Your right thats why a retopologise low poly with manual UV Unwrapping is need no big deal it's possible combining blender unwrapping and Painter baking system, 3D coat seem to do that fine too...
I'm new to all this, a photographer. I like the idea of combining a photographed environment and being able to bring in specific 3d objects such as furniture etc which I may have imported using this Photo-import technique. What industry - jobs, is this part of?
I was just playing with Substance Sampler because of the Embroidery layer filter (it’s just too good), was delighted to see that sbs files can be exported and further enhanced in Designer (like adding an image input slot as an exposed parameter). The 3D capture is yet another jaw dropping feature and curious to see how it integrates further down the line… can the mesh be enhanced with substance model graphs one day? The .sbs file I do see it as an option to export. Kudos for the USDZ, apple fined viewers, AR, Motion, etc prefer that self contained format.
Mmm that UV generation is not nice, with great loss of pixels, would be interesting if the post-processing section recreates a tri-planar projection and make better use of the pixel space and better would be to have some udim options
Lightroom and Substance Sampler combo.. thats 2 different plans. Why not just integrate Substance into the main plan along with photoshop premiere pro etc
If i can reduce the polycount i probably rebuild a more efficient UV unwrapping with Blender to keep more texel resolution with less need for big texture ressources, especialy for cylinder and planned shape objects , painter is great value for that kind of baking and retouche especialy if i work with UDIM ...
Great 👍 I grabbed a good plan with all Adobe softwares, Im a Reality Capture user originally but I was excited to test this! Do you have anymore Tutorials regarding scans and painter?
Hi Wes thanks for this, Can you tell me would it be possible to change colours once processed? For example in the Adobe bag scan example could you change the blue pattern to a red? The hue/saturation filter only works globally not like photoshop where you can select a target colour to change.
Love how they added a photogrammetry process into the suite, which is amazing, but we are still limited to 4K. Why!? Do not say VRAM because I can almost guarantee at least 60% of developers have 8GB or more of VRAM on their main machines with more than a few with 16Gb up to 48GB. I am expecting 16K in these apps and Sampler still is only 4K.
Sampler Layer stacks are limited to 4K for the moment due to technical constraints. However, the photogrammetry result generates textures higher than 4K. When adding the result to the project, the first layer 'model.obj" has textures of 8K and even 16K in some cases.
I'm getting an unexpected error when trying to submit my photos during the processsing stage. That's all it gives me though. An unexpected error. Any idea what might be causing this?
Just saw on another video that importing as RAW may be buggy if your doing that. Just an idea. Seems JPG best option for importing, but that is not verified, just a thought.
kinda of a fail that the lightroom app on android doesnt do focus peaking. Also.. by default the lightroom app saves stuff to the adobe cloud.. yet in substance sampler you cannot grab the files from there... clunky.
It is high poly mesh so just like any high poly to low poly workflow. Take HP mesh and retopologize it. UV unwrap new mesh and bake HP details on the LP.
There isn't a limitation in set number of photos. It is the total number of pixels of all images that is 6G pixels. It can be 500 images of 12MP or 180 images of 36MP
There's not much to set here. Sampler should automatically recognize different cameras and put them into different photogroups, you can then manually edit lens and sensor info if needed. We aim to keep the settings friendly and easy.
Not really, it depends on your subject. You just want sharp-high quality gear. We have an advanced series explaining the reasoning behind it, but we don't get into recommending specific lenses as every case is different.
Yes theoretically, but photogrammetry has a lot of problems with thin objects like branches and leaves, I'm not sure you would get a good result. Almost anybody doing this professionally will recreate branches and leaves manually afterwards, and only scan the main trunk.
Saddly it's not working, your software is not grouping photos in the same group, but is creating one photogroup per photo. Tested with multiple photosets that are working on reality capture. The photosets are perfect, shot with an automated rig and already masked. Something is wrong with the version you released! Also, no more than 6 giga pixels? Wtf? How do you want to have accurate results, and how can we use two sets with cross polarized light with so few pixels?
Hey, the grouping for photos is done automatically. Sometimes it misinterprets based on metadata. I find editing in Photoshop or other applications is not ideal and can cause this. We'll improve this for sure in the future.
