LinAlgErorr: SVD did not converge

[lrjam]

I'm getting the following error when attempting to extract glottal features from voiced segments (pasted below).

Wondering if anyone has encountered this issue, and has any solutions? Thanks in advance!