Funny thing. In this video I mentioned the "graffiti" on the rock. A year later I'm just now realizing, it may not be graffiti at all. The rock was brand new so it might have been marked XXL in the store they bought it from.
Hmm very strange.In Meshlab after i hide the so called point cloud model i click file/import mesh but i do not have a .ply file.I followed everything correctly.I am using your rock images just to try it.What i did wrong?
George Ginev All I can think of is that there was an older version of Visual SFM that didn't save that file. 2 options are either make sure you have the latest version. or you can also do it like this. Instead of looking in the root for the PLY look in [filename].nvm.cmvs\00\models\ (filename is whatever you called it) and load the file called option-0000.ply
Phil Nolan This is my directory: Rock_Video.nvm.cmvs\00\models and there is also no file inside.Maybe i need the newest version?Can you give me a link for the newest one?I searched in google but its full with links and i get confused.
The current one that I have is 0.5.26 You can see that in the Help / About menu of VSFM. The link is in the video description., But here it is again. ccwu.me/vsfm/
Thank you Phil!!!! I just bought a CNC router for my hobby and 3D scanning was definitely something I was beginning to research. Your tutorial is awesome and was very useful to me
One of the best stuff I found on UA-cam in years! I really mean it. Just loved your tutorial. The technology is not new but the accessible way you have showed to us is amazing. Congratulations man. It's a shame I just watched it so late after posted.
Olá Ricardo. Tudo bem? Será que vc pode me ajudar? Instalei o VisualSFM e ele funciona corretamente mas o CMVS for windows explode na tela e desliga imediatamente. Não sei se é minha versão do windows10 que não aceita. O que devo fazer? Me ajuda?
I am amazed what individuals in these communities give away for free. I would have to spend several hundreds (if not thousands) of dollars in software and classes for this information. At the same time, you've made it seem so easy with your step-by-step presentation. Thank you.
What a great tutorial. This is a plus for me I'm in the middle of the beginning of a project that uses photogramography and have been struggling with the filters in Meshlab . Also now I have two new softwares to use. Thanks
As someone who is very knew to photogrammetry (as in, like, today new) this tutorial was extremely helpful and I appreciate immensely the lengths you went to in order to make this video! Thanks a ton, seriously!
I have one question for you in regards to this process-- after using VisualSFM to create my dense mesh and saving it, and then opening it within Meshlab and following your process as was in the video, I don't have, for whatever reason, a .ply file. Is this something Meshlab automatically generates? I'm curious and a little frustrated because I feel like I'm soooooooo close to being able to successfully replicate the tutorial in your video with the exception of this particular hangup. Any ideas?
***** The PLY should be created by Visual SFM when CMVS is finished running. Check what version you have I think older versions may not have done that. The one I have is 0.5.26. For now you should still be able to find it in the folder with the pics. Look in the folders [filename].nvm.cmvs\00\models\ And you 'll find "option-0000.ply " That should work.
Sir, I am trying to get into game developement, and I am a real noob. I am (still) not able to draw objects in a game engine, but simply fotographing (clay made) objects, and inserting them (using your method) into the engine will help me to get started. phantastic tutorial/demo video, thank you
@michael lowe If you were to use this as a game asset for instance, you would want to create a low poly version in your 3d software. This can he done in many ways, decimating the geometry or retopology. Then you would use the high poly model to create a normal map which basically tells light and shadows how to behave over top the image as of the model was still high poly. Games use normal maps so lower poly models appear to have a higher polygon count. The trick is to use more polygons around the edges that will appear visible while rotating around the object and less on the surfaces you will be mainly looking at from straight on.
While it's taking "a little more time" I wanted to take the opportunity to thank you on this great tutorial. I did my thesis using fringe pattern projection. I remember we would look for objects without texture... so I reached for a white bottle with horrid results. This is really fun Again thanks
Eduardo Alejandro del Corral Lira The results of your scan were horrid your you spilled milk all over? lol Thanks. Hopefully the rest of your scan goes well.
Oh it has, second was much better. I didn't want it to sound negative my results back at the lab were often quite bad but that was part of what kept me tinkering and enjoying the process :)
Especially liked author's drawing attention to options, filenames, and running commentary on program execution (what it was doing). Puts new colours in your paintbox ;)
I really appreciate how they visualized the construction/alignment process in Visual SFM. It has that same sort of appeal like the old drive defragmentation visualizations. I don't know how many others like to watch stuff like this, but I was one of those kids who could watch a lava lamp for a long time. ;) I always appreciate when a program lets you observe some of what's going on while you wait for the output.
Great tutorial video Phil, very helpful. I like the way you make a mistake or two and correct it. This shows me what can go wrong and what to look out for + your explanation of the problem. I am thinking of trying this to do some bas reliefs on my cnc router. Rather than scouring the internet for appropriate 3d images I can photograph a subject and turn it into a 3d image myself. Wish I had found this tutorial sooner. PaulM
Thanks, I learned a great deal from this video. It seemed this methodology was a ton of work, but then you weren't working with an expensive laser scanner either. I can't believe software can do all this. Very impressive but so time consuming.
I did a post on 3D scanning with photogrammetry before but I wanted to do a full video tutorial. So this is how you do 3D scanning with just your own camera (even your phone!) and some free, open-source software. There are simpler ways to do it but this way gives really good results and more control.
zwiadowca88 I don't think I've ever used .iges so I checked and MeshLab does not have it as an export option. it's a NURBS format isn't it? Scans come in as polygonal models so if you want .iges you'd have to find a converter or just load into Solidworks as polygons if possible.
Phil Nolan how can i make and scan a completley 3d model? I want to print Parts for beyblades, but need first of all 3d models. s1233.photobucket.com/user/TTUL8TR/media/DragoonPPcopy.jpg.html any idea??
worldofbeys If you are trying to scan a full object the only way I know of would be to somehow support it in the air so that you can take pictures from all sides. The other option is to make two scans of the top then the bottom then merge them together in 3D software. For something like that it might be easier or better to just model it in 3D.
You sir deserve my sincerest regards! You've just opened a whole new opportunity for making a living out of what I love! I can go back to my passion - the landscape photography and making money out of it, without being super lucky working for Nat Geo :D . Adios to the studio portraits, the children's birthdays and weddings! YOU ARE MY HERO! *Deep bow*
Great tutorial on how to craft skillfully with free tools! Thanks for explaining me how to fix those annoying scan holes which is always a problem even with professional software.
the future of 3d modeling of natural items. Dice already uses this a lot in their games, and they have the most stunning graphics of any studio. I was actually kinda dissapointed when I found out much of their natural art was photogrametry. I thought they had some superhero artists or something hahahaha.
Oh yeah photogrammetry can do a really well job if you have a detailed model. Reality Capture costs a little bit but does an even better job and it's a lot easier.
