Closed mairaksi closed 8 years ago
1) I think that's fine. Just print a message that tells how many LSF vectors needed fixing.
2) I think sorting is a better approach. The averaging approach makes the assumption that the disordered LSFs are computed/generated close to each other and the values should be within their mean range.
There seems to be a problem with this function. It stays in a perpetual fixing loop (using WLP).
Fixed the bug, added sorting and modified the error message.
Added StabilizeLsf based on the old Fix_LSF code. A couple of questions:
1) I removed the verbose warning messages (what type of LSF problem was fixed)
2) The old code handles non-increasing LSFs by taking their average and shifting the two LSFs slightly apart. Do we want to do this by sorting instead?