A while back when I was studying at AIE Sydney I did a photogrammetry shoot of an Apple.

I prepped everything as I would if it was going in my portfolio, but really it's just an apple. So it can go here for now until I have enough food examples for something worthy. I shot about 80 photos with a Canon 77D then ran it through Reality Capture, then the standard realtime pipeline process.

Realtime View

The scan results were pretty clean, I didn't really have to do any cleanup work.

Getting clean results like this is just practice and experimentation until you figure out the most ideal process. But having the right equipment and using the right settings makes a huge difference between a clean result and a noisy or unusable one. Thankfully objects like this with alot of surface detail are an easy scan compared to say, A blank plastic apple.

Reality Capture, 3d Zephyr and Metashape results compared.

Im beginning to think Zeph is going to be a cleaner result for things like action figures and hard surface at least.

The photo consistence filter in Zephyr seems to be similar to clay polish in zbrush.it feels like it does some kind of denoise/smoothing.
The downside of course, is because its a much lower density than RC, along with the filter.. its not as sharp in fine details.

RC on the left, Zephyr on the right

Metashape thrown into the mix

I tried Metashape and found it fell somewhere in between RC and Zeph.
Additionally, the resolution difference is interesting.

RC just throws 22mill at it to get its level of detail with point cloud method.

Metashape on the other hand I did both point cloud and depth map and found depth map basically the same but had less garbage mesh floating around, came out at 3.3mill.

Zephyer was the smallest even after a x2 division in the photo consistence filter at 1.5ish mill.
Zeph falls behind just in regard to pure resolution for surface detail, the mesh just isn't dense enough and this was on max settings I think.

RC left / Metashape middle / Zephyer right.

Here's a closer compare of metashape (left) vs RC (right)

My overall thoughts of the processing of all 3

RC seems to be the quickest out of the three.

Zephyer however did feel like it was reasonably fast and basically around the same.

Metashape however is damn painful for its processing time and now I remember why i switched to RC from it, in fact I couldn't even process in its max presets as it would run out of memory. I had to take it back to high. even then, 7.5 hrs for depth map processing.It seriously has problems when it comes to memory requirements by comparison to the other 2. I have 32gb ram with an RTX 3070 and I couldn't fully utilize the highest presets because it would crash.

Id say if Zephyer could produce a far denser mesh result like in the 10mill mark it could be closer to being on par with RC regarding sharpness of details. Id also like to be able to process the mesh without any filters at all to get a true result.

3D print, now in Realtime

So several years ago I was heavily into 3d printing and upon finding a don post statue scan, I decided to build a symmetrical working helmet replica with most components modeled in fusion 360, with the front and back face shells re-worked in Zbrush to my head scale.

There's still a few fixes to do, I realized mirroring half the front and back head shells isn't ideal if I want to texture it well, as the grunge pattern is making it quite obvious. The texture was really just a quick throw together so Id like to do that again.

Setting up the refraction and lights for the eyes was interesting. Unfortunately Toolbag still has a limitation on how many lights can be used. So they didn't render.

Realtime viewer

Exploded view

Realtime viewer

Stills

Disclaimer: A couple of the components are reworked from others found within the C3PO builder community. (the eye back plate and lens predominantly as it was a design improvement to help see better) However, even my own 360 models along with everything else had to be retopo'd and virtually remodeled to work for both low poly and subdivision. Fusion exports everything in Tri's.

The main purpose of this exercise was to display a real-time version of my 3D printing work.

So, over the weekend I decided to pull out my old box of decades old illustrations packed away I did at 8 - 23 years old.

While I've been a digital designer and now a 3d artist, my foundations started in illustration and character design.

These are just popping up now, because I took photos of all my old illustrations recently, about 700 from ages 12-23 and I finally have digital copies instead of a big box of paper packed away for decades. So now I can sift through my work and cherry-pick what could be made better and its actually motivating me to draw again.

Dinosaurbirdthing

Something I did as a kid at 16 on the left, I decided to refine the body and use another head I did around that time that was more bird-like. Didn't use reference, so anatomically probably still wrong, particularly the legs.

E.L.V.I.S

Another one of my archived character designs I did in the late 90s on the left. Never felt right because my proportions/posing sucked, I never knew to use reference back then either. Feels good to refine my old work, Particularly since I've not drawn much in the past 10 years.