From someone who knows nothing about this, I thought this was a great video! You have a nice style of commentary, and you seemed to show primary features and a good end result. Very good demonstration video! Thanks! Subscribed.
Phil were experimenting using your 3D scanning with our drones, although it requires some work, it seems to be working great! Thank you for this! Great job
Excellent Video. very nicely done.. I don't need to buy a expensive scanner. I already have a 3D router and Mastercam.. Now, it just gotta learn these softwares and put my camera to work. I love it. Thank you.!
Phil, thanks so much for putting this together!! The yard around two sides of my house needs improvement, but I'm the kind of guy that likes to draw things up in Autocad 3D first and play with ideas there before I start moving dirt. The challenge before me has been surveying the terrain (the hard part when I know next nothing about surveying) accurately enough to re-create it as a mesh in Autocad (the easy part). Enter photogrammetry and your handy tutorial that uses FREE applications! I did a trial run and believe I can create a more polished version that will serve my landscape design needs, without having to do the tedious survey work. Thanks Again!!
+Monte - Yeah this should work well for that. I think I mentioned in the video I scanned a hillside in the part once and it perfectly showed where all the rocks, trees, and grass were.
Hi Phil - I took another series of higher res photos and did another run of the landscape and side of the house. When I imported the mesh into Meshlab it made a beautiful reconstruction, but refused to cooperate when trying to do the surface reconstruction poisson. I suspect the dense point cloud is too much for it to handle, so I removed a bunch of the points, but it still won't do it. Is there a way to 'optimize' the mesh before attempting this operation? I'm using Windows 7 x64 with 8GB of ram, and a decent 1gb video card....
This was my first look at 3D modeling I was always under the impression you had to draw the model first then texture it. This gives me hope. Thank you for the video.
Well this isn't modeling, it's scanning. Like using a photo copier for physical objects. Modeling is more similar to building something out of wood (just on a computer).
Hi, Thank you for your awesome video, I'm new to 3D scanning. I am following you step by step, but when I want to import my .ply file to meshlab , I couldn't find that file !!!!! I'm wondering where I missed ? could you please help me.
Tanya Shirazi When you start the CMVS part ( 6:21 in the video) you save out the file. I can't say where you chose to save it. I would also make sure that you have the latest version of everything. The older version of CMVS didn't save the file in the root where you choose, it creates a folder structure there like this \filename.nvm.cmvs\00\models\ (filename is whatever you picked). The problem with this is that in that folder it may create more than one PLY file and there's an extra step to merge them. The new version always creates just one file.
Phil Nolan Hi, Thanks for quick reply, I just download the last version which is available , as you said it creates the 00\models folder but nothing is available in that folder as well !!! could you please post the link that you are using, maybe I downloaded the wrong version. appreciate.
Phil Nolan sorry Phil, I downloaded cmvs and put it in the same folder as visualsfm , but nothing changed, I still don't have .ply file !!!Any suggestion!!
Tanya Shirazi The exact ones that I'm using are VisualSFM Cuda, x64 version: ccwu.me/vsfm/download/VisualSFM_windows_cuda_64bit.zip CMVS for WIndows: github.com/pmoulon/CMVS-PMVS/archive/master.zip
Wow this is a much better workflow than the one I follow! Thanks for the vid! I usually used agisoft photo scan, and then cleaned the mesh in zbrush, then UV'd in Maya, then brought everything back into substance painter, and finally into 3ds max, and from there into cryengine. You saved me ALOT of time.
Well as it is you wouldn't bring these directly into a game engine, the poly count is too high. When I'm finished here I normally bring it into 3D-Coat and adjust any sculpting, then retopologize it and UV, bake the textures to the new mesh , and touch up any painting. Still, that's only one additional program.
UA-cam has cut what they pay creators by a lot. If you like these videos please consider becoming a patron. Patrons also get bonus content that isn't on my channel and I'll be sending out 3D printed minis to some patrons. www.patreon.com/philnolan3DJ
Glassic Gamer I was making double what I am now. It was enough to pay my phone bill each month. Since they cut it I only get a check from them every few months.
it would be awesome if UA-cam integrated way to tip creators for a nice, sending 1$ would be the same as watching 500 adds (last time I checked). Patreon also works of course.
Wemja They used to have a "tip jar" with a link to donate on the channel page but they took it away" Maybe because nobody used it. You could always try UA-cam Red, $10/mo for no ads. It doesn't go directly to a UA-camr though. I think like .5% of my revenue is from Red users.
Sorry to whomever's comment I just deleted by accident. You were asking how I get less than $100 a month from UA-cam. I don't. That's what I was saying earlier. I usually only get a payment from them every few months when that small amount adds up to more than $100.
What an excellent tutorial! I'm not sure whether the technology has changed but you made the process so clear especially to novices like me. Thank you!
Mesh lab has changed slightly. I believe I noted it in the description. Otherwise I believe it's the same. These days I prefer Reality Capture. I have videos on that too. It's not free but very cheap.
Hi Phil, I should have taken more notice of your description text below the video. It's all there. I have used the links, followed everything, and it is all working. Thank you so much for this video and the extra help you gave me !
Never thought a video about a rock could be so interesting :D Since i'm completely talent-free when it comes to creating textures and 3D models, this might prove really worthwhile in the future. Thanks a lot.
TheAmazeer To scan a person? Sure. Loose hair and clothing like an open jacket can be difficult. But as long as you can get the person to stand totally still you could scan them. It would be best to have them in a T pose, perhaps with arms slightly lowered.
Thanks for this excellent video. It helped me doing my very first 3D scan which out worked amazingly well for an STL out of a a bunch of cell phone shots. My honest kudos to programmers of open source software like VisualSFM, PMVS/CMVS, meshlab, and OpenSCAD which really make all the difference when creating 3D models and 3D prints for free!
Thanks, Phil. Not used this program or any like it but have extensive experience at photo manipulation software. When you were deleting the unwanted ground . . the icons buttons to the left and right of it look typically like alternatives to using the square cage to select.
this is great! there are 3d scanners that are sold for thousands of dollars. Your technique's results better quality 3d scanning and it is free! thanks for sharing! :)
Very good looking result - and all for free ! Now I only need to know a free or inexpensive way to turn the mesh back into a real (scaled?) object again - I'm planning on making my personal shoe last ! Thanks for the information ! Thumbs up !
I'm not sure what you mean. Did I read it? I wrote it. The image is all of the files that come with it. If you see different files then you must have downloaded something else.
that's awesome stuff man... pretty intimidating to a guy like me... doesn't take very long to be way behind the times on software or any tech at all...thanks for the video.
Great video man! I appreciate the fact too that you didn't talk fast or "rush" through the video. A lot of people doing tutorials on stuff similar to this always seem to rush through it, forcing me to go back and watch a part several times...lol. I do have a few questions for you. 1. Can I use a DSLR camera for this? Manual or Automatic settings? 2. Could I import the file into ZBrush to clean it up prior to 3D printing? 3. Is it a must to shoot the pictures outside? I have bright daylight white "florescent" lighting in my room, which is why I asked. You're a life saver man as I was about to invest in a 3D scanner haha!
