Wrong tonic identification result returned in joint analysis about dunya HOT 6 CLOSED

This Mbid are not recordings they are works, so they don't have audioanalysis nor jointanalysis.

from dunya.

from dunya.

@andrebola, sorry I couldn't write earlier because I was traveling. I couldn't run the checks for the complete audio collection yet, but I wanted to provide an example in the meantime: http://dunya.compmusic.upf.edu/document/by-id/5e2153e4-03fc-47a7-b91c-67c085ada644/jointanalysis?v=0.1&subtype=tonic

This is the output of audioanalysis tonic identificaiton method, which somehow propogated into the jointanalysis: http://dunya.compmusic.upf.edu/document/by-id/5e2153e4-03fc-47a7-b91c-67c085ada644/audioanalysis?v=0.1&subtype=tonic

from dunya.

@sertansenturk looking at the extractor, I guess they have the same value due to some failure on jointAnalyzer. Can you check if this code it's right?
https://github.com/MTG/pycompmusic/blob/master/compmusic/extractors/makam/jointanalysis.py#L95-L100

or should we store the output of summarized_features?

from dunya.

I found out that the problem is in the score-informed tonic identification in joint analysis. Somehow the tonic value is not returned but the alignment etc. are done without a problem. It's quite strange.

I couldn't pinpoint the problem yet because the problem seems to be on the MATLAB side. I'll probably investigate it thoroughly next week, when I return to Barcelona.

from dunya.

Since it's a tomato related issue, I moved it to: sertansenturk/tomato#73

from dunya.