
simplify/clean IO_read_frame: use  vec=FrFileIGetV(f,channel,epoch,fs*N);

remplace IO_read_ascii par IO_hdf5

in write_hdf5, add a preliminary test ... dataset exist?

add automatic documentation: use TeXinfo

plans for fixing current bugs :

2) complete writing up of resampling_test.c, check reconstruction in-band part.
check the out-band part: does that fix the spurious peaks appearing in the reject region?