+Kirkland Shunk Thanks for the compliment. You can definitely use a DSLR, if you watch my other video, linked at the end of this one, I was using a DSLR. There is an older video that a guy did for another scanning program, where he mentions settings for a DSLR that seems to work well. ua-cam.com/video/NsBg-m2hrIM/v-deo.html I don't use zbrush for my sculpting but it should be no problem to import the model. I don't know if zbrush will accept the color though. I know in 3D-Coat I can import the model with color if I import to Surface mode with Voxelization turned off. It does not have to be outside at all, actually florescent lighting would be great I think because it gives very few shadows. I wish I had it for my own scanning. Some of my scans were just done in regular lighting though and it's OK as long as it has even light all around.
Phil Nolan Awesome! I actually did watch your video after I commented so I found that out right away haha. What I'm planning on doing is making sculptures of characters using Super Sculpey(which is a tan color), so I think ZBrush won't have a hard time with it, not sure. Yeah that's what I thought. Thanks for the information though! Really appreciate it! Subscribed as well!
+Phil Nolan I have downloaded and expanded the visual sfm windows cuda 64 bits with no problem. I have some problem with the cmvs-pmvs-master. It has several folders - binaries, license, program. Shall I put the whole folder (cmvs-pmvs-master) or shall I just put the windows 64 vs2010 into the sfm folder? sorry for the stupid question
+Phil Nolan thanks a lot. I did add the binaries files. It worked. But when I press the tab after the reconstruction, only a poor quality picture. I do not know whether the problem is because of my pictures or the program. I shall repeat the program with your pictures. Shall get back
+CP HO Dear Phil I have downloaded the viusal SFM windows cuda 64 beat and the CMS tool. I used your rock photos for a trail. the whole process ran smoothly, including the full 3D and dense 3D reconstruction and the CMVS tool. On pressing the tab key, a nice rock photo appeared. However, it is only 2 D and not 3 D, can not rotate is as in your demo. In fact when I was in full 3D reconstruction before CMVS tool, the rock was also in 2 D, cannot roatate. Any idea what went wrong? Shall I reinstall the SFM? I have a cuda Geforce FX card. thanks
Thanks for making this video. I looked at the SFM web site and I was very happy to see that it will run on Linux. I haven't checked the other software yet. Anyway, this looks really excellent. I just started with 3D modeling in Blender. This SFM and CMVS could really help me to build up a library of original models to use in animations. Now I'm going to watch your newer video. :D
Hey greate tutorial! One thing I found out is, instead of using/taking a bunch of photos, you could just film your object, and then in for example in Adobe Media Encoder render the movie to a jpeg sequence, depending on the framerate you are using, you could also change the number of images really easily. Just something I found out recently :)
Neon Artworks I mentioned this in the description. People ask about doing it over and over. lol I think still photos are higher resolution for better quality.
great tutorial thanks. As a follow up, can you describe how one might provide fairly accurate and/or accurate dimensional measurement to scans? (1:1 dimensions). What are different ways this could be done? If you took measurement(s) can you apply them? Or if you had ruler object in photos etc ? Does focal length affect proportions and make it harder for software to work correctly , if FL was not consistent? Or does it kinda work out anyway because of how it puts it together in 3 dimensions? Could you 'safely' change FL during shoot for you convenience etc?
Sorry that's a little beyond my knowledge. You could definitely put a ruler in the scene or just any item which you know its size, then adjust the final model's scale to match.
Thanks, ruler tape stuck to object if wanting most accurate version of that I suppose. Or maybe on ground would be same accuracy idk. But is keeping the same focal length kinda important in general? I wonder if there might be some guidelines about FL vs distance to subject for more accurate/proportional results. I haven't tried this yet but I could find out for myself by taking multiple sets of photos, one with a radical range of FL, a couple others with different fixed lengths. I good size symmetrical object that had some smaller symmetrical features would be a good test I suppose. I wonder if a longer lens would end up having the software 'think' the camera was closer than it was in shots with shorter length. With most zoom cameras like mine its hard to document the length setting since is not indicated., and you'd have to be careful to keep away from the zoom control. That is of course if that mattered much, which it may not. I kinda guess that some kinds of fencing is really hard to do?, (possibly epically hard) But a small enclosed fenced area of chain link fence would probably tell a lot about accuracy. (like a 4x4 foot enclosure of 4' tall fence). Curious if you have experience with screen-like subject. Does it make larger file size in software than 'normal'?
Thank you so much for such an elaborate video. I am getting into aerial mapping with drones and this software wont break the bank like Pix4D. Going to mess with this tomorrow :)
Awesome tutorial mate! I been wanting to try this stuff out for ages. You explained everything in perfect detail and I managed to get from taking the photos to the finished model on my first ever go at this. Subbed and Liked :)
Great video. But what do you do when you end up with two pointcloud models in VFSM? I took two good sets of photos around a little figurine. One from an even level and one from a higher level. And I did everything up to 7:30. I ended up with two models. One model made from the set of photos taken at the lower level and one model made taken from the set of higher leveled photos. Do you know how this can be fixed?
Thanks Phil!!This is awesome! Can one scan the models at a lower density/lower poly, but still high quality (Like low poly models) If i will use 10+ models in my scene like this it with crash, so looking at making it lowish poly but realistic looking...?
There are some decimation options in Meshlab but i have not used them so I couldn't say how they work. I really think the best option would be retopology, as I mention in the video. IMO the best option for that would be 3D-Coat, though there are some other programs that do it too.
First off, thanks for the tutorial, it was well put together and easy to follow. A few questions/comments/ideas: ** Do you think putting the item on a turntable or swivel chair and rotating the item would help? -- Wouldn't it be easier, you'd have more consistency with distance from the object as well as intervals around the object? -- I could see inconsistent shadows/lighting causing an issue; as you turn the object, the shadows would change, but if you're lighting was good maybe it wouldn't be too bad? -- I'm also not sure if the unchanging background and camera position would cause any issues when creating the point cloud. Do you think if the background was all white it might work better or do you think this would even be an issue at all? I assume that VisualSFM generates the camera position for each shot so having a stationary camera shouldn't be an issue. ** What effect, if any, do you think using a phone camera over an actual camera would have? ** Do you think using a burst shot feature on a camera would work as well as normal single shot? What about taking a video of the object and then dicing it into frames using VLC (or other software)? ** I'd like to use this 3D modeling technique to get a 3D model of my head to put into videogames; right now it looks like uraniom (uraniom.co/) would be the way to go. Uraniom says that it's compatible with zip archives containing three files (.obj, .mtl and .jpg files); are you familiar with what type of export that would be? Thanks, and sorry for the question dump
-All of these questions have been brought up several times in the comments below except the last one. From what I understand from people who've tried it the turntable inside a light box with no shadows and white background can work but the software uses the background so it's better to have the BG moving with the object. -Phone camera works fine for me, that's what I usually use. -Burst might be too many pictures, I would think, sounds like a waste of smemory to me. -video... I think it would be hard to keep it in focus and get everything composed right. -As for OBJ/mtl/jpg, those are the files that are exported at the end of this so it should be fine. I personally use Collada (DAE) file but that's just because I'm going to edit it further. You could be able to export OBJ (which includes MTL).
