Closed GoogleCodeExporter closed 9 years ago
As discussed with antonin we decided to update file5 tiff file, see issue 305.
However this makes the test fails now
Original comment by mathieu.malaterre
on 24 Mar 2014 at 5:20
Original comment by mathieu.malaterre
on 24 Mar 2014 at 5:20
To see what was done for file5 in the past, see issue 281
Original comment by mathieu.malaterre
on 24 Mar 2014 at 5:25
Attaching ICC profile
Original comment by mathieu.malaterre
on 25 Mar 2014 at 9:07
Attachments:
Kakadu 7.3.3 does not seems to be compatible with lcms2. Steps:
$ cat test.sh
kdu_transcode -i data/input/conformance/file5.jp2 -o file5.j2k
# Can't create 3 components TIF directly, use PPM as intermediate format:
kdu_expand -i file5.j2k -o rgb.ppm
kdu_compress -i rgb.ppm -o rgb.ppm.jp2 Creversible=yes
kdu_expand -i rgb.ppm.jp2 -o rgb.ppm.jp2.tif
tificc -ifile5.jp2.icm rgb.ppm.jp2.tif file5.j2k.tificc.tif
If we use kdu_expand directly:
kdu_expand -i data/input/conformance/file5.jp2 -o file5.kdu.tif
Using compare_image:
$ crc32 data/baseline/conformance/jp2_5.tif
f76eb229
$ bin/compare_images "-b" data/baseline/conformance/jp2_5.tif" "-t"
file5.j2k.tificc.tif "-n" "3" "-p" "4:4:4" "-m" "1:1:1" "-s" "b_t_"
******Parameters*********
base_filename = data/baseline/conformance/jp2_5.tif
test_filename = file5.j2k.tificc.tif
nb of Components = 3
Non regression test = 0
separator Base = _
separator Test = _
MSE values = [ 1.000000 1.000000 1.000000 ]
PEAK values = [ 4.000000 4.000000 4.000000 ]
Non-regression test = 0
NbFilename to generate from base filename = 3
NbFilename to generate from test filename = 3
*************************
Step 1 -> Header comparison
Step 2 -> measurement comparison
<DartMeasurement name="PEAK_0" type="numeric/double"> 59.000000
</DartMeasurement>
<DartMeasurement name="MSE_0" type="numeric/double"> 19.002998
</DartMeasurement>
ERROR: MSE (19.002998) or PEAK (59.000000) values produced by the decoded file
are greater than the allowable error (respectively 1.000000 and 4.000000)
while:
$ bin/compare_images "-b" data/baseline/conformance/jp2_5.tif" "-t"
file5.kdu.tif "-n" "3" "-p" "4:4:4" "-m" "1:1:1" "-s" "b_t_"
******Parameters*********
base_filename = data/baseline/conformance/jp2_5.tif
test_filename = file5.kdu.tif
nb of Components = 3
Non regression test = 0
separator Base = _
separator Test = _
MSE values = [ 1.000000 1.000000 1.000000 ]
PEAK values = [ 4.000000 4.000000 4.000000 ]
Non-regression test = 0
NbFilename to generate from base filename = 3
NbFilename to generate from test filename = 3
*************************
Step 1 -> Header comparison
Step 2 -> measurement comparison
<DartMeasurement name="PEAK_0" type="numeric/double"> 116.000000
</DartMeasurement>
<DartMeasurement name="MSE_0" type="numeric/double"> 2862.819649
</DartMeasurement>
ERROR: MSE (2862.819649) or PEAK (116.000000) values produced by the decoded
file are greater than the allowable error (respectively 1.000000 and 4.000000)
Original comment by mathieu.malaterre
on 25 Mar 2014 at 9:29
Attachments:
There is basically no difference whether we use lcms2 or lcms1:
$ tifficc -ifile5.jp2.icm rgb.ppm.jp2.tif file5.j2k.tifficc.tif
$ bin/compare_images "-b" data/baseline/conformance/jp2_5.tif" "-t"
file5.j2k.tifficc.tif "-n" "3" "-p" "4:4:4" "-m" "1:1:1" "-s" "b_t_"
******Parameters*********
base_filename = data/baseline/conformance/jp2_5.tif
test_filename = file5.j2k.tifficc.tif
nb of Components = 3
Non regression test = 0
separator Base = _
separator Test = _
MSE values = [ 1.000000 1.000000 1.000000 ]
PEAK values = [ 4.000000 4.000000 4.000000 ]
Non-regression test = 0
NbFilename to generate from base filename = 3
NbFilename to generate from test filename = 3
*************************
Step 1 -> Header comparison
Step 2 -> measurement comparison
<DartMeasurement name="PEAK_0" type="numeric/double"> 59.000000
</DartMeasurement>
<DartMeasurement name="MSE_0" type="numeric/double"> 19.058179
</DartMeasurement>
ERROR: MSE (19.058179) or PEAK (59.000000) values produced by the decoded file
are greater than the allowable error (respectively 1.000000 and 4.000000)
Original comment by mathieu.malaterre
on 25 Mar 2014 at 9:36
Original comment by mathieu.malaterre
on 25 Mar 2014 at 10:37
For reference:
[Lcms-user] Applying ICC profile
https://sourceforge.net/p/lcms/mailman/message/32141849/
Original comment by mathieu.malaterre
on 25 Mar 2014 at 11:56
For reference jasper and kakadu produces the exact same output.
Original comment by mathieu.malaterre
on 25 Mar 2014 at 12:05
file5.jp2 has been updated to the latest version of the conformance test set
(1.5) and test is passing now.
Closing.
Original comment by antonin
on 3 Apr 2014 at 3:50
Original issue reported on code.google.com by
mathieu.malaterre
on 24 Mar 2014 at 5:17