Hello,
It is a 32-bit / 64-bit problem (a C "long" used to be 4 bytes but is now
8). When I sorted that out with a quick hack then the 2D processed ok.
Well, the phasing was off and the spectrum might be cycled by npts/2 in
D2, but otherwise looked plausible. I'll sort it out properly and send
you the patch (it will need re-compiling though of course).
On the baseline front, I'll look at that paper on Monday (hopefully it can
be looked at online).
Wayne
On Fri, 10 Dec 2010, Justin Lecher wrote:
> On 10/12/10 18:23, Brian Smith wrote:
>> On Fri, 10 Dec 2010, Justin Lecher wrote:
>>
>>> Word to look at (< 0 to stop): 1
>>> float = 0.000000, int = 16777216, hex = 1000000
>>> swapped: float = 0.000000, int = 1, hex = 1
>>
>> So the ntraces parameter is 1 (in the swapped:int field) and the problem
>> comes from the endinanness with wich AZARA reads the header. I'm pretty
>> sure that if you recompile (don't forget to make clean) with the
>> endianness block in ENVIRONMENT set as follows
>>
>> # set up ENDIAN_FLAG as required
>> # WRITE_ENDIAN_PAR if want endianess written to par file
>> # BIG_ENDIAN_DATA if want data files to be big endian by default
>> # LITTLE_ENDIAN_DATA if want data files to be little endian by default
>> # FLIP_SWAP if want input data files to have swap flipped
>> # not all of these make sense in combination
>> # some examples of what you might do:
>> #ENDIAN_FLAG = -DWRITE_ENDIAN_PAR
>> ENDIAN_FLAG = -DBIG_ENDIAN_DATA
>> #ENDIAN_FLAG = -DLITTLE_ENDIAN_DATA
>>
>> you will be able to process your data with the varian keyword.
>>
>
> I compiled it correctly all the time. But w/o success.
> I removed the call the the varian header check function from the code
> and it works fine now.
>
> big_endian now!
> no swap needed
>
> no extract "head 8" specified. npts corrected to the original value of
> 1794. And also the "range 8 1802" removed.
>
> But I am still worrying whether I read in the data correctly now. Wayne,
> could you be so kind and take a look into it.
> My original concern, finding a better baseline correction or less
> baseline distortion, than what I get with nmrpipe is fullfilled with
> azara, as far as I can judge it now. So I really like to use azara from
> now on.
>
> justin
> --
> Justin Lecher
> Institute of Structural Biology and Biophysics
> ISB-3 Structural Biochemistry
> Research Centre Juelich
> 52425 Juelich, Germany
> phone: +49 2461 61 2117
>
>
>
|