Very cool demo. Thanks so much. Can you recommend 3D modeling software? In particular, I'm interested in being able to output a series of frames, rotated precisely in a sequence about a third of degree per "frame", from a particular point of view, or "camera location" in grayscale format.
Thank you Mr. Nolan. I thoroughly enjoyed that video. I had no idea that kind of detail could be had with free software. Are there any special settings that must be on the phone when taking the pictures?
Just a tip for any time you are doing heavy CPU processing, but want your computer to continue to preform normally (like in the case of your audio glitches) If you have a multi-core processor (every processor these days are at LEAST dual core) navigate to the process that is hogging all the CPU usage in Task Manager under the 'Processes' tab. Right-click on the process, and choose 'Set Affinity'. This will let you select which CPU cores that process is allowed to use. Uncheck one of them... and now you computer will have one core free to to be used for other things, restoring your computers responsiveness back to normal.
Phil Nolan 16 cores - Nice! Even better. All the more reason to uncheck one of the cores. A single core wont make much difference in your processing time, but being free will make all the difference in your computers responsiveness if you still need to use it for other things. When a CPU intensive application is well designed, it doesn't matter how many cores you have... it's going to suck ever last bit of processing power your computer has, as it should.
+FPEvideos Well, I didn't notice any hit to my processing power while it works. So while a good tip over all I just mean that it isn't affecting me personally.
Hi phil, awesome tutorial. My name is James and Im really interested in creating 3d models ( mainly figures ) to import in to Blender for rigging and skinning and then exporting for use in Iclone 5, my chosen platform. I just wondered if its possible to use the Kinect sensor for the XBOX 360 as the sensor or will be SLR suffice? I looked at Skanect software which utilises the motion technology of kinect to provide the data, which was interesting but as i'm only requiring static 3d models I thought this might be more applicable. Any advice you can give me on this would be great. Keep up the good work sir..
James Harvey I have a Kinect that I've tried before I discovered this technique. It gives very, very low quality scans. I would recomend using the techniques in this video, photrogrammetry, instead. pretty much any digital camera will work. In this video I used my phone.
Super video. Really well presented. Two questions. For areas of the rock with limited detail, could you go back, retake pictures and incorporate them into an existing model? Second question, can you export into Cinema4d with this software?? Thankyou
+Nick Kingstone Thanks Nick. In theory I guess you could go back and shoot more photos, then start the process over. However in cases like this rock, it might be tricky since the lighting would change by the time you went back, especially if it was a different day. As for export. Sure, it exports as OBJ format, which almost all 3D software accepts. I don't use C4D but I would be very surprised if it didn't take such a univresal format.
Thank you Phil for the spectacular tutorial, I love it. I wonder if taking the picture for the rock while it sits on a glass penal would made it easy to get rid of unwanted background meshes? Ideally this should create a totally separated model from the bacground.
Funny thing. In this video I mentioned the "graffiti" on the rock. A year later I'm just now realizing, it may not be graffiti at all. The rock was brand new so it might have been marked XXL in the store they bought it from.
Hmm very strange.In Meshlab after i hide the so called point cloud model i click file/import mesh but i do not have a .ply file.I followed everything correctly.I am using your rock images just to try it.What i did wrong?
George Ginev
All I can think of is that there was an older version of Visual SFM that didn't save that file. 2 options are either make sure you have the latest version. or you can also do it like this. Instead of looking in the root for the PLY look in [filename].nvm.cmvs\00\models\ (filename is whatever you called it) and load the file called option-0000.ply
Phil Nolan This is my directory:
Rock_Video.nvm.cmvs\00\models and there is also no file inside.Maybe i need the newest version?Can you give me a link for the newest one?I searched in google but its full with links and i get confused.
The current one that I have is 0.5.26 You can see that in the Help / About menu of VSFM. The link is in the video description., But here it is again.
ccwu.me/vsfm/
Phil Nolan Thanks for the help.I will check it.
onestly I can’t imagine why anyone would dislike this,is clear and very helpful kudos to this dude
Thank you Phil!!!! I just bought a CNC router for my hobby and 3D scanning was definitely something I was beginning to research. Your tutorial is awesome and was very useful to me
"Not a rock music video, A video about an actual rock"
Made me lol :)
I quietly lol'd too. Hehe
+Garrett Stevens you beat me to it. loved that part. :)
One of the best stuff I found on UA-cam in years! I really mean it. Just loved your tutorial. The technology is not new but the accessible way you have showed to us is amazing. Congratulations man. It's a shame I just watched it so late after posted.
+Ricardo Pereira Thanks! I'm thinking up ideas for other tutorials I can do.
Olá Ricardo. Tudo bem? Será que vc pode me ajudar? Instalei o VisualSFM e ele funciona corretamente mas o CMVS for windows explode na tela e desliga imediatamente. Não sei se é minha versão do windows10 que não aceita. O que devo fazer? Me ajuda?
I am amazed what individuals in these communities give away for free. I would have to spend several hundreds (if not thousands) of dollars in software and classes for this information. At the same time, you've made it seem so easy with your step-by-step presentation. Thank you.
What a great tutorial. This is a plus for me I'm in the middle of the beginning of a project that uses photogramography and have been struggling with the filters in Meshlab . Also now I have two new softwares to use. Thanks
As someone who is very knew to photogrammetry (as in, like, today new) this tutorial was extremely helpful and I appreciate immensely the lengths you went to in order to make this video! Thanks a ton, seriously!
Glad you liked it.
I have one question for you in regards to this process-- after using VisualSFM to create my dense mesh and saving it, and then opening it within Meshlab and following your process as was in the video, I don't have, for whatever reason, a .ply file. Is this something Meshlab automatically generates? I'm curious and a little frustrated because I feel like I'm soooooooo close to being able to successfully replicate the tutorial in your video with the exception of this particular hangup. Any ideas?
*****
The PLY should be created by Visual SFM when CMVS is finished running. Check what version you have I think older versions may not have done that. The one I have is 0.5.26.
For now you should still be able to find it in the folder with the pics. Look in the folders [filename].nvm.cmvs\00\models\ And you 'll find "option-0000.ply " That should work.
