I'm having a scaling problem when using DEMto3D and extracting STL's for later manipulation.

I have a merged TIFF file for Australia and am attempting to extract models of each state so I can print out. I want to print each state individually and mount them on a frame later so the scale of each model is important.

CRS for everything is GDA94.

My workflow is as follows.

  • Use clipper to extract state raster (using state boundary shapofile) and input as a separate layer.
  • Use Demto3d on each layer to extract the states to a STL.

Whilst I can get the models saved as STL without problems, when I try to assemble them in a 3d modelling program, they are about 5-7% different in size and do not line up.

Checking each TIFF file and overlaying in GIMP shows them aligning perfectly. DEMto3D settings are the same for each model as are the clipper settings.

There must be a projection or scaling issue but I cannot figure out, where exactly I went wrong. Nor understand why because everything I can find is set for GDA94.

  • using GDA94 might actually be the problem; a geographic projection of a (geocentric) datum simply projects the curvature of the spheroid (the lat/lon values in degrees) onto a plane, resulting in 'distortions' growing non-linear with the distance to the center coordinate. a datum is not meant for planar cartographic visualizations; try reprojecting your initial data to e.g. the GDA94/Australian Alberts projection (EPSG:3577) for a uniform minimal planar distortion over the whole continent.
    – geozelot
    Feb 4, 2019 at 10:29
  • @ThingumaBob, The reprojecting made no difference to the generated STL files. However I have since discovered the scale error is likely due to differences in what reference DEMto3D uses when converting latitude and longitude co-ordinates to meters and saving them to the file. Demto3D requires you specify latitude and longitude, but the output of the file is in mm so some conversion system is being used. I am seeking feedback from the author and will post up if I find a solution that works Feb 5, 2019 at 2:47

1 Answer 1


I reached out to the Author of DEMto3D and received a response that addressed my problem. I repeat his solution here for reference:

"The difference that you notice is due to DEMto3D adapts the distance of the model in degrees to the max/min latitud of the data. It means that the escale in the max latitude is quite different to the escale in the min latitude as the surface of reference is not plane is an ellipsoid. To avoid this efect I recommend you to reproject te data to a system in meters. In this way the relation between the size of the model in milimeters is direct to meters.

Reprojecting the data using a projection that utilises meters resolved this issue. The original answer by ThingumaBob ended up being correct. I am still unsure why my attempt (and hence comment) failed the first time, however I put this down to inexperience.

Thanks for the assistance.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.