Free 3D Photogrammetry with MacOS Monterey on M1 + Intel chips

Apple just released MacOS Monterey and on M1 chip macs users can now use apples new build in object capture.

It works impressively fast on my M1 macMini and delivers pretty fantastic results.

Online I read that it works also on Intel CPUs but might not be as fast as on the M1 chip.
Maybe some of the special cpu unites like neural engine etc could accelerate the photogrammetry algorithm compared to the Intel CPU.

Here is a great summery of the workflow and also how to setup the scene.

I did some test too and compare it with AgiSoft Metashape. I will post those in tonight. Is a free UI for the Apple API.

Very impressive comparison

macOS API RAW does not have mesh dense as Metashape High but in terms of details very close in some instances better.

What quite surprised me is that the M1 macMini takes 12 min while the macPro with 2 GTX 1070Ti and CUDA takes 18 min …

That is insanely impressive!

Here is the Blend File:


You can also use Preview to view your USD files; just click on the file and boom the entire scene in the preview window that you can navigate in.

1 Like

Ja I saw that - preview is much better now - amazing!

To bad, it cannot export / convert 3D files …


Here is a top and bottom pass- software auto-stiched it together.

AgiSoft Metashape cannot use the image sequences with the show rotated.
This, even more, makes the macOS API a killer tool.

Screenshot below: iPhone 12 Pro max 12 MP images.

Blender file is here:


Nice. Not bad at all.
Did you use the iPhone LiDAR or photos?

The LiDAR is yunk - fine for room scans and that is it.

This are normal photos iPhone 12 Pro Max.

With a 24 MP Canon the level of details is obviously even better.
But I noticed one can even upscale images to get the mesh detail.
Pixelmator Pro has a good ML upscale function.


This is just ridiculous

141 Images
24 MP
3 passes

Blend File :


These look great nice results.

1 Like

The results are astonishing for a build-in feature.

I cannot understand why Apple doesnt play this more and offers a nice UI.
PhotoCatch does the job and is a simple app.

I noticed where Metashape and Trnio fail Apples API just does it.
Mind blowing.

What Photocatch cannot do is deal with 48 MP images.
The Win10 macPro with two GTX 1070Ti shuts off lol

1 Like

yeah it is a shame that they did not make their own ui.

I have to try it sometime again, last I did it was on beta one of Monterey.

even thin bodies work well

canon 24 MP images
two passes


blend file :


Here is a good case for Metashape : It can deal with large images

10 Images
24 MP images ML Upscaled to 48 MP

500 MB scan

Insane Detail level - this is fricken nuts !

The last scan for today, I did a pass

Canon 24 MP
Multiple Passes including flipping chair
Blend File:

I am impressed with how flawless the system works, also creating usable scans from thin injection molded parts!


Great cekuhnen

How many photos for the shovel results?

Does it work with movies?

I guess Apple still don’t have their heads in 3D, they should promote this much more.

1 Like

Dust pan is 25 images top 25 images bottom part.

You can use a movie too - the app extracts the frames - but I find it easier to make photos and plan / carefully execute my detail shots

Plus a movie does not work if you want to shoot 24MP image format :wink:


Can you upscale images to get more mesh in the scan?
Yes one can ! And the gained fabric details are astonishing!
I did not see any difference between normal upscaling and ML upscaling.

BTW the 24 MP images upscaled were 174.96MP 16200*10800 pixels :wink:

Blend File:


I did a little writeup with some more examples and results compared to some other software I use. Posted in on reddit

1 Like

what about transparent and reflective objects?

you need to make the solid - non reflective

dry shampoo spray is great !

1 Like

it is impressive that apples API delivers quality like with reality capture or metashape

what I love the most is the ability to flip objects photograph them from different angles and the API and stitches this together into one scan !!!

no more manual alignment of scans and seams …