Hi Gael,

The conversion from Matrix3 to quaternion you use returns a non-normalized quaternion. I can see from the code (Quaternion.h, line 784 from sha f3137d73 which we use) that you are using Ken Shomake's code (available in here http://www.cs.ucr.edu/~vbz/resources/quatut.pdf) which is all good but although Ken Shoemake states in his document that the quaternion is unitary, it is well know that it is not in the general case and that the output quaternion should be normalized.

You can check this using matrix:

V = [ 0.42303329706192017, -0.1743864119052887, 0.8891751170158386;
-0.13570299744606018, 0.9580454230308533, 0.25245535373687744;
0.8958948850631714, 0.22746068239212036, -0.38162028789520264 ];

What do you think?

Kind regards,

Antoine.