Horace-3 file was storing ubmatrix wich was describing connection between pixel coordinate system and image coordinate system in Horace-3
When Horace-4 founds Horace-3 data, it tries to restore line_proj from this ubmatrix. It takes b-matrix from lattice and u-matrix by dividing ubmatrix by b-matrix. From u-matrix is used to recover u and v (and w) vectors for line_proj.
Problem occurs when legacy data are aligned. u-matrix in this case is production of rotation matrix, defined by u,v vectors and alignment matrix. Horace-4 makes assumption about nature of u-matrix to separate u-matrix and alignment matrix but these assumptions are not 100% valid.
There are no generic way of separating these matices especially in nonorthogonal== true case.
The soulution to support aligned data is introducing ubmat_proj which would be identical to line_proj except being constructed from ub_matrix directly.
This would allow to use it for old data recovery and work alongside with line_proj and other projections without any constrain.
Additional benifit of this approach would be possibility to define projection from goniometer (or ubmatrix) directly if somebody needs it.
abuts
commented
1 month ago
Horace-3 file was storing ubmatrix wich was describing connection between pixel coordinate system and image coordinate system in Horace-3
When Horace-4 founds Horace-3 data, it tries to restore
line_projfrom thisubmatrix. It takes b-matrix from lattice andu-matrixby dividing ubmatrix by b-matrix. From u-matrix is used to recoveruandv(andw) vectors forline_proj.Problem occurs when legacy data are aligned.
u-matrixin this case is production of rotation matrix, defined byu,vvectors and alignment matrix. Horace-4 makes assumption about nature ofu-matrixto separate u-matrix and alignment matrix but these assumptions are not 100% valid.There are no generic way of separating these matices especially in
nonorthogonal== truecase.The soulution to support aligned data is introducing
ubmat_projwhich would be identical toline_projexcept being constructed fromub_matrixdirectly.This would allow to use it for old data recovery and work alongside with
line_projand other projections without any constrain.Additional benifit of this approach would be possibility to define projection from goniometer (or ubmatrix) directly if somebody needs it.