

Thanks Dave, I will, just killing time waiting for the lazy susan.
I thought I'd try the tripod with turntable for calibration, as I didn't want to keep unscrewing the scan head from the stand.
First one failed, and on lowering the tripod the second one didn't.
Anybody used the tripod for calibration?
Seriously, impressed with your scans, looking hopeful it will be good enough for what I want.
Also, I'm liking this scanner a lot.
Mike

I always use my scanner on my tripod when I calibrate. I just make sure the distance and height is as per the OEM stand and position template position provided.

Lazy susan bearing is dirt cheap at Lee Valley tools if you have anything similar where you are. $10, 1000lb rating.

Wait, i'm lost guys !

I thought we only need to calibrate the scanner once ?! isn't it what the instructions says, I'm sure i read it somewhere...re-calibrate only if the scanner got bumped or something like this.

YOu guys calibrate before each scan ?

No, I just recalibrate when the software gets updated. :)

Sorry Hugues, I'm an ignorant noob, just assumed you had to...if you moved it! LMAO
and Dave, I did exactly the same replicated the positions on the template, but the extra height threw it the first time, even though I tried to land the base of the white projection on the second ring as recommended. I take it, all the scanner does when it calibrates is align the rotational axis wrt to the reference template, and modify it's built in 3D model. Plus free scan is the way to go, after having a go, but just need to work out the alignment references...those matt balls look like a good idea, now i've seen those purple alignment marks in the software.
Alfred, it's a shame you can't tab or resize the three viewing boxes to make it easier to rotate/zoom etc. I kept trying to hold down the mouse and resize them, but a tab option to bring a window full screen whilst you found the alignment point would be useful. Just a 2¢ suggestion :) I'm liking the scans.

You can enlarge the view by Mouse Scroll Wheel.
Add one thing, you can rotate the scanning by press Mouse Left Button and move Mouse.
And move by adding pressing a Ctrl key.

Thanks Alfred_t_s, I wondered if there were any "hotkeys" I have a trackpad on my MacBook so it doesn't quite behave as it should. Also the text in the left hand column is wider than the column, so I don't see all the text, that mentions the mouse controls. I don't think, it's a resolution issue as I have 1920 pixels wide, so maybe you can look at this, in later software revisions. (y)