@@Substance3D It's not sometimes, it's always misinterpreting, I tested more than 10 sets including non masked, and it has never worked a single time. Also, I always mask the pictures I take with photoshop, to reduce the storage space needed by 80-90%, which is equivalent to several tb now, a lot of loading time and processing time, and I will not stop because a stupid IA can't group the photos. You should not release a software untested or in alpha stage, or you should add more manual settings to adjust everything. Grouping photos should not be something done automatically, but manually, I don't know how it's came to your mind!
I N C R O Y A B L E - Cloud Point before its this operation was only use into Autodesk Maya, Max, Flame or Other high end program for film industry or video games...the guy on video got a Iphone PRO to got the focus peaking (green line) only on PRO Phone or DSLR in video mode (Blackmagic, RED, CANON, Sony Alpha) The best solution was to film stabilised and convert to image (38 or more)
Let me know when video is the standard capture for photogrammetry. Going around in circles for 15 mins just to capture a small to medium size object too me is tedious.
You can but you’ll need to crank up your shutter speed; motion blur really messes things up. Also you’re capturing way more data than you need; recording at 30fps for like a minute will get you 1800 images. Too much data. Also, again, you’ll need to go through all those images and remove motion blurred ones.
It is very unfortunate to see Adobe waiting for this feature to ship it in 2023. 2022 version pretty much got no proper improvements but this such important feature is shipping with 2023 only. It is obvious this feature will grow on in the future, so why not at least ship the new scan feature with 2022. Very unfortunate.
Not a good idea, the images that you feed need to be extremely clear, and should have no artifacts while moving around. The better the input, the better the output. And a moving camera video will definately create some of that blur causing artifacts.
If i can reduce the polycount i probably rebuild a more efficient UV unwrapping with Blender to keep more texel resolution with less need for big texture ressources, especialy for cylinder and planned shape objects , painter is great value for that kind of baking and retouche especialy if i work with UDIM ...
![HIWWHEE | OHM [Official MV]](http://i.ytimg.com/vi/OJJLtTLqLag/mqdefault.jpg)
![เขามัทรี - เอ็กซ์ ศุภกฤต - จอนนี่มิวสิค [ Official MV ]](http://i.ytimg.com/vi/AS9ufKuQIVs/mqdefault.jpg)
Incredible how the world of 3D is improving. 20 years ago to carry out this tutorial you would have had to mortgage your house
Today, free tutorials, no house
hahahaha
I have Substance 4.1.0 installed but there is no button for 3D Capture. Running it on a Mac. Can you help me wit that issue?
Why can i not find the 3d capture import option on the version of Substance3d sampler i just downloaded?
Textile density looks very low, there is a lot of wasted texture space. It needs to have new UV’s.
True story, although less seams
Your right thats why a retopologise low poly with manual UV Unwrapping is need no big deal it's possible combining blender unwrapping and Painter baking system, 3D coat seem to do that fine too...
can we export high detail 3d mesh to Zbrush do some clean up, then reimport to sampler do re-baking the textures?
Bringing in arbitrary meshes is not possible yet, but it's something we are looking at.
I'm new to all this, a photographer. I like the idea of combining a photographed environment and being able to bring in specific 3d objects such as furniture etc which I may have imported using this Photo-import technique. What industry - jobs, is this part of?
why i don't have 3d capture item? i have steam version of sapler
I was just playing with Substance Sampler because of the Embroidery layer filter (it’s just too good), was delighted to see that sbs files can be exported and further enhanced in Designer (like adding an image input slot as an exposed parameter).
The 3D capture is yet another jaw dropping feature and curious to see how it integrates further down the line… can the mesh be enhanced with substance model graphs one day? The .sbs file I do see it as an option to export.
Kudos for the USDZ, apple fined viewers, AR, Motion, etc prefer that self contained format.
Mmm that UV generation is not nice, with great loss of pixels, would be interesting if the post-processing section recreates a tri-planar projection and make better use of the pixel space and better would be to have some udim options
Meshroom does it with UDIMs and it has a couple of other UV options, including importing a model with its UVs, and it is free.
Does this work well for small detailed 3d models?
Same here. I cant find the import option in version 4.1.1
Lightroom and Substance Sampler combo.. thats 2 different plans.
