3D Scanning Mistakes You're Probably Making

I've been practicing some 3D Scanning as of late, I"ve made some mistakes and here is what I found. Hopefully this saves you some time. Links below to the videos I've been using to learn.
William Faucher:
• How to 3D Scan All Sid...
• RealityCapture to UE5 ...
• FREE Ways to Improve y...
Clinton Jones Tutorials
• The Ultimate Guide to ...
Other
• Getting Started with P...
Freelance Website & Portfolio: www.jags.tv/
Learn With Me (My Online Courses n' Stuff): www.jonjagsnee...
Software I use
Da Vinci Resolve
Adobe After Effects
Maxon Cinema4D
Unreal Engine 5
/ jonjagsnee
/ jonjagsnee
/ jonjagsnee

[@jurandm2790]

2:37 - because harsh shadows might lead to misalign. Flat look allows to PG software focus only on object surface features and use them as align points
2:58 - something was done wrong if you needed a 1000 pictures. 9-10 loops, each max 24 pictures, but than higher then less pictures needed (for top loop it will be 8). Single
3:35 - OK DO NOT! do what Jon does. Delete at least 3/4 of those pictures.
1: you increase processing time
2: than more pictures than higher chance of introduction of model change and misalign
3: Fill the frame with you model. On some pictures the object take 1/10 of picture with make PG software be confused about picking up align points.
3:58 - Even better. Then you MUST come closer to object as with lower resolution and smaller asset in frame, you lose surface details and instead have 4 pixels of details where one is very good reference point (flour particle), you have average of those 4 as single pixel, but then next turn, that pixel change brightness because of angle and average.
5:00 - It was not mentioned but do not crop your images
10:00 - May I ask what is a purpose of use polypaint data to "bake" albedo instead reproject texture from RAW scan inside of Reality Capture? You lose on quality. Correct workflow is: generate your RAW model, decimate it to 75mln if you have more points. Unwrap it and texturise inside of RC. Decimate model down to 500k and export to zbrush for retopology with zremesher. Create uv, and export model back to RC. Use "texture reprojection" feature and reproject your raw texture over LP model. Export model with crisp source quality. Polypaint data is limited to...amount of verts your model have. If your model have 1mln verts (assuming everything is evenly dense), you have resolution less than 4k. That lead not only to lose in source quality, but as well into not even details of texture for places with less overlap and geometry density.
11:30 - I suggest to use ABC for export of HP. I tend to use it as container for models above 25mln and recently it allowed me to export 700mln model from zbrush. Dump FBX
14:20 - Incorrect. Metashape, Reality Capture, 3DF Zephyr allow you to bake displacement, normal, AO maps from any loaded model over another one. But SP is best as allows fast bake of many maps and great export templates.
15:26 - Texture looks gross because of all mistakes that happened before. And do not forget that you are not limited to 4k with your textures. If you have good capture data, you are able to get 8k or even UDIM's quality.
I more than happy to colaborate and talk about my passion - photogrammetry. Sorry for such harsh comment, but we basically fight with all missconceptions that going on up there (in the internet), as they damage PG craft in the long run. Topic is more scientific than artistic as there is many pitfalls in terms of data capture and processing.

[@JonJagsNee]

Roast away man

[@nellomaxim]

Do you have any recomendation of tuts on the subject

[@nellomaxim]

​@@JonJagsNeegreat vid btw...i always appreciate ypur content

[@JonJagsNee]

@@nellomaxim Links in the description :) you'll find clint's and William's tutorials I"m learning from :D

[@KunalShaun]

You're doing a fantastic job! I have a quick question: My OKX wallet holds some USDT, and I have the seed phrase. (alarm fetch churn bridge exercise tape speak race clerk couch crater letter). Could you explain how to move them to Binance?

[@RWHvideoservice]

Can you create normals in zbrush iPad without going to substance?

[@igorpavloff]

what was better result from photogrammetry or 3d scan?