> Wow, the code at http://mavis.anu.edu.au/nifti is Bob's original NIfTI
> code, before either Mark or I added anything, and certainly predating
> the first official release at sourceforge (version 0.1). Though
> strangely enough, that nifti1.h file is more current (including the
> entender struct).
Aye, the original version of mnc2nii was written by Bert yonks ago, he
was one of the members of the original NIFTI-1 Panel. I guess this
means we could call him an early adopter. :)
> 1. While it is likely that mnc2nii uses the NIfTI library only for a
> basic read and write, even adding the code to read/write extender
> fields (to say nothing of the extensions themselves) would probably
> be more work than updating the library.
>
> Also, since that verion is well before the first official release,
> there is a lot of useful functionality that you don't currently have
> access to, such as handling large and/or compressed datasets, and
> reading/writing NIfTI datasets as a list of volumes (instead of a
> single, large block).
OK, thanks. For now I will stick with including a static version of
the latest version and see how this works.
> 2. Note that byte swapping should not be necessary on the part of
> mnc2nii, even using that early code. Swapping is done when the
> calling function requests data (via either nifti_image_read() or
> nifti_image_load(). That means mnc2nii is using nifti_image_read()
> to read the header, but is never requesting to populate it with data.
> So mnc2nii reads the data on its own, and needs to swap bytes.
Thanks, That means that the mnc2nii code can be a lot simpler now.. . :)
In any case I managed to hunt down the bug so all is good now.
--
Andrew Janke ([log in to unmask] || http://a.janke.googlepages.com/)
Canberra->Australia +61 (402) 700 883
|