3D model exporting is coming to Trimensional v1.02. For many 3D artists and 3D printing enthusiasts, this is the feature that turns Trimensional from an interesting toy into a useful tool, so the plan is (and always has been) to make exporting available to these advanced users as an in-app purchase for a few bucks. (Note that this is the only feature I ever intend to charge extra for, all other improvements will come free of charge.)
Based on feedback so far, OBJ is a format that would work for a lot of people. If that’s not the case for you, please let me know in the comments. As pictured above, I have some example output here (untextured at the moment) if anyone would like to test it out or get a feel for the resolution or how noisy the raw data is: http://www.trimensional.com/models/Trimensional_15.obj.zip
Any reactions at all are useful to me in building the 3D model exporting feature, so feel free to post any thoughts about whether this output suits your needs. If I understand correctly, the 3D printing folks need airtight models. Boxing in the back of the model is one approach — is there something better?
Thanks, and I look forward to a constructive discussion.
This is great! I had a similar idea to use an iPhone as a structured light 3D scanning device. I wouldn’t have thought that you’d be able to get anything out of just holding the screen to the object. My idea was to place the iPhone in a box with fresnel lenses to project the pattern more clearly. It looks like you’re doing a kind of structured light scanning that I’m not familiar with – instead of projecting lines, you’re illuminating the object from different angles? It would be awesome if you added different pattern options that might be more useful with a DIY projector-like setup or even for people with pico projectors so that all the scanning could be done on the iPhone instead of requiring a computer to process the images.
So far as model exporting options go – OBJ would be fine, STL would be even better. Also it might be useful to be able to export just a point cloud for me to mesh myself.
As a future wishlist – I’d like to see the ability to scan an object from multiple angles (perhaps placing the object on a turntable) and have Trimensional combine them into a single object.
Hi Tony,
Thanks for the links to your iPhone scanner idea — I was previously unaware of your plans but they seem feasible so I’d still love to see someone try a method such as yours. Trimensional actually was born out of a paper I published in 2008 on using computer screen lighting for reconstruction. When the iPhone 4 came out with the screen and camera on the same side, it was the right time to try it on a smaller device. And you’re right, the method I’m using is related to Photometric Stereo.
Thanks for the input on file formats, and yes, fusing scans from multiple angles would indeed be a great idea for the future.
Pingback: Tweets that mention 3D Model Exporting Coming to Trimensional | Trimensional Blog -- Topsy.com
So the in-app purchase will be for the ability to export, not charge per export, correct? Just wanted to make sure. 🙂 Also, when the ipad eventually gets a camera, will you make this an universal app? I’m so excited to try this, but I may wait if you’re going to create a separate ipad version.
Finally, as far as exporters, why not either, include a bunch of types in one in-app purchase, or do in-app purchases for the type the user wants? FBX, OBJ, 3DS, STL in that order would be my preference. Will the exporting be done through email, itunes, both?
Hi Chip,
Thanks very much for your input! Don’t worry — it’s a one-time purchase of the ability to export 3D models, after which you can export as much as you want. Export is implemented via email in the latest build. I’m leaning toward multiple export types in one in-app purchase so that I can start small (OBJ, STL) and expand the list over time without much hassle.
I’ve made no decisions about a universal iPad version yet, though you can bet I’ll be buying one as soon as they put FaceTime (and therefore the front-facing camera) on it in order to start development and testing.
Great looking App!
As for output geometry files go, most 3D apps can import OBJ, STL, DWG/DXF meshes.
The Apps I use can import far more of these too and translate to other formats as well, so my advice is don’t get to hung-up on output.
If your concern is for a non-professional user, then OBJ would be a good start since most 3D printing services can make use of those file types.
Lastly, I have a iPhone 3G. Is there a way to use TRIMENSIONAL as one poster suggested, inside a light sealed box with lighting applied inside? Getting the iPhone4 just to do this is a little overkill for me.
This is an outstanding use for the iPhone. Congratulations!
I’m an archaeologist and the ability to have a ‘take anywhere’ object scanner (even at relatively low resolution) is thrilling.
On exporting, OBJ is fine but I could really make use of just the point cloud – just a csv file with the 3d coordinates of each point. Optionally, three additional coordinates for each point giving its RGB value would mean you build your own colored meshes in something like Meshlab. This would be tremendous. Finally, the ability to export the depthmap or heightfield image of the data would be useful as it would allow modelling via displacement mapping in other 3D software. So, OBJ, (colored) point cloud, and the heightfield would turn this into a tool I would use constantly.
One issue with the current version is the ‘big nose’ effect due to the distortion of the tiny camera lens. I do photogrammetry and can generate a calibration file based on my iPhone that would give you a good idea of the lens’ distortion. Based on that you could remove the distortion during your image processing and reduce the ‘big nose’ and greatly increase the accuracy of the point clouds. Perhaps cut down on noise a bit too.
If you’d like a calibration file, let me know.
I’ll be keeping a close eye on this as you develop it.
Well done!
Thanks! It’s great to hear that there are even more uses than I had imagined for this technology (and its future iterations). Output format preferences are duly noted.
You are absolutely correct that lens distortion is not something I’m taking into account yet, which leads to the effects you’ve noted. There are currently a few such unaccounted-for factors which leave room to increase the accuracy of scans over time. Thanks again, and I’ll be in touch with you about calibration information.
