From zijdenbos at gmail.com  Tue Dec 13 15:42:30 2011
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Tue, 13 Dec 2011 15:42:30 -0500
Subject: [MINC-users] mincaverage bug
Message-ID: <CABEEk14XXX-fmhkwYMd05Fqr-ZNeFn-3CgncGdztgrFub8yFow@mail.gmail.com>

Hello all,

I just stumbled across a bug in mincaverage. If you try to average a
list of volumes that have a small range for the data type (e.g., byte
0-3), the average (correctly) inherits that data type- and range. But
obviously a 4-valued average is hardly useful; so I thought the output
type options would come to the rescue. Alas, they only seem to affect
the sd volume, not the average. In other words, running something like
this:

   mincaverage -short -range 0 1000 in1.mnc in2.mnc avg.mnc -sdfile sd.mnc

will give sd.mnc a short type and the advertised range; but not avg.mnc.

-- A

From a.janke at gmail.com  Tue Dec 13 15:53:27 2011
From: a.janke at gmail.com (Andrew Janke)
Date: Wed, 14 Dec 2011 06:53:27 +1000
Subject: [MINC-users] mincaverage bug
In-Reply-To: <CABEEk14XXX-fmhkwYMd05Fqr-ZNeFn-3CgncGdztgrFub8yFow@mail.gmail.com>
References: <CABEEk14XXX-fmhkwYMd05Fqr-ZNeFn-3CgncGdztgrFub8yFow@mail.gmail.com>
Message-ID: <CAG0CB23ezedTRj6r2bqiGABXMAuG8igTWnnwZ9aaiaHT=4rhEg@mail.gmail.com>

Hi Alex,

On 14 December 2011 06:42, Alex Zijdenbos <zijdenbos at gmail.com> wrote:
> I just stumbled across a bug in mincaverage. If you try to average a
> list of volumes that have a small range for the data type (e.g., byte
> 0-3), the average (correctly) inherits that data type- and range. But
> obviously a 4-valued average is hardly useful; so I thought the output
> type options would come to the rescue. Alas, they only seem to affect
> the sd volume, not the average. In other words, running something like
> this:
>
> ? mincaverage -short -range 0 1000 in1.mnc in2.mnc avg.mnc -sdfile sd.mnc
>
> will give sd.mnc a short type and the advertised range; but not avg.mnc.

I can't seem to reproduce this behavior you speak of?

$ mincaverage -double bp-native.mnc a.mnc
$ mincaverage -double bp-native.mnc d.mnc
$ mincaverage -short -range 0 1000 bp-native.mnc s.mnc

cai-harold:~$ mincinfo a.mnc d.mnc s.mnc
file: a.mnc
image: unsigned byte 0 to 255
image dimensions: zspace yspace xspace
    dimension name         length         step        start
    --------------         ------         ----        -----
    zspace                    267      1.01563     -135.079
    yspace                    217      1.01563     -109.688
    xspace                    267      1.01563     -135.079


file: d.mnc
image: signed__ double 0 to 5.4909507159588706315
image dimensions: zspace yspace xspace
    dimension name         length         step        start
    --------------         ------         ----        -----
    zspace                    267      1.01563     -135.079
    yspace                    217      1.01563     -109.688
    xspace                    267      1.01563     -135.079


file: s.mnc
image: unsigned short 0 to 1000
image dimensions: zspace yspace xspace
    dimension name         length         step        start
    --------------         ------         ----        -----
    zspace                    267      1.01563     -135.079
    yspace                    217      1.01563     -109.688
    xspace                    267      1.01563     -135.079

?

mincaverage uses voxel_loop so if there is a bug in there it will
affect a number other minc tools as well.


