Thanks, from the responses it's clear there are fast ways to convert
between these formats using FSL and MRICRO, so the dual file/single
file issue has an easy solution using freely available software.
I agree that having faster routines is better than having the gzipped
format option. The .img/.hdr format also has the minor advantage of
not requiring users to start thinking in terms of 4D multi-volume
files. Looking forward to the full SPM5 release in November.
Regards,
Stephan
On Sep 22, 2005, at 5:34 AM, John Ashburner wrote:
> Dear Stephan,
> I'm afraid that there is no option to do this, although the NIfTI
> utilities
> within SPM would allow it to be done. My choice of using the dual
> file
> system was based on the FIL having an automated backup system.
> Having a
> separate .hdr and .img would mean that only the .hdr files would be
> backed up
> after tweeks are made to the headers (e.g. after realignment), which I
> figured would save us a lot of tape.
>
> I also chose to not work directly with gzipped data - mostly
> because of
> lazyness. The .gz files would also make a number of routines
> slower. For
> example, if you just want to read a single voxel in a file, then
> all the data
> before that voxel would need to be uncompressed.
>
> The AFNI developers are part of the Data Format Working Group who
> developed
> NIfTI, so I'm surprised that AFNI doesn't deal with the .hdr+.img
> format.
>
> Best regards,
> -John
>
>
>> Is there an option in SPM5b (or the upcoming SPM5) to output to the
>> single-file (.nii) NIfTI-1 format?
|