Probably the most straightforward tutorial for this workflow out there. I run a studio and don't really need new staff getting into the nitty-gritty of ZBrush, as this is literally all it is used for in-house. Thank you very much, this video is a great resource and a lot of companies still fumble when it comes to photogrammetry optimization.
Like you, most of what I scan is used for inhouse, so the only reason I learnt ZBrush was to clean up a few bits, use the ZRemesher, and project details back. I never dove into the know how of texture and normal projection as I use either Blender or Reality Capture for that, but this tutorial may just have saved me a few extra steps for the diffuse projection. Thanks for the info.
Im not 100% sure. But if all you use Zbrush for is the Zremesher, Blender has Quad Remesher, which is made by the same persoon who made Zremesher. Idk if it has the guideline feature or not though, but its worthl looking into id you wanna pay like, nothing for it.
the only fast way is to produce them in substance by using the other maps. It might not be accurate but it definitely beats shooting with and without cross polarization!
Ho, thanks for your awesome tutorial, I have a question, why did not you project the details from photogrametry model directly on to the final unwraped low poly model? Thanks in advance
Great work, and clear presentation, thank you! Curious if you just dumped the unoptimized high res scan into UE5 if it’s Nanite technology would take care of all of this manual optimization.
Yes for opaque objects like a tree log you don't need to optimize the scan, you simply import the high res mesh's (less than 10mio polys is better for the import not to crash) and it works directly! You just need to create roughness map. You can see examples of wood debris and gravel scans on my latest video test using UE5
The model that you apply the texture to (source object) and the object you're going to project the texture to (target object), have to have a dense mesh in order for the polypaint texture to be high res. If your source object has a low res mesh the texture applied to it will be low res too. As a result even if your target object is high res it will still have a low res texture.
Hi there, I love your tutorial, very clear and concise. I'm just a bit unclear on the elements between the jumps between Zbrush, and Cinema 4D. For example the filing system and what we're exporting/importing, when and how? For example when you say we take it into Cinema 4D, how do we do that? And what file type did we bring into Zbrush originally? What's the export process from Zbrush, and the import process into Cinema 4D? Also, I'm not very experienced in Zbrush but I'm a bit mistified regarding how we're making subtools, and how we're getting the high res subtools and overlaying it perfectly over something we drew in (the original mesh we drew into the viewport) my apologies if this is confusing!
Hi there. Glad you enjoyed the tutorial. Here's how the process goes. Keep in mind that the first step for you is not necessary since most software doesn't export in USDZ 1) import the USDZ scan in to Cinema 2) strip it out of any textures and export as obj 3) import the obj in to ZBrush 4) Retopologize and export as OBJ 5) import that obj in to Cinema and create UV. Export again as obj 6) import obj in to ZBrush and import the high res mesh for the projection to happen Since everything is the same size (low res mesh and high res mesh) there's nothing you need to do when adding them as sub-tools. The size of the object in the viewport doesn't impact the size of the sub-tools. Hope that clears things up!
@@marvelousdecay hi there, another question I'm afraid! Is there an automatic way to place the high res mesh subtool atop the low res subdivided tool? I find myself having to try to carefully place them on each other myself which ofcourse can lead to human error, is there an automatic placement method you use? many thanks xx
Hi Dimitri! Great and thorough tutorial but I had a question for you. What's the reason behind going to C4D to generate the UVs instead of staying in ZBrush and using the UV Master to generate them? Thanks again!
You could definitely do it all in ZBrush. But I got some really weird results when I tried. I don't have as much control of the result even when I draw the UV guides. It might just be a matter of experimenting more with Zbrush. So if you feel more comfortable doing the UVs in Zbrush definitely do that! No need to get another program for it!
@@marvelousdecay I see. Yeah, I was just wondering if you have experimented with the attract UV brushes Zbrush offers. But yeah, sometimes it just won't give you the result you're going for. Thanks for clarifying that!
