[FieldTrip] Problems with "filter_with_correction"

[maxime ferez]

Dear all,

I have a question about Fieldtrip artifact-rejection. I find myself having
a problem that I do not know how to solve. I am currently working on a pc a
little old (~ 5 years) and my analyzes worked properly, or at least I had
no bug. When I changed my pc the problem has arrived. I have an error
message "Calculated filter coefficients have poles on or outside the unit
circle and will not be stable. Try a higher cutoff frequency or a different
type/order of filter" that appears in the "filter_with_correction" function
.The problem seems to occur during the "roots" function of Matlab. I
verified and the vector that is introduced is exactly the same in both
cases, but after the function “roots” runs the result differs slightly and
induces the error message above. Has anyone ever had this problem or know
where the problem might come from? On both pcs I have exactly the same
versions of Matlab (R2017a) and fieldtrip (fieldtrip-20171004 and
fieldtrip-20180626). Scripts and data are also identical in all cases. I
have tried several configurations but failed to find the solution:

    - Ubuntu 16.04 machine about 5 years -> OK

    - Ubuntu 16.04 machine about 2 years -> Not OK

The two machines above are identical in all softwares (deployed with script
in PXE)

    - Fedora 27 machine about 2 years -> Not OK

    - Ubuntu 16.04 about 2 years -> Not OK

    - Fedora 17 about 8 years old -> OK

    - Centos7 machine less than a year -> Not OK

To conclude, the script only works on old hardware but this does not seem
logical as an explanation.

Best regards,



Maxime Ferez


-- 
Maxime Ferez, ingénieur d'étude
Centre de Recherche en Neuroscience de Lyon (CRNL)
Inserm
équipe Dycog
Bron, France
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20180703/6eacc1fe/attachment-0001.html>