"Transformation Matrix not invertible"

[Godzilla]

Upon a quick search for this bug on the forums, it seems Whaat had the same issue a few months back, but a solution was never found as far as I can tell.

Anyone have any idea what causes this error and how I can fix it?


Blendigo 2.3.1
Indigo 2.2 Stable

[galinette]

Could you post the igs source? If there really is a non invertible matrix, this may be an exporter bug

Etienne

[dougal2]

this may be an exporter bug

Either that, or the exporter is simply 'passing-on' some bad geometry.

Trying to isolate the object in question might help shed some light on where the problem is.

[Godzilla]

this may be an exporter bug

Either that, or the exporter is simply 'passing-on' some bad geometry.

Trying to isolate the object in question might help shed some light on where the problem is.

Well, I found what the bug was - when a mesh had a particle system applied to it which emitted another mesh (grass), one of the particle options caused this bug.

It could be a bug with the exporter, though-

When the 'Rand' (Random) option in the blender particle menu, under 'extras' was set to 1.00 ( the max) , I got this error, but when I set it to 0.99, the error did not come up.

[dougal2]

When the 'Rand' (Random) option in the blender particle menu, under 'extras' was set to 1.00 ( the max) , I got this error, but when I set it to 0.99, the error did not come up.

I'm guessing here that when using exactly 1.00 some matrix elements may become 0.0 or inf or NaN, whereas at 0.99 you will get an extremely small or large number - probably as the result of a division somewhere.

Interesting that that can happen - posting here the exact faulty matrix from the IGS would help to identify why 1.0 is a bad value to use.

[Godzilla]

Sorry but I no longer have the IGS file..