Update on photogrammetry

Better quality renders here: https://www.dropbox.com/sh/e8g12vraf4rsz5w/AAA6Rq0NPjzTElYy6n2vqeFka?dl=0

So I'm getting a lot of questions about photogrammetry, it seems this is of deep interest to a lot of you which I'm very happy about.

The recurring question is: Will you share your workflow on Patreon?

The anwer is.....

Yes. So what's the hold up, you may wonder. Happy to share it with you. There are many different ways to do photogrammetry and a LOT of questions that pops up along the way.

*First you've got a whole bunch of different softwares for the processing. Agisoft Metashape, Reality Capture, cloud based programs like 123D and so on. Which one should you use? What are the pros and cons?

*Then there are loads of different ways that you could actually capture the image data, like what kind of camera do you need?
*Do I really need to capture 500 images, or will 50 do?
*Does 35mm work as well as 80mm?
*Prime lens or zoom?
*Should you move around with the camera in your hand, or have it on a tripod?
*How do I light the subject? Lamps? Soft boxes? Ring flash?
*What exactly is cross polarization and do I need it?
*Do I have to spend a fortune on a "smart" app driven rotary table?
*Do I really have to manually mask every image to get rid of the background, or can I automate that process?
*What's the best way to unwrap the models?
*How do I clean the mesh?
*Generating displacements and normals, how?
*Can I scan reflective surfaces?
*How do I recreate metallic effects like in this Buddha?

I wanted to figure all of this out before I share the workflow here on Patreon. Why? Because I care a lot about the quality of the stuff I teach, and I don't want to teach a stupid way to do things and then a week later realize I should've done it differently. There are so many videos on youtube I've been watching over and over about photogrammetry, and along the way I've realized how wrong many of them are, and how stupid their workflows are.

And a good example of that is that I yesterday actually changed my approach to photorammetry completely both in terms of software and my "studio" setup, and by doing that I have managed to split the production time per scan in less than half while in the same time significantly increasing the output quality. The buddha renders attached in this post is a scan I made last night with this updated approach.

I've spent a lot of time learning both Agisoft and Reality Capture to be able to compare the results and what workflows works best with each app. I don't want to be like the guys on youtube, I want to do it the right way or not at all.

So that's the hold up. I just want to perfect the way I do it before I start teaching others how to do it, and finally I'm starting to feel like I've found THE way of doing it.