Potential data precision problem

[arnodelorme]

I noticed that an exported BV marker file with bva-io has a bug in Mk1 as follows:

Mk1=New Segment,,1,8.111800e+04,0,0

This prevents BVA-IO to re-load it again (and could have caused some other troubles elsewhere...)This is probably because of resampling which changes the data starting point (this is an old bug which has been causing several other bugs, isn't it?)

My solution was to overwrite 8.111800e+04 with 0. I confirmed that the modified marker file was recognized fine and could be imported by BVA-IO.

Makoto

[SvennoNito]

I just ran into the problem 7y later.
My fix was in parsebvmrk.m

Namely changing

[mrkType mrkDesc EVENT(idx).latency EVENT(idx).duration  EVENT(idx).channel EVENT(idx).bvtime visible] = ...
strread(MRK.markerinfos{idx, 1}, '%s%s%f%d%d%d%s', 'delimiter', ','); 

to

[mrkType mrkDesc EVENT(idx).latency tmpDuration tmpChannel tmpBvtime visible] = ...
strread(MRK.markerinfos{idx, 1}, '%s%s%f%f%f%f%s', 'delimiter', ',');
EVENT(idx).duration = round(tmpDuration);
EVENT(idx).channel  = round(tmpChannel);
EVENT(idx).bvtime   = round(tmpBvtime);