Sir, I am trying to get into game developement, and I am a real noob. I am (still) not able to draw objects in a game engine, but simply fotographing (clay made) objects, and inserting them (using your method) into the engine will help me to get started. phantastic tutorial/demo video, thank you
keep in mind that these models have some many vertices that it will make loading screens and performance horrible
Yes I may have missed this when it was first posted. You will definitely need to retopologize these scans if you plan to use them in a game engine.
@michael lowe If you were to use this as a game asset for instance, you would want to create a low poly version in your 3d software. This can he done in many ways, decimating the geometry or retopology. Then you would use the high poly model to create a normal map which basically tells light and shadows how to behave over top the image as of the model was still high poly. Games use normal maps so lower poly models appear to have a higher polygon count. The trick is to use more polygons around the edges that will appear visible while rotating around the object and less on the surfaces you will be mainly looking at from straight on.
While it's taking "a little more time" I wanted to take the opportunity to thank you on this great tutorial.
I did my thesis using fringe pattern projection. I remember we would look for objects without texture... so I reached for a white bottle with horrid results. This is really fun
Again thanks
Eduardo Alejandro del Corral Lira The results of your scan were horrid your you spilled milk all over? lol Thanks. Hopefully the rest of your scan goes well.
Oh it has, second was much better. I didn't want it to sound negative my results back at the lab were often quite bad but that was part of what kept me tinkering and enjoying the process :)
Eduardo Alejandro del Corral Lira As long as you're having fun!
Especially liked author's drawing attention to options, filenames, and running commentary on program execution (what it was doing). Puts new colours in your paintbox ;)
I really appreciate how they visualized the construction/alignment process in Visual SFM. It has that same sort of appeal like the old drive defragmentation visualizations. I don't know how many others like to watch stuff like this, but I was one of those kids who could watch a lava lamp for a long time. ;) I always appreciate when a program lets you observe some of what's going on while you wait for the output.
I used to get mesmorised by watching my HDD defrag. lol
Great tutorial video Phil, very helpful. I like the way you make a mistake or two and correct it. This shows me what can go wrong and what to look out for + your explanation of the problem. I am thinking of trying this to do some bas reliefs on my cnc router. Rather than scouring the internet for appropriate 3d images I can photograph a subject and turn it into a 3d image myself. Wish I had found this tutorial sooner.
PaulM
Paul Marshall That's a great use for it. Need something? Just make a copy for yourself!
Thanks, I learned a great deal from this video. It seemed this methodology was a ton of work, but then you weren't working with an expensive laser scanner either. I can't believe software can do all this. Very impressive but so time consuming.
Yeah the amount of time taken is a lot more than using a laser, but that's the trade off for a free technique.
I did a post on 3D scanning with photogrammetry before but I wanted to do a full video tutorial. So this is how you do 3D scanning with just your own camera (even your phone!) and some free, open-source software. There are simpler ways to do it but this way gives really good results and more control.
Nice tut Phil!
***** ***** Thanks. :)
zwiadowca88 I don't think I've ever used .iges so I checked and MeshLab does not have it as an export option. it's a NURBS format isn't it? Scans come in as polygonal models so if you want .iges you'd have to find a converter or just load into Solidworks as polygons if possible.
Phil Nolan how can i make and scan a completley 3d model? I want to print Parts for beyblades, but need first of all 3d models. s1233.photobucket.com/user/TTUL8TR/media/DragoonPPcopy.jpg.html
any idea??
worldofbeys If you are trying to scan a full object the only way I know of would be to somehow support it in the air so that you can take pictures from all sides. The other option is to make two scans of the top then the bottom then merge them together in 3D software. For something like that it might be easier or better to just model it in 3D.
You sir deserve my sincerest regards! You've just opened a whole new opportunity for making a living out of what I love! I can go back to my passion - the landscape photography and making money out of it, without being super lucky working for Nat Geo :D . Adios to the studio portraits, the children's birthdays and weddings! YOU ARE MY HERO! *Deep bow*
Brilliant, good thing I stumbled upon your channel :)
Great tutorial on how to craft skillfully with free tools! Thanks for explaining me how to fix those annoying scan holes which is always a problem even with professional software.
the future of 3d modeling of natural items. Dice already uses this a lot in their games, and they have the most stunning graphics of any studio. I was actually kinda dissapointed when I found out much of their natural art was photogrametry. I thought they had some superhero artists or something hahahaha.
Tomi Ajayi-Dopemu - the best way to have photorealistic games is to use real photos :D
"the most sunning graphics of any studio" lol
in this case the superhero artists are the software developers who can make this happen :)
It’s too bad they stopped making good games!
Thanks mate; Still rivals cheap "3d Scanners" over a decade later.
Oh yeah photogrammetry can do a really well job if you have a detailed model. Reality Capture costs a little bit but does an even better job and it's a lot easier.
This is impressive, I must say
From someone who knows nothing about this, I thought this was a great video! You have a nice style of commentary, and you seemed to show primary features and a good end result. Very good demonstration video! Thanks! Subscribed.
This video _ROCKS!!!_
Phil were experimenting using your 3D scanning with our drones, although it requires some work, it seems to be working great! Thank you for this! Great job
Robert Nicholson I'm glad you like it. Yes there are easier ways but they all cost more $$$. Haha
19:50 - I think that's "per-AM-it-er-ization", like the word parameter. "Are we operating within established parameters, lieutenant?"
ReverendTed That's quite possible!
Excellent Video. very nicely done.. I don't need to buy a expensive scanner. I already have a 3D router and Mastercam.. Now, it just gotta learn these softwares and put my camera to work. I love it. Thank you.!
Great Informative Video. Thank you sir
firstly I would like to wish you well and to say a huge thank you for uploading these videos as they have been an invaluable resource to
Thanks it's really clear and easy to follow.
Phil, thanks so much for putting this together!! The yard around two sides of my house needs improvement, but I'm the kind of guy that likes to draw things up in Autocad 3D first and play with ideas there before I start moving dirt. The challenge before me has been surveying the terrain (the hard part when I know next nothing about surveying) accurately enough to re-create it as a mesh in Autocad (the easy part). Enter photogrammetry and your handy tutorial that uses FREE applications! I did a trial run and believe I can create a more polished version that will serve my landscape design needs, without having to do the tedious survey work. Thanks Again!!
+Monte - Yeah this should work well for that. I think I mentioned in the video I scanned a hillside in the part once and it perfectly showed where all the rocks, trees, and grass were.
Hi Phil - I took another series of higher res photos and did another run of the landscape and side of the house. When I imported the mesh into Meshlab it made a beautiful reconstruction, but refused to cooperate when trying to do the surface reconstruction poisson. I suspect the dense point cloud is too much for it to handle, so I removed a bunch of the points, but it still won't do it. Is there a way to 'optimize' the mesh before attempting this operation? I'm using Windows 7 x64 with 8GB of ram, and a decent 1gb video card....