ta


a

From zijdenbos at gmail.com  Tue Dec 13 21:58:56 2011
From: zijdenbos at gmail.com (Alex Zijdenbos)
Date: Tue, 13 Dec 2011 21:58:56 -0500
Subject: [MINC-users] mincaverage bug
In-Reply-To: <CAG0CB23ezedTRj6r2bqiGABXMAuG8igTWnnwZ9aaiaHT=4rhEg@mail.gmail.com>
References: <CABEEk14XXX-fmhkwYMd05Fqr-ZNeFn-3CgncGdztgrFub8yFow@mail.gmail.com>
	<CAG0CB23ezedTRj6r2bqiGABXMAuG8igTWnnwZ9aaiaHT=4rhEg@mail.gmail.com>
Message-ID: <CABEEk17TnofQ4myQGZ1S9SOLKJr057tu5iFVNsT4tbbes1yuAA@mail.gmail.com>

Hi Andrew,

False alarm - user error :) I had started with this command line

mincaverage .... avg.mnc -sdfile sd.mnc

and somewhere along the way of experimenting the '-sdfile' opt dropped
off, resulting in

mincaverage ... avg.mnc sd.mnc

which evidently was not what I had intended :)

Sorry

-- A

On Tue, Dec 13, 2011 at 3:53 PM, Andrew Janke <a.janke at gmail.com> wrote:
> Hi Alex,
>
> On 14 December 2011 06:42, Alex Zijdenbos <zijdenbos at gmail.com> wrote:
>> I just stumbled across a bug in mincaverage. If you try to average a
>> list of volumes that have a small range for the data type (e.g., byte
>> 0-3), the average (correctly) inherits that data type- and range. But
>> obviously a 4-valued average is hardly useful; so I thought the output
>> type options would come to the rescue. Alas, they only seem to affect
>> the sd volume, not the average. In other words, running something like
>> this:
>>
>> ? mincaverage -short -range 0 1000 in1.mnc in2.mnc avg.mnc -sdfile sd.mnc
>>
>> will give sd.mnc a short type and the advertised range; but not avg.mnc.
>
> I can't seem to reproduce this behavior you speak of?
>
> $ mincaverage -double bp-native.mnc a.mnc
> $ mincaverage -double bp-native.mnc d.mnc
> $ mincaverage -short -range 0 1000 bp-native.mnc s.mnc
>
> cai-harold:~$ mincinfo a.mnc d.mnc s.mnc
> file: a.mnc
> image: unsigned byte 0 to 255
> image dimensions: zspace yspace xspace
> ? ?dimension name ? ? ? ? length ? ? ? ? step ? ? ? ?start
> ? ?-------------- ? ? ? ? ------ ? ? ? ? ---- ? ? ? ?-----
> ? ?zspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
> ? ?yspace ? ? ? ? ? ? ? ? ? ?217 ? ? ?1.01563 ? ? -109.688
> ? ?xspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
>
>
> file: d.mnc
> image: signed__ double 0 to 5.4909507159588706315
> image dimensions: zspace yspace xspace
> ? ?dimension name ? ? ? ? length ? ? ? ? step ? ? ? ?start
> ? ?-------------- ? ? ? ? ------ ? ? ? ? ---- ? ? ? ?-----
> ? ?zspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
> ? ?yspace ? ? ? ? ? ? ? ? ? ?217 ? ? ?1.01563 ? ? -109.688
> ? ?xspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
>
>
> file: s.mnc
> image: unsigned short 0 to 1000
> image dimensions: zspace yspace xspace
> ? ?dimension name ? ? ? ? length ? ? ? ? step ? ? ? ?start
> ? ?-------------- ? ? ? ? ------ ? ? ? ? ---- ? ? ? ?-----
> ? ?zspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
> ? ?yspace ? ? ? ? ? ? ? ? ? ?217 ? ? ?1.01563 ? ? -109.688
> ? ?xspace ? ? ? ? ? ? ? ? ? ?267 ? ? ?1.01563 ? ? -135.079
>
> ?
>
> mincaverage uses voxel_loop so if there is a bug in there it will
> affect a number other minc tools as well.
>
>
> ta
>
>
> a
> _______________________________________________
> MINC-users at bic.mni.mcgill.ca
> http://www.bic.mni.mcgill.ca/mailman/listinfo/minc-users