[FieldTrip] ft_freqanalysis warning (correcting numerical inaccuracy in time axes) after ft_redefinetrial on continous data

Jean-Baptiste Eichenlaub jb.eichenlaub at gmail.com
Fri Feb 16 10:38:02 CET 2018


Dear Fieldtripers,

After having segmented continuous EEG data (ft_redefinetrial, cfg.length  =
2; cfg.overlap = 0.5), I compute the power spectrum for each trial/segment
(ft_freqanalysis).
However, I am getting the following warning message (see below in blue).
Based on several sources, it seems to be explained by "the inaccuracy in
binary representation of floating point values". I do not get the error
message with cfg.length  = 5 in the previous step (ft_redefinetrial).
Should I be concerned?

Thanks!
Best,

Warning: correcting numerical inaccuracy in the time axes
> In ft_warning (line 184)
  In ft_datatype_raw>fixtimeaxes (line 310)
  In ft_datatype_raw (line 262)
  In ft_checkdata (line 236)
  In ft_freqanalysis (line 217)
  In DefineSleepPeriodsBlock>computeSpectrumAveragedChan_fieldtrip (line
237)
  In DefineSleepPeriodsBlock (line 91)
Warning: correcting numerical inaccuracy in the time axes
> In ft_warning (line 184)
  In ft_datatype_raw>fixtimeaxes (line 310)
  In ft_datatype_raw (line 262)
  In ft_checkdata (line 236)
  In ft_selectdata (line 134)
  In ft_freqanalysis (line 232)
  In DefineSleepPeriodsBlock>computeSpectrumAveragedChan_fieldtrip (line
237)
  In DefineSleepPeriodsBlock (line 91)
the call to "ft_selectdata" took 1 seconds
processing trials
processing trial 1199/1199 nfft: 2048 samples, datalength: 2000 samples, 1
tapers
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20180216/7f0a1b5c/attachment-0001.html>


More information about the fieldtrip mailing list