+Monte - Hmm. Sorry, I haven't seen that happen.
That is brilliant.
Thanks! It's the generosity of people like you who make the internet a great community. Thanks again for a very well explained video!
you rock!!
This was my first look at 3D modeling I was always under the impression you had to draw the model first then texture it. This gives me hope. Thank you for the video.
Well this isn't modeling, it's scanning. Like using a photo copier for physical objects. Modeling is more similar to building something out of wood (just on a computer).
Nice video, not many out there like this.
+John Odd Thanks.
Wow! I will be studying this video by watching several times and taking notes. Thanks so much for this post. BTW your calm speaking voice is a plus.
Thanks. I've made several other video tutorials professionally.
Hi, Thank you for your awesome video, I'm new to 3D scanning. I am following you step by step, but when I want to import my .ply file to meshlab , I couldn't find that file !!!!! I'm wondering where I missed ? could you please help me.
Tanya Shirazi When you start the CMVS part ( 6:21 in the video) you save out the file. I can't say where you chose to save it. I would also make sure that you have the latest version of everything. The older version of CMVS didn't save the file in the root where you choose, it creates a folder structure there like this \filename.nvm.cmvs\00\models\ (filename is whatever you picked). The problem with this is that in that folder it may create more than one PLY file and there's an extra step to merge them. The new version always creates just one file.
Phil Nolan Hi, Thanks for quick reply, I just download the last version which is available , as you said it creates the 00\models folder but nothing is available in that folder as well !!! could you please post the link that you are using, maybe I downloaded the wrong version. appreciate.
Phil Nolan sorry Phil, I downloaded cmvs and put it in the same folder as visualsfm , but nothing changed, I still don't have .ply file !!!Any suggestion!!
Tanya Shirazi The exact ones that I'm using are VisualSFM Cuda, x64 version:
ccwu.me/vsfm/download/VisualSFM_windows_cuda_64bit.zip
CMVS for WIndows:
github.com/pmoulon/CMVS-PMVS/archive/master.zip
Phil Nolan Dear Phil , now I can see two .ply file in 00 folder, could you please let me know how to merge them to use it in Meshlab ?
Wow this is a much better workflow than the one I follow! Thanks for the vid! I usually used agisoft photo scan, and then cleaned the mesh in zbrush, then UV'd in Maya, then brought everything back into substance painter, and finally into 3ds max, and from there into cryengine. You saved me ALOT of time.
Well as it is you wouldn't bring these directly into a game engine, the poly count is too high. When I'm finished here I normally bring it into 3D-Coat and adjust any sculpting, then retopologize it and UV, bake the textures to the new mesh , and touch up any painting. Still, that's only one additional program.
Yeah. Have you heard of memento,its amazing. It even decimates your mesh for you upon export, and bakes normal's, displacement, and diffuse.
Prithvi Boinpally it _can_ decimate the mesh but it my experience that's a little buggy. It is a beta though.
thank you for this informative video! great job!
Wow! this was an awesome video! You've got me interested in 3d scanning. Please keep making these videos !
Nathan Holbrook I actually have a new one in the works. In talks with the related company now.
UA-cam has cut what they pay creators by a lot. If you like these videos please consider becoming a patron. Patrons also get bonus content that isn't on my channel and I'll be sending out 3D printed minis to some patrons.
www.patreon.com/philnolan3DJ
With 4.5k subs you won't make much if any to start with. less than 100 a month.
Glassic Gamer I was making double what I am now. It was enough to pay my phone bill each month. Since they cut it I only get a check from them every few months.
it would be awesome if UA-cam integrated way to tip creators for a nice, sending 1$ would be the same as watching 500 adds (last time I checked).
Patreon also works of course.
Wemja They used to have a "tip jar" with a link to donate on the channel page but they took it away" Maybe because nobody used it. You could always try UA-cam Red, $10/mo for no ads. It doesn't go directly to a UA-camr though. I think like .5% of my revenue is from Red users.
Sorry to whomever's comment I just deleted by accident. You were asking how I get less than $100 a month from UA-cam. I don't. That's what I was saying earlier. I usually only get a payment from them every few months when that small amount adds up to more than $100.
Watched the full video. Very clear and entertaining. I work on Mac, but have Windows in Bootcamp, so I'll just work from there. Thanks, Phil!
I didn't know I was entertaining. Thanks. ;)
+Phil Nolan yeah, but I am a 3D nerd, so things like these are always entertaining. ;)
excellent-- many thanks
What an excellent tutorial! I'm not sure whether the technology has changed but you made the process so clear especially to novices like me. Thank you!
Mesh lab has changed slightly. I believe I noted it in the description. Otherwise I believe it's the same. These days I prefer Reality Capture. I have videos on that too. It's not free but very cheap.
Amazing!
Hi Phil,
I should have taken more notice of your description text below the video. It's all there. I have used the links, followed everything, and it is all working. Thank you so much for this video and the extra help you gave me !
I had to subscribe once I learnt this 😎
So did I!
Never thought a video about a rock could be so interesting :D
Since i'm completely talent-free when it comes to creating textures and 3D models, this might prove really worthwhile in the future. Thanks a lot.
is it usable for human body reconstruction?
TheAmazeer To scan a person? Sure. Loose hair and clothing like an open jacket can be difficult. But as long as you can get the person to stand totally still you could scan them. It would be best to have them in a T pose, perhaps with arms slightly lowered.
BROTHER, YOU ARE THE BEST!!! You oooh really helped me!! THANK YOU VERY MUCH!
Very Nice!
Thanks for this excellent video. It helped me doing my very first 3D scan which out worked amazingly well for
an STL out of a a bunch of cell phone shots.
My honest kudos to programmers of open source software like VisualSFM, PMVS/CMVS, meshlab, and OpenSCAD
which really make all the difference when creating 3D models and 3D prints for free!
Glad you liked it!
Thank you for the video.
Pokemon Go You're welcome. :)
Thanks, Phil. Not used this program or any like it but have extensive experience at photo manipulation software. When you were deleting the unwanted ground . . the icons buttons to the left and right of it look typically like alternatives to using the square cage to select.
it's so nice!
this is great! there are 3d scanners that are sold for thousands of dollars. Your technique's results better quality 3d scanning and it is free! thanks for sharing! :)
+numan duman No problem, glad you liked it. BTW I'm doing a giveaway starting tonight, be sure to check my channel for the video!
+Phil Nolan definitely!!
Nice video. THanks.
Very good looking result - and all for free ! Now I only need to know a free or inexpensive way to turn the mesh back into a real (scaled?) object again - I'm planning on making my personal shoe last !
Thanks for the information !
Thumbs up !
Thanks for sharing this....
This is so much fun, thanks! (I haven't even gone beyond Visual SFM yet)
Glad you like it!
Don't forget to check out the rest of the videos on my channel!
CogitoErgoCogitoSum Check the description, there's more info.
why so mad?