Why not just integrate Substance into the main plan along with photoshop premiere pro etc
cause they make more money this way 💀
If i can reduce the polycount i probably rebuild a more efficient UV unwrapping with Blender to keep more texel resolution with less need for big texture ressources, especialy for cylinder and planned shape objects , painter is great value for that kind of baking and retouche especialy if i work with UDIM ...
Great 👍 I grabbed a good plan with all Adobe softwares, Im a Reality Capture user originally but I was excited to test this! Do you have anymore Tutorials regarding scans and painter?
Awesome! We do have a more advanced series. Please check them out here. adobe.ly/3WktITR
any idea if and/or when the UV generation will let us create multiple UV tiles, i.e. UDIMS? Pretty please, with tons of sugar on top.
Great content 👍🏼
thank you just what i needed
Great but how many years we should wait to add a diffuse map do bake/projection on substance painter high res on low res?
I have followed the steps, but I don't seem to have the materials output window next to my 3D model.
Trial version doesn't have a 'Mesh reconstruction' step? How come?
Hi Wes thanks for this, Can you tell me would it be possible to change colours once processed? For example in the Adobe bag scan example could you change the blue pattern to a red?
The hue/saturation filter only works globally not like photoshop where you can select a target colour to change.
Found it, Colour Variation filter.
My ultra 23 phone can invert picture at 200mp. If I delete backgrounds, can this still work since other side is the same?
Love how they added a photogrammetry process into the suite, which is amazing, but we are still limited to 4K. Why!?
Do not say VRAM because I can almost guarantee at least 60% of developers have 8GB or more of VRAM on their main machines with more than a few with 16Gb up to 48GB. I am expecting 16K in these apps and Sampler still is only 4K.
Still waiting on Unity to get their heads out of their butts with ArtEngine so I can go back to 16k without having to switch to Mari.
I’m sorry, but where the hell do you need 16K texture resolution?
@@Settiis VFX clients
Sampler Layer stacks are limited to 4K for the moment due to technical constraints. However, the photogrammetry result generates textures higher than 4K.
When adding the result to the project, the first layer 'model.obj" has textures of 8K and even 16K in some cases.
@@Substance3D That is cool! Thanks for messaging that.
I'm getting an unexpected error when trying to submit my photos during the processsing stage. That's all it gives me though. An unexpected error. Any idea what might be causing this?
Just saw on another video that importing as RAW may be buggy if your doing that. Just an idea. Seems JPG best option for importing, but that is not verified, just a thought.
Ok but what kind of smartphone i have to get to do photogrammetry ? Apple ones ?
what's the name of that iphone app?
no udims/16K or custug marking same points on diferent photo`s ? ?
Can I have my camera fixed in one place but have the object rotate?
Can I scan bigger elements like a car ??
A car will be difficult with all the reflections. But worth a shot.
44 GB of memory required for the Mesh Reconstruction? it's possible?
Great video :)
Thats really cool ! 😍✌
Wooooooow i love adobe!
Who was 3D capturing you in this video?
kinda of a fail that the lightroom app on android doesnt do focus peaking. Also.. by default the lightroom app saves stuff to the adobe cloud.. yet in substance sampler you cannot grab the files from there... clunky.
Can you do this with an Android phone as well?
Yes it works with any phone camera. The phone is just used for taking photos. No app is required.
Only 4K output texture? really?
Awesome
Is mesh reconstruction limited to windows? It's not available for me.
It is not, it works on OSX too. Verify you are using latest versions of the app and OS.
That's pretty cool but how can we create new UV's and new topology for the mesh to be better optimized ?
It is high poly mesh so just like any high poly to low poly workflow. Take HP mesh and retopologize it. UV unwrap new mesh and bake HP details on the LP.
So we have to bake the textures on the low poly ? I never done that but I guess it's possible.
Do you know if there is a maximum amount of photos allowed to be added to the scan?
There isn't a limitation in set number of photos. It is the total number of pixels of all images that is 6G pixels.
It can be 500 images of 12MP or 180 images of 36MP
@@Substance3Ddoes DNG work or i need to convert in PNG or JPEG or EXR first ? I want to preserve as much color as possible for an ACES CG workflow.
How are the setting for manuel alighnment from multiple cameras ?
There's not much to set here. Sampler should automatically recognize different cameras and put them into different photogroups, you can then manually edit lens and sensor info if needed. We aim to keep the settings friendly and easy.