Hi Dimitris, A very informative video and essential to my position on understanding the photogrammetry process. I’ve enjoyed your presentation and performance style in all of the videos I’ve watched. I saw a video the other day where the presenter utilized Adobe Lightroom and PS for post processing the image files before sending them into Metashape. I checked your channel today and boom step two in the post process workflow appeared. I do have a question thats been driving me crazy regarding the photogrammetry process for a while now. I’m able to create decent models with my very basic equipment (iPhone SE, LED desk lamps, a backdrop and my repurposed 45 year old Dual 701 turntable @ 10 degree increments) provided that I can fill the frame with the object, the object does have some reflectivity and no post processing applied. Your step two video today may serve to make the models perform all the better by reducing the mesh size that the photogrammetry process creates as well at texture maps. So here’s the question. Is it possible for the photogrammetry process to capture highly reflective and odd shaped objects? Some people say that there are some objects where photogrammetry will not work. In my case it’s musical instrument. A Telecaster body style with a maple neck and a mirror finish. I need to go back and review your video that covers this topic. I know that you polarize both the camera and the light source in the video as well as an example of a plastic doll body with and without polarization. If the elusive high quality model that I crave can be created I would certainly purchase the equipment to achieve the result. As per your request I would certainly use the associate links you list for the information you provide. Thanks One of many attempts to scan the guitar. This one actually did output a somewhat identifiable model. The images have been scaled down for this example only. storage.googleapis.com/photogtree/imgAck01.mp4 Here’s a few examples of models I’ve created none of which are 100% but that will be addressed…. sketchfab.com/3d-models/haldex-90054007-neway-900-54-007-medium-res-fbf3f5ca40c44efdbe993efb64a3d651 sketchfab.com/3d-models/haldex-90054007-neway-900-54-007-5f4461ea904c473ebf0928fa2a14c44c sketchfab.com/3d-models/sealco-110380-relay-valve-7d7778c235f948d6b7490a5d5429a09f sketchfab.com/3d-models/worn-running-shoe-6a5536cc8edc49b7a48a1a713ef44706 sketchfab.com/3d-models/big-kitty-a92ab878242945aaa9231a6ddd0da88d sketchfab.com/3d-models/stone-face-09b9978b4c5b48ebbfa3702069be9dd6 sketchfab.com/3d-models/birminghams-bull-ring-a740e7193e524926be432928f8171f01 sketchfab.com/3d-models/kwakiutl-band-totem-test-03-72e3900fce844e50b681a08e17244681
Hi. impressive work, all shown simply and clearly. I'm just wondering how do you bake your diffuse color map with brown background? Mine is all black. I will be grateful for your response. Once again great work!
Hello! Thank you for this amazing walk through. I was hoping to get some help. When I export the displacement map from Zbrush, it is just grey, there doesnt seem to be any detail. Is there something I should be doing differently if that happens compared to this tut? Thank you!
what if you have a photogrammetry model that does not have any UVs? I need to be able to place a texture created from Illustrator that is a costume for a character.
You can use a different projection other than uvw. Cylindrical, frontal etc. The result will vary depending on the complexity of the mesh but you can give it a try.
Hey, I follow all the steps but my diffuse maps come in bad quality. And it also a bit oversaturated when I import on photoshop. What is the issue ? The way is to project diffuse on the high poly original mesh with polypaint, and then subdivide a lot the low res mesh, and project all onto the low res and then export all the maps, that’s it right ? Thanks
Yep that's exactly the procedure. You got it right. I suspect that your original high res mesh might not be as high-res as required for the diffuse map. Keep in mind that polypaint resolution relies on geometry density. So I would suggest to subdivide the original mesh a couple of times and then using polypaint to project the texture. Finally make sure that the subdivided low-res mesh has approximately the same poly count as the mesh with the polypaint.
Great video! Question. Is there a way to create a single texture maps from multiple seperate scans? Say you scan and process two objects separately but want them on the same texture maps. Is this possible?
In theory yeah you could do that. Each part of the object could take a small section of the map. But it would have a significant impact on the texture quality since the 4k or 8k texture would be split into two or more objects.
Hi thanks a lot for the tutorial. For scanning small object I use a rod to put the model on, but how to remove the rod on the final scan with proper texture with any hole? thannks again
@@marvelousdecay thanks so I as used Metashape, you recommand to export the mesh with the rod into Zbrush ? Not removed the rod directly into metashape
@@peadly4328 You could remove the bulk of the rod in Metashape and then do some minor cleanup in Zbrush. Just pick the method that feels more comfortable to you.
Hi Dimitri, I really enjoyed the video and I was wondering if you have ever seen 3D scanning? I have seen people do it with Xbox Kinects and would like to know your opinion on it, I'm personally thinking of giving it a try myself. Keep up the great work!
Hey Manoli. Yeah Kinect could work absolutely fine. The only problem is fidelity. There's not enough resolution but if you already own one definitely give it a try. It's gonna be a nice weekend project!
@@marvelousdecay Thanks for your response, I have seen people use up to three in order to get around this issue, This is escpecially fascinating considering the Kinect is a depth sensing camera . It could also make quite an interesting video comparing the results of it to photogrammetry considering how budget friendly this method is
Blender works for it all. I've used blender since owning a 3d Portrait studio 5 years ago. Between the infrastructure expansion and blender updates it's a great time to start. You've got this!
Blender is open source and extremely capable - substance is extortionately priced when there are alternatives like quixel mixer which is flat free. I wish the industry would move away from these overpriced 3D packages that are exclusive, sluggish, and slow to adapt.