I'm not sure what you mean. Did I read it? I wrote it. The image is all of the files that come with it. If you see different files then you must have downloaded something else.
Thank you for sharing! The finished result is very impressive. Excited to try it out.
that's awesome stuff man... pretty intimidating to a guy like me... doesn't take very long to be way behind the times on software or any tech at all...thanks for the video.
Great video man! I appreciate the fact too that you didn't talk fast or "rush" through the video. A lot of people doing tutorials on stuff similar to this always seem to rush through it, forcing me to go back and watch a part several times...lol. I do have a few questions for you. 1. Can I use a DSLR camera for this? Manual or Automatic settings? 2. Could I import the file into ZBrush to clean it up prior to 3D printing? 3. Is it a must to shoot the pictures outside? I have bright daylight white "florescent" lighting in my room, which is why I asked. You're a life saver man as I was about to invest in a 3D scanner haha!
+Kirkland Shunk Thanks for the compliment. You can definitely use a DSLR, if you watch my other video, linked at the end of this one, I was using a DSLR. There is an older video that a guy did for another scanning program, where he mentions settings for a DSLR that seems to work well.
ua-cam.com/video/NsBg-m2hrIM/v-deo.html
I don't use zbrush for my sculpting but it should be no problem to import the model. I don't know if zbrush will accept the color though. I know in 3D-Coat I can import the model with color if I import to Surface mode with Voxelization turned off. It does not have to be outside at all, actually florescent lighting would be great I think because it gives very few shadows. I wish I had it for my own scanning. Some of my scans were just done in regular lighting though and it's OK as long as it has even light all around.
Phil Nolan Awesome! I actually did watch your video after I commented so I found that out right away haha. What I'm planning on doing is making sculptures of characters using Super Sculpey(which is a tan color), so I think ZBrush won't have a hard time with it, not sure. Yeah that's what I thought. Thanks for the information though! Really appreciate it! Subscribed as well!
+Kirkland Shunk Glad you got the info and thanks for the sub. Good luck with your project.
Thanks took hours of frustration away from what I was doing cheerz.
this is the best Mesh Lab teaching video I have ever seen
thanks very much Phil
+CP HO And thank you very much!
+Phil Nolan
I have downloaded and expanded the visual sfm windows cuda 64 bits with no problem. I have some problem with the cmvs-pmvs-master. It has several folders - binaries, license, program. Shall I put the whole folder (cmvs-pmvs-master) or shall I just put the windows 64 vs2010 into the sfm folder? sorry for the stupid question
CP HO I said in the video description. Extract the files from "BinariesWIn-Linux / Win64-VS2010" and put them in the same folder as Visual SFM.
+Phil Nolan
thanks a lot. I did add the binaries files. It worked. But when I press the tab after the reconstruction, only a poor quality picture. I do not know whether the problem is because of my pictures or the program. I shall repeat the program with your pictures. Shall get back
+CP HO Dear Phil
I have downloaded the viusal SFM windows cuda 64 beat and the CMS tool. I used your rock photos for a trail. the whole process ran smoothly, including the full 3D and dense 3D reconstruction and the CMVS tool. On pressing the tab key, a nice rock photo appeared. However, it is only 2 D and not 3 D, can not rotate is as in your demo. In fact when I was in full 3D reconstruction before CMVS tool, the rock was also in 2 D, cannot roatate.
Any idea what went wrong? Shall I reinstall the SFM? I have a cuda Geforce FX card.
thanks
Phil Thanks for this tutorial how can I work with the object in blender, what exported format does it work? have you done this?
I don't like blender but the export is OBJ which just about any 3D app will accept, including blender.
Amazing, I was always questioning if there is a simple way of doing a 3D scanning. Thank you for the tutorial, really appreciate.
never going to use this or do it but i have sat here facinatted. nice video
wow thanks for the video dude :O i didnt know this was possible with free software. So going to try it right now!
will get it. Just don't get burnt out. Whenever you need a break, take one.
thats pretty amazing for just a bunch of photos. great work you got a new sub for sure cant wait for more.
Thanks Phil, I was into 3d mapping and I ran into this. You are awesome!
Thanks for making this video. I looked at the SFM web site and I was very happy to see that it will run on Linux. I haven't checked the other software yet.
Anyway, this looks really excellent. I just started with 3D modeling in Blender. This SFM and CMVS could really help me to build up a library of original models to use in animations.
Now I'm going to watch your newer video. :D
EdWittenable Good luck, I've only used it on Windows so I assume my tutorial will work for you too.
Hi Phil!
My profession is connected to the conservation of cultural heritages. This was a great tutorial, thank you!
+Miklós László Pap Oh yeah, this would be very good or that. Good luck with it.
Hey greate tutorial! One thing I found out is, instead of using/taking a bunch of photos, you could just film your object, and then in for example in Adobe Media Encoder render the movie to a jpeg sequence, depending on the framerate you are using, you could also change the number of images really easily. Just something I found out recently :)
Neon Artworks I mentioned this in the description. People ask about doing it over and over. lol I think still photos are higher resolution for better quality.
Super helpful video man! Thanks so much! I have a photography background so this will be awesome to apply to my workflow!
great tutorial thanks. As a follow up, can you describe how one might provide fairly accurate and/or accurate dimensional measurement to scans? (1:1 dimensions). What are different ways this could be done? If you took measurement(s) can you apply them? Or if you had ruler object in photos etc ? Does focal length affect proportions and make it harder for software to work correctly , if FL was not consistent? Or does it kinda work out anyway because of how it puts it together in 3 dimensions? Could you 'safely' change FL during shoot for you convenience etc?
Sorry that's a little beyond my knowledge. You could definitely put a ruler in the scene or just any item which you know its size, then adjust the final model's scale to match.
Thanks, ruler tape stuck to object if wanting most accurate version of that I suppose. Or maybe on ground would be same accuracy idk. But is keeping the same focal length kinda important in general? I wonder if there might be some guidelines about FL vs distance to subject for more accurate/proportional results. I haven't tried this yet but I could find out for myself by taking multiple sets of photos, one with a radical range of FL, a couple others with different fixed lengths. I good size symmetrical object that had some smaller symmetrical features would be a good test I suppose. I wonder if a longer lens would end up having the software 'think' the camera was closer than it was in shots with shorter length. With most zoom cameras like mine its hard to document the length setting since is not indicated., and you'd have to be careful to keep away from the zoom control. That is of course if that mattered much, which it may not.
I kinda guess that some kinds of fencing is really hard to do?, (possibly epically hard) But a small enclosed fenced area of chain link fence would probably tell a lot about accuracy. (like a 4x4 foot enclosure of 4' tall fence). Curious if you have experience with screen-like subject. Does it make larger file size in software than 'normal'?
Very nice result! beautiful texture!
