update notes

This commit is contained in:
John Cupitt 2014-07-16 14:54:51 +01:00
parent 87698cc961
commit 9d096c39f1

76
TODO
View File

@ -1,76 +1,3 @@
- butterworth item in filter menu in nip2 makes a mask, it doesn't filter the
image
- butterworth highpass is wrong! actually makes a lowpass filter
$ ~/vips-7.26/bin/vips-7.26 vips im_create_fmask x2.v 32 32 2 42 0.5 0.5 0 0
is correct: a black circle on a white bg, with a white dot top left
2 because we have
vips7compat.h: IM_MASK_BUTTERWORTH_HIGHPASS = 2,
in vip8 this is
$ vips mask_butterworth x.v 32 32 42 0.5 0.5
generates a highpass filter, or
$ vips mask_butterworth x.v 32 32 42 0.5 0.5 --reject
for a lowpass filter
doesn't match ideal! there we have --reject to make a highpass filter
im_freq_mask has:
IM_MASK_IDEAL_HIGHPASS: 0
if( vips_mask_ideal(
IM_MASK_IDEAL_LOWPASS: 1
if( vips_mask_ideal( "reject", TRUE,
IM_MASK_BUTTERWORTH_HIGHPASS: 2
if( vips_mask_butterworth(
IM_MASK_BUTTERWORTH_LOWPASS: 3
if( vips_mask_butterworth( "reject", TRUE,
IM_MASK_GAUSS_HIGHPASS: 4
if( vips_mask_gaussian(
IM_MASK_GAUSS_LOWPASS: 5
if( vips_mask_gaussian( "reject", TRUE,
IM_MASK_IDEAL_RINGPASS: 6
if( vips_mask_ideal_ring(
IM_MASK_IDEAL_RINGREJECT: 7
if( vips_mask_ideal_ring( "reject", TRUE,
IM_MASK_BUTTERWORTH_RINGPASS: 8
if( vips_mask_butterworth_ring(
IM_MASK_BUTTERWORTH_RINGREJECT: 9
if( vips_mask_butterworth_ring( "reject", TRUE,
IM_MASK_GAUSS_RINGPASS: 10
if( vips_mask_gaussian_ring(
IM_MASK_GAUSS_RINGREJECT: 11
if( vips_mask_gaussian_ring( "reject", TRUE,
IM_MASK_IDEAL_BANDPASS: 12
if( vips_mask_ideal_ring(
IM_MASK_IDEAL_BANDREJECT: 13
if( vips_mask_ideal_ring( "reject", TRUE,
IM_MASK_BUTTERWORTH_BANDPASS: 14
if( vips_mask_butterworth_ring(
IM_MASK_BUTTERWORTH_BANDREJECT: 15
if( vips_mask_butterworth_ring( "reject", TRUE,
IM_MASK_GAUSS_BANDPASS: 16
if( vips_mask_gaussian_ring(
IM_MASK_GAUSS_BANDREJECT: 17
if( vips_mask_gaussian_ring( "reject", TRUE,
IM_MASK_FRACTAL_FLT: 18
// not implemented
- maxpos_avg seems to give variable results
@ -92,6 +19,9 @@ $ vips max --size 10 spec/samples/wagon.v --x-array --y-array --out-array
seems to be a race in maxpos_avg, it's still some Tom code, replace with a
wrapper over vips_max()
hard to do, since vips_max() needs to be told in advance the number of maxes
it should search for