[Eeglablist] Load ANT CNT data
Maarten van de Velde
mvelde at ant-neuro.com
Tue Dec 6 01:58:21 PST 2005
The ANT CNT format was originally designed by developers in the
EEG/MEG/ERP field at the Max Planck Institute for Human Cognitive and
Brain Sciences, Leipzig. The developments were necessitated by the use
of new 22-bit amplifier around 1996. The CNT extension was kept from
their earlier Neuroscan system. Since then, the dataformat has been
improved to allow for 32-bit data storage, including a simple Huffman
compression scheme (compression over channels and time, quite efficient).
The format has not changed much over the past years; changes that have
been released are, and will be, fully backwards compatible.
We have actively addressed the incompatibility issues that have arisen
with the new releases of EEGLAB and releases of Matlab v6.0 R12 and v7.0
R14. The recent problem with v6.5 is new and will be addressed.
For those who wish to work with the mex code by Robert Oostenveld,
please e-mail me at mvelde at ant-neuro.com for the 'cntopenlib' code. The
library 'cntopenlib' is released under the BSD license. We do not
publish this on our website, because we would like to provide personal
assistance when using this library.
We are in the process of creating a more extended version called
'LIBEEP' under LGPL, with more documentation, but that will take some
more time. Before this is released, only the "cntopenlib" is the public
Maarten van de Velde
ANT (Advanced Neuro Technology)
7521 PT Enschede
+31 53 4365175
Robert Oostenveld wrote, On 05/12/2005 9:41 am:
> Dear Joseph and others
> Recently I noticed both on and off the list that multiple people are
> having problems with the ANT format and the mex files.
> I originally wrote the mex code for the ANT matlab importers, and
> also compiled the functions for various OSes and Matlab versions. ANT
> wrote some "glue" around it to make it into an EEGLAB plugin and is
> providing the collection of functions to their customers as a service
> (with my permission, I also gave them my c-code). See also http://
> oase.uci.kun.nl/~roberto/index.php/eeprobe/ on my homepage.
> The c-code underlying the mex files is not too elaborate (few hundred
> lines of code), but it calls several important functions from a c-
> library with the actual file i/o functions. Those functions have to
> do the dirty work like decompressing the data, since it is a complex
> and compressed dataformat.
> I don't mind sharing the c-code of the mex files, so that you can
> compile it yourself, but it would be incomplete without the library.
> I got a copy of the C-library code from ANT when I was a customer
> from them (I moved to another lab since then, and am not using ANT
> software or hardware any more). Although the library with the i/o
> functions is called cntopenlib, I am not 100% sure about the
> agreement under which it was given to me, but I suspect that it is
> covered by the BSD license. I am also not sure whether the fileformat
> has changed since then, and/or whether the library has changed.
> Since ANT does not seem to support the matlab mex files too actively,
> and since I am not using their format any more, it would be nice if
> someone else would take over the responsbility. I am cc'ing this to
> Maarten van der Velde from ANT in the hope that he will pick up this
> request and also take action in clarifying the legal status of their
> c-library. Later today I will put the c-code of the mex files on my
> homepage (the webaddress that I mentioned above) and you (or someone
> else) can compile it yourself and fix the version related bugs that
> recently appeared.
> best regards,
> PS this also applies to people using BIOSIG, since that is also using
> my mex files for the EEProbe formats.
> Robert Oostenveld, PhD
> F.C. Donders Centre for Cognitive Neuroimaging
> Radboud University Nijmegen
> phone: +31-24-3619695
More information about the eeglablist