Playground Posted May 31, 2018 Report Share Posted May 31, 2018 Hello I have a large mapping project coming up. We have been doing extensive tests on a 1/10th scale model of the scenic piece. I have mapped on cars in the past using the 3d OBj technique. So I am familiar with the approach. However I am finding it very difficult to align my projections to the set. I can align some of the projection but not completely snapped perfectly. I spend hours tuning the projections and find It less than perfect all the time I find I don't always understand what calibration points interact with what. I am not able to isolate small areas and lock them in place without skewing the projection some where else. I am wondering if other users have tips and tricks for doing this. How can i use the numbers in the position window to my advantage better. We are hitting the set piece with 2 4K 35000 lumen projectors. I am unable to attach a rendered png of the object but it is a series of "boxes" with a 11 by 20 flat screen inserted. Any tips are welcome Ben Chaisson Lead Watchout engineer FMAV central Branch Toronto 0 Quote Link to comment Share on other sites More sharing options...
Dataton Partner Walter Posted May 31, 2018 Dataton Partner Report Share Posted May 31, 2018 Hi Ben, you state the object is a series of boxes... so not a separate object for each box? I’d go with separate elements and, if needed, split up your source content in watchout using virtual displays to map onto the separate boxes. Would give you exact control per box. 0 Quote Link to comment Share on other sites More sharing options...
Playground Posted May 31, 2018 Author Report Share Posted May 31, 2018 Hi Unfortunately that's not an option. The OBJ file and the flattened video files are created by the clients content team. I really am just looking for advice on where and how to pick my calibration points. I seem to need quite a few before calibration actually works but then it seems to interfere with itself? 0 Quote Link to comment Share on other sites More sharing options...
Dataton Partner RBeddig Posted June 1, 2018 Dataton Partner Report Share Posted June 1, 2018 The calibration points for the alignment of the texture in a 3D projector do not allow any isolation of areas. The points are actually used to triangulate the position of the projector in relation to the position and orientation of the 3D object. We always start with 5 points first and align those since up to this point the texture will stay unmoved and the projector will not "fly into space". Then we add the 6th point which will start to move the texture. The position of the texture itself depends solely on the information in the 3D (OBJ) file. This is defined in softwares like 3DS, Blender, Cinema 4D etc. If the 3D file is not accurate, i.e. it differs a bit from the real world object, the alignment won't be perfect at this point. WATCHOUT sees in the 3D file where the texture needs to go and not in the real object. You can optimize the calibration by using a few (try to keep the number as low as possible) more calibration points. They do not need to be at exact corners or so since they are then just used as handles to drag the texture from the correct position. This only works with small corrections. If you drag too far, WATCHOUT will tell you that the texture can not be calculated in relation to the geometry of the 3D file anymore and show red lines. You can also use the standard grid for correction. In this grid you can easily isolate a region and prevent it from spilling over. Just place 4 points in a rectangular way (same lines vertical and horizontally). Then you can add extra points inside which will not influence anything outside the box. Of course, you can break your texture down into smaller regions as Walter suggests. This can help too. We usually ignore the numbers in the position completely. You can use them to preposition the projectors in 3D space but they will change automatically once you have 6 alignment points and start to move them. Important: on the general tab you have to define the vertical and horizontal lens shift. If you do not have correct values in there (lock them after defining them) the calibration will not work accurately since it makes a difference for the underlying formulas whether the lens is shifted or not. Purely optical physics! 0 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.