Great🔥
is there a recommended lens focal length for the photo gathering?
Not really, it depends on your subject. You just want sharp-high quality gear. We have an advanced series explaining the reasoning behind it, but we don't get into recommending specific lenses as every case is different.
@@Substance3D OK, good to know. thanks!
would it be possible to 3d capture a tree say 50m tall? possibly with a drone
Yes theoretically, but photogrammetry has a lot of problems with thin objects like branches and leaves, I'm not sure you would get a good result. Almost anybody doing this professionally will recreate branches and leaves manually afterwards, and only scan the main trunk.
@@Substance3D yeah fair call. very interesting tech though many different applications other than games
What app do you use to take photos on your phone?
Adobe Lightroom Mobile
bravo..
Saddly it's not working, your software is not grouping photos in the same group, but is creating one photogroup per photo. Tested with multiple photosets that are working on reality capture. The photosets are perfect, shot with an automated rig and already masked. Something is wrong with the version you released! Also, no more than 6 giga pixels? Wtf? How do you want to have accurate results, and how can we use two sets with cross polarized light with so few pixels?
Use metashape or reality capture instead. Miles ahead of this at this stage
Hey, the grouping for photos is done automatically. Sometimes it misinterprets based on metadata. I find editing in Photoshop or other applications is not ideal and can cause this. We'll improve this for sure in the future.
@@Substance3D It's not sometimes, it's always misinterpreting, I tested more than 10 sets including non masked, and it has never worked a single time. Also, I always mask the pictures I take with photoshop, to reduce the storage space needed by 80-90%, which is equivalent to several tb now, a lot of loading time and processing time, and I will not stop because a stupid IA can't group the photos. You should not release a software untested or in alpha stage, or you should add more manual settings to adjust everything. Grouping photos should not be something done automatically, but manually, I don't know how it's came to your mind!
yep this shit slow af
@@Lootzuh Nah everyone who does photogrammetry knows 3DFZephyr is the best.
Which you are use app for take photo?
Adobe Lightroom Mobile.
@@Substance3D thank you so much!
where is the option 3d capture beta ?
The update is available in the latest release - Sampler 4.0
I N C R O Y A B L E - Cloud Point before its this operation was only use into Autodesk Maya, Max, Flame or Other high end program for film industry or video games...the guy on video got a Iphone PRO to got the focus peaking (green line) only on PRO Phone or DSLR in video mode (Blackmagic, RED, CANON, Sony Alpha)
The best solution was to film stabilised and convert to image (38 or more)
Is this an app or am I missing something?
This is a new feature that's now part of Substance 3D Sampler.
@@Substance3D thanks! I thought you guys were also making a 3D scanning app for phones?
what is the name of app
Adobe Substance 3D Sampler.
This does not work. 3D Capture is removed.
Let me know when video is the standard capture for photogrammetry. Going around in circles for 15 mins just to capture a small to medium size object too me is tedious.
You can but you’ll need to crank up your shutter speed; motion blur really messes things up. Also you’re capturing way more data than you need; recording at 30fps for like a minute will get you 1800 images. Too much data. Also, again, you’ll need to go through all those images and remove motion blurred ones.
@@sayrith Yea that's a good point
On top of that, you need at least 4K video, 1080p is way too low res. Even 4K is only 8Mp, 12Mp minimum is recommended.
This guy sounds like an Anime character.
I would prefer using drone to take photos tbh.
You can use a drone to take photos. There is no requirement on the capture hardware as long as you import photos in Sampler.
It is very unfortunate to see Adobe waiting for this feature to ship it in 2023. 2022 version pretty much got no proper improvements but this such important feature is shipping with 2023 only. It is obvious this feature will grow on in the future, so why not at least ship the new scan feature with 2022. Very unfortunate.
Why not record a vidoe and export each frame as a png? would that work
Not a good idea, the images that you feed need to be extremely clear, and should have no artifacts while moving around.
The better the input, the better the output. And a moving camera video will definately create some of that blur causing artifacts.
If i can reduce the polycount i probably rebuild a more efficient UV unwrapping with Blender to keep more texel resolution with less need for big texture ressources, especialy for cylinder and planned shape objects , painter is great value for that kind of baking and retouche especialy if i work with UDIM ...