Don't know how I got here but absolutely worth it 😀 thank you
Thank you so much for such an elaborate video. I am getting into aerial mapping with drones and this software wont break the bank like Pix4D. Going to mess with this tomorrow :)
Awesome tutorial mate! I been wanting to try this stuff out for ages. You explained everything in perfect detail and I managed to get from taking the photos to the finished model on my first ever go at this. Subbed and Liked :)
Glad you enjoyed.
Great video. But what do you do when you end up with two pointcloud models in VFSM? I took two good sets of photos around a little figurine. One from an even level and one from a higher level. And I did everything up to 7:30. I ended up with two models. One model made from the set of photos taken at the lower level and one model made taken from the set of higher leveled photos. Do you know how this can be fixed?
You are dope bro! Damn , thank you so much, subbed
Thanks Phil!!This is awesome! Can one scan the models at a lower density/lower poly, but still high quality (Like low poly models) If i will use 10+ models in my scene like this it with crash, so looking at making it lowish poly but realistic looking...?
There are some decimation options in Meshlab but i have not used them so I couldn't say how they work. I really think the best option would be retopology, as I mention in the video. IMO the best option for that would be 3D-Coat, though there are some other programs that do it too.
Phil Nolan cheers Phil!!
First off, thanks for the tutorial, it was well put together and easy to follow. A few questions/comments/ideas:
** Do you think putting the item on a turntable or swivel chair and rotating the item would help?
-- Wouldn't it be easier, you'd have more consistency with distance from the object as well as intervals around the object?
-- I could see inconsistent shadows/lighting causing an issue; as you turn the object, the shadows would change, but if you're lighting was good maybe it wouldn't be too bad?
-- I'm also not sure if the unchanging background and camera position would cause any issues when creating the point cloud. Do you think if the background was all white it might work better or do you think this would even be an issue at all? I assume that VisualSFM generates the camera position for each shot so having a stationary camera shouldn't be an issue.
** What effect, if any, do you think using a phone camera over an actual camera would have?
** Do you think using a burst shot feature on a camera would work as well as normal single shot? What about taking a video of the object and then dicing it into frames using VLC (or other software)?
** I'd like to use this 3D modeling technique to get a 3D model of my head to put into videogames; right now it looks like uraniom (uraniom.co/) would be the way to go. Uraniom says that it's compatible with zip archives containing three files (.obj, .mtl and .jpg files); are you familiar with what type of export that would be?
Thanks, and sorry for the question dump
-All of these questions have been brought up several times in the comments below except the last one. From what I understand from people who've tried it the turntable inside a light box with no shadows and white background can work but the software uses the background so it's better to have the BG moving with the object.
-Phone camera works fine for me, that's what I usually use.
-Burst might be too many pictures, I would think, sounds like a waste of smemory to me.
-video... I think it would be hard to keep it in focus and get everything composed right.
-As for OBJ/mtl/jpg, those are the files that are exported at the end of this so it should be fine. I personally use Collada (DAE) file but that's just because I'm going to edit it further. You could be able to export OBJ (which includes MTL).
Very cool demo. Thanks so much. Can you recommend 3D modeling software? In particular, I'm interested in being able to output a series of frames, rotated precisely in a sequence about a third of degree per "frame", from a particular point of view, or "camera location" in grayscale format.
Thank you Mr. Nolan. I thoroughly enjoyed that video. I had no idea that kind of detail could be had with free software. Are there any special settings that must be on the phone when taking the pictures?
Time Lord you're welcome and nope. I just pull out my phone and start shooting like any other pictures.
Great video. I use LW and 3dCoat also. Cant wait to give this a try. Thanks!
Just a tip for any time you are doing heavy CPU processing, but want your computer to continue to preform normally (like in the case of your audio glitches) If you have a multi-core processor (every processor these days are at LEAST dual core) navigate to the process that is hogging all the CPU usage in Task Manager under the 'Processes' tab. Right-click on the process, and choose 'Set Affinity'. This will let you select which CPU cores that process is allowed to use. Uncheck one of them... and now you computer will have one core free to to be used for other things, restoring your computers responsiveness back to normal.
I actually have 16 cores so I haven't thought about the affinity setting in a long time.
Phil Nolan 16 cores - Nice! Even better. All the more reason to uncheck one of the cores. A single core wont make much difference in your processing time, but being free will make all the difference in your computers responsiveness if you still need to use it for other things. When a CPU intensive application is well designed, it doesn't matter how many cores you have... it's going to suck ever last bit of processing power your computer has, as it should.
+FPEvideos Well, I didn't notice any hit to my processing power while it works. So while a good tip over all I just mean that it isn't affecting me personally.
That was great Phil, thanks for sharing. Got to give this a try! :)
Hey man I just want to say thank you and that this is an awesome tutorial! Kudos subbed!!
Wonderful tutorial, and all with free software, thank you!
Thank you so much this helped a lot!!!! You saved my life
Exactly what I was looking for! Thank you!
+zcqw4a1 Glad you like it!
Hi phil, awesome tutorial. My name is James and Im really interested in creating 3d models ( mainly figures ) to import in to Blender for rigging and skinning and then exporting for use in Iclone 5, my chosen platform. I just wondered if its possible to use the Kinect sensor for the XBOX 360 as the sensor or will be SLR suffice? I looked at Skanect software which utilises the motion technology of kinect to provide the data, which was interesting but as i'm only requiring static 3d models I thought this might be more applicable. Any advice you can give me on this would be great. Keep up the good work sir..
James Harvey I have a Kinect that I've tried before I discovered this technique. It gives very, very low quality scans. I would recomend using the techniques in this video, photrogrammetry, instead. pretty much any digital camera will work. In this video I used my phone.
Super video. Really well presented. Two questions. For areas of the rock with limited detail, could you go back, retake pictures and incorporate them into an existing model? Second question, can you export into Cinema4d with this software?? Thankyou
+Nick Kingstone Thanks Nick. In theory I guess you could go back and shoot more photos, then start the process over. However in cases like this rock, it might be tricky since the lighting would change by the time you went back, especially if it was a different day. As for export. Sure, it exports as OBJ format, which almost all 3D software accepts. I don't use C4D but I would be very surprised if it didn't take such a univresal format.
+Phil Nolan Thank you for your reply
Thank you Phil for the spectacular tutorial, I love it. I wonder if taking the picture for the rock while it sits on a glass penal would made it easy to get rid of unwanted background meshes? Ideally this should create a totally separated model from the bacground.
Yeah maybe. I never thought to try that.
I'm student I've been using maya trust me I just saw this first time and you just blown my mind I'm gonna try it sure thanks for sharing
kishor vandan Good luck! Maybe with some practice you can scan an action figure and animate it!
Phil Nolan yay nice suggestion thank you ;)
Outstanding, very nicely explained, and an awesome project for printing in considerable detail. Well done.
Thanks, glad you liked it.