Awesome program – being able to export the scans to 3DS Max would take this from a cute toy to something amazing 🙂
I hate to drip on this parade, but as far as being a useful tool goes…
…well if the models we will end up being able to export are as noisy and sloppy as the sample you published, I’m afraid it will still be more of a toy then a tool. Even after cleaning it up a bit in my favorite 3d Modelling app, it’s still more jagged than the grand canyon.
Sorry to be so negative. I would love for this to work well. I’ll be keeping up I’m sure.
Thanks for this perspective — part of my motivation for posting the example was to get exactly this kind of feedback. I do expect the quality of scans to improve with continued development, both by taking into account additional camera parameters (as suggested above) and by performing post-processing on the raw data.
I’m considering the inclusion of a variety of export options, including the choice of raw or smoothed output. This way, those who want to do their own clean-up can do so, but those who want smoothed models (at the possible expense of some detail/resolution) have a ready-to-go export option as well.
I’d be curious to see the results of your attempted clean-up of the model in order to compare to the processed output I’m producing. Thanks again!
Grant,
Kudos for the concept and the execution!
I would like to have to the data exported as a point cloud.
That way, it would be possible for programmers (like myself and many others) to mash-up the acquired data again and again.
And if we talk wish-list: I would also like to have an option to scan from the front camera. Without such an option, I find it difficult to scan arbitrary objects.
Great Stuff!!
I can just only stand in the line and ask the same as the rest.
– Obj export via. file file sharing (mail is too slow but of course useful non the less)
– Combine multiple scans into 1 for increased quality (and lens correction)
– Be able to mass delete (or select) which scans i want to delete in the gallery section.
– When I delete in the 3D view it should not jump back to the previous scan. It should skip forward so I can continuously delete scans 🙂
– If possible similar smoothing as the Laplacian HC in Meshlab
– The setting for wireframe/shaded toggle could be moved to the info page where the rest of the options are (took some time to find it the first time)
Keep up the good work!
What happened to the app? No longer available in App Store. And no blog entries or comments in more than 2 weeks… Neat concept….
Thanks, I’m still here and Trimensional is still on the App Store. I’ve heard reports from iPad users that it doesn’t show up on the iPad app store, so perhaps this was what you encountered as well?
You’re right that I should post more often here, though I’d like to have something of substance to say 🙂 I’ve been developing the next update over the past month and it should be ready soon.
As a 3D printer owner, you’re right in that closing off the back is important. I’d imagine that just squaring off the back is sufficient. If you want to get really fancy, maybe an option to contour the back the same as the front (e.g. all back points are just the front point minus some value on the Z axis), but I’d say this is only useful as a default-off option (e.g. if you were going to incorporate it into a larger model or do some other sort of artsy thing to it) and most people would be best with a flat back. You can probably accomplish the same sort of thing by hand by extruding a non-closed model, but I don’t know how many think to do that.
OBJ is fine. Blender can import that and then export just about any other format. For what it’s worth, most people share their 3D printer objects as STL files (which Blender exports). I expect most people with printers are going to fine-tune in Blender (or some other program) anyway, so I don’t see this posing a problem.
Great work!
hi, juz installed the apps, however, i do find the features such as “select the 3d rendering modes” nor “rotate and zoom your 3d modes” as shown in your apps description, please advice?
Hi Herman,
When viewing a scan, touch the bottom-left “action” button. Underneath “Email Image” and “Save to Camera Roll” you’ll see the third button “Toggle Render Options”. This puts two buttons on screen that let you choose the type of coloring (Textured, Gray) and geometry (Points, Polygons, Wireframe) of the displayed model.
You can also drag a single finger back and forth to rotate the model, or use two fingers to perform a standard pinch-to-zoom gesture. I’ll have to make these options more clear in a future release or on a FAQ page 🙂
Thanks for downloading,
Grant
Are there any plans to be able to stitch together scans from multiple angles? I’m interested in scanning small tabletop models such as chess pieces.
Stitching is definitely something I would like to include in a future release. Thanks.
I am also interested in stitching multiple angle scans. Let us know if there are any advancements and many thanks for the work so far. Very impressive to have this technology so accessible 🙂
Ditto on Tony B’s feedback. OBJ format’s fine, since I can take that to a CAD editing software and make refinements from there and convert it to an STL that I can send to a 3D printer. Resolution’s pretty noisy, though. Try scanning a simpler object and see how the scan looks like. Being able to scan 360° would be awesome.
This look great, well done.
The scans look distorted by the lens – is there any plans to correct this in a future release – or any tools that can do this?
I have tried the Autodesk PhotoFly cloud app, which works pretty well:
http://labs.autodesk.com/utilities/photo_scene_editor/
Oops, sorry I let your comment sit here for 1 month unanswered! Yes, I have plans to improve the results and minimize distortion, so stay tuned. That cloud app looks really nice — I’ll have to give it a try. Thanks!
I bought your app. How do I select different renderings?
BC
When viewing a scan, touch the bottom-left “action” button. Underneath “Email Image” and “Save to Camera Roll” you’ll see the lowest button “Toggle Render Options”. This puts two buttons on screen that let you choose the type of coloring (Textured, Gray) and geometry (Points, Polygons, Wireframe) of the displayed model. Thanks for trying the app!
Meshlab is a free download that is used to patch up messy meshes. Can export from it into lots of different formats too
What about importing image into Sculptris, that might work