Thanks

@@user-wy3ng5hs3i RC is king for detail and speed.

@@user-wy3ng5hs3i Rc is worst for estimating camera poses though

Very good one, also you can use polarizing light. Also instead of baby powder you can use something like watercolors dispersed with a toothbrush for example. Anything goes really as long as it is uneven and adds contrast, mud, spray paint etc.

William faucher released one as well today !!!

@@bkentffichter I did, I literally have an over three year old video on the subject as well 😄

people on youtube get so pumped about commercials. we should be critical and skeptical. like show me a livestream of a complete process.

Bryan Mitchell huh? Reality Capture didn’t pay me anything, nor are they a sponsor of my video in any sense. I paid for all my scans myself. I did get a license from 3DF Zephyr to their software, but they also had no insight or oversight into my video 🤷‍♂️ Before you question someone’s objectivity it helps to do your research first 😉

v b read what I wrote to Bryan above, I can’t speak for Clint, but my video is not a “commercial” and all I said in my video are my own opinions, nothing anyone has told me to say. Question objectivity all you want, but questioning things just because you have trust issues, and not actually valid concerns, seems like a good way to go through life paranoid and unhappy 🤷‍♂️

👓👓🏧🏧🇨🇿🇨🇿🇨🇿🇨🇿

Thats a perfect comparison!

Absolutely true.

@@pwnisher Hi Clint! one thing that i don't really understand is that some videos forbid you of moving the object you are trying to capture. however you turned your one wheel upside down and sideways. Does this have any negatice influence when the pictures get ported into realitycapture? this is my first time ever doing this and i just want to make sure i am doing it right :)

Feel free to write me too.

why are ppl so pumped about this vid . it’s a commercial and of course we don’t see the mess.

@@MrMadvillan you k?

That means so much! Thank you!

Heeeeck yeah!!

I believe in you!

Many thanks for the kind words and support! I’m glad your enjoying the training :)

I'm wondering the same thing

Hopefully I'm not too late to this, but what I do is to bake a high to low normal map also, and do a basic roughness map using a colorramp to roughness in Blender. That way, when I bring the model into Painter I have 3 basic maps. I'll then tweak them like you said, giving each material it's properties using tons of masks. It can be a little laborious, but the results are amazing.

that really depends, for the most part this is much faster. Modeling the onewheel from scratch would take 6+ hours while this seems like it took around 1 minus the rendering process

Gotchu my guy :)

Hey Thomas - absolutely you can do this. I've captured some interiors with incredible detail with photo-only scans. That said, if you can ALSO capture with a decent lidar scanner (and be sure to have it grab textures) you can combine both datasets and get the best of both. Photos alone can go a long way though but you will need to be disciplined in your capture - much more involved than walking around an object or turntable. Also likely will want to utilize a tripod and choose exposure carefully - multiple brackets if needed to capture full range, etc. You can run tests in RC for free so I'd suggest start by trying and seeing what you get. Your suggested approach is more or less what I do.. Circle around the room, back to the wall facing in, different heights, orbit around things in the room at varying heights, make sure you have a camera 'path' to go from wide to close up and try to think of a giving a lot of overlap and path for solver to join all the photos. If you have a big jump in position make sure you have a lot of overlap with other 'photo paths' so it can solve. Avoiding having to connect the dots later will make it all work much better. Try to get at least 3 or so angles on any feature you want, so if moving accross beamed ceiling make try to visualize every 'face' having at least a couple photos of coverage. Good Luck!

you can do it without repo directly in RC by simplifying your model and reprojecting textures and normal maps.

@@danieljackson182 surely you'd have a much more efficient poly count by retopologising though? or is RC really that good?

I've managed to scan 5 objects in under one hour on a cake spinner. Then processing varies on what technique you use, your computer and also image resolution. For a model with udims and normal maps it would take on average 3-5 hours.

Yeah it just depends on number of pictures you take and computer speed. My scans take me a couple hrs tops right now.

@@12Konst I straight pay for a Vagon Galaxy for large scale clients. It absolutely works.