If you good with Lightroom or DxO i strongly recommend forgot about JPGs and use only RAWs. Even from iPhone RAW files have a bit wider dynamic range, and this mean you can extract more details from shadows. But noise from smartphone sensor is a big problem. on JPEGs you see only good result of system/hardware image optimisations. But with RAW you’ll see how much in reality have great but small smartphone sensor.
BTW, ShadowTail.
Looks like on both scans you have bad coverage on object head. On second image because of this you have strongly visible misalinment.
Götz Echtenacher nice captures!
The darker tealight I did not get the top of the head. It was more or less just a test setup.
On both images you can see the seams of the mold the figurines were made with on the side of the head. It has not been properly cleaned up.
Also, for the hair parts the camera was ever so slightly out of focus.
But there was actually no misalignment there at all.
Here’s another project I did over the weekend:
193 photos, 35M tris decimated to just below 50k.
drg_pick_nose.jpg
There are still a few flaws in that mesh due to some hard to photograph areas. I may have to add some more photos focusing on those. Overall that is the result I got without cleaning up the reconstructed mesh.
RC just does its magic and reliably delivers what all the competition fails at 
Over and over again.
Ha. So that was just a plastic replica. 
This heads was always gave me pain on outdoor scans
https://www.artstation.com/artwork/EggL0
Thanks Vladlen!
Nothing compared to yours though!
But I’ll get there…
Nice work with the thumb too!
What gives you "head"aches? The intricate detail?
Thanks for sharing btw.
ShadowTail, also nice one!
Yeh, the nooks and crannies are the worst!
About the suspected misalignment:
With your buddha the scale can be really confusing! 12 cm is tiny!!!
Vladlen, nice retopo - is that zbrush?
How much time did you spend on that?
Also, which mapping method did you use for the detail?
That’s really not bad - I thought it was high res at first.
About the “orange” skin texture:
I am sure in most places it is the real surface (compared with texture) but in others it seems like artifacts.
This is something I am struggling with quite a bit and since I am interested in stone tooling it is quite vital for me.
Is that the limit of the technology? I am expecting better results with RAW, but since you already do that, I was wondering…
Hi Götz.
"head"aches:
Repeatable pattern, small details and as result any small error (low coverage, blur, etc) - problem with alignment that really hard to fix especially with manual control points (sometime hard to understand this this dot on image is that on other… even as we smarter than any Ai)
Regarding zBrush:
I hate if people use zBrush automatic quads retopology when not needed, and this is 99% of all scans. (hate more only InstantMeshes retopo).
For static solid organic shaped objects quads give usually more errors and produce meshes with higher poly counts.
I Use quadric edge error decimation in MeshLab that produce low poly mesh that work much better with combination with baked from high resolution scan tangent space normals.
And 3DCoat (UV Path tool) for manual UV unwrap. Because this tool allow broke addiction to quads that have better edge loops.
And last time using Marmoset Toolbag for bake normals/color and Ambient occlusion from high res mesh to low poly.
Vladlen Great Job!!!
I don’t like zBrush. I prefer Autodesk products. Definitely I agree with you and I recommend RAWs than JPG. You don’t need Lightroom, RAWTherapee is good enough as Blender and for free:] You can bake AO or normal maps in Blender as well. Blender has good retopology and UV tools as well.
https://blendermarket.com/products/retopoflow/
Ok this is external plugin but much cheaper than zBrush, Mudbox or Maya. In my opinion, maybe I am wrong, but it is much easier to create low poly meshes objects like yours or ShadowTail than gothic or renaissance buildings with many details.
Once Again Great Job:]
Vladlen
Well, the headache was worth it though!
The head looks really good - very crisp!
Oh-oh. That baking-workflow I am happy to leave it for other people… 
Robert
What Autodesk product do you use, mesh-mixer?
And did you use the RAW-developer on Affinity Photo yet?
Götz
Yes I’ve tried. Even I thinking buy license. So Affinity is good idea. I’m trying to reduce costs. But now I’m using Lightroom + Photoshop, Mudbox, Maya and Blender. RAWT only when I need to process RAWs from LX100. I never tried mesh-mixer.
Here is another good example for accuracy and detail fresh from the processor.
The railing andsteps are only 1 cm thick.
Much better than I would have hoped for and only at Normal!
With my 10 mpx JPEGs from a 1/2,3" sensor…
One more Reconstruction I have for you:
26 photos taken with my Galaxy S7 phone with the camera locked to ISO 100.
The reconstruction was 8.4 million tris, this render is from the simplified mesh (480k tris).
The quality is surprisingly good.
metal_structure.jpg
Cheers.
Indeed it is not!
Would you mind uploading a wireframe?
How big is the model in real life? Like half a meter wide?
Did you have any problems with alignment and did you have to add CPs?
Get the source files and project here:
https://dl.dropboxusercontent.com/u/255 … ucture.zip
The alignment went through without problems. No CPs needed at all.
The main beams are about a meter or so apart.
Robert,
What flight mission planning software did you use for this project?
Robert,
What flight mission planning software did you use for this project?