[Pkg-exppsy-maintainers] Was: ridge... now VALGRIND time

[Yaroslav Halchenko]

> Typically we collect continuous data for about 1 hour at around 500Hz.
>  That gives us 1.8 million samples per channel with up to around 128
> channels.  Definitely doable.
if my calculations correct it would be
*octave:2> (1.8*10^6*128*8)/1024^3
ans = 1.7166
1.7GB of data... not too bad... should be managable to work on in memory
;-)

> However, I think one very viable option is to store EEG in the NetCDF
> data format, which has some good python wrappers that allow you to
> access the dataset as a numpy array, but not load the entire thing
> into memory.  There are even wrappers for matlab (if people would like
> to do such a thing.)
That sounds great! Someone should write NetCDFDataset! We should add
that to TODO ;-) Shouldn't be too difficult I think...

> The only problem is that the latest NetCDF is not at a stable release
> the last time I checked.  Still possibly usable, but a moving target.

Is API changing wildly?

-- 
Yaroslav Halchenko
Research Assistant, Psychology Department, Rutgers-Newark
Student  Ph.D. @ CS Dept. NJIT
Office: (973) 353-5440x263 | FWD: 82823 | Fax: (973) 353-1171
        101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102
WWW:     http://www.linkedin.com/in/yarik