Re: Installation Question: Debian 12, GraphicsMagick 1.3.42, JasPer 4.0
Matthias Schmitt via Graphicsmagick-help <[email protected]> Sat, 30 Sep 2023 16:57:00 +0200
| Newsgroups | gmane.comp.video.graphicsmagick.help |
|---|---|
| Message-ID | <[email protected]> |
Hello Bob, > On 29 Sep 2023, at 22:25, Bob Friesenhahn <[email protected]> wrote: > > Matthias, I allowed this email through even though you are not subscribed. But just this once. Please subscribe to the list before posting to it. Ooops, sorry. I was sending with a different mail address. I subscribed in 2010. >>> not ok 12 - JP2 truecolor16 >>> FAIL: tests/rwfile_deep.tap 12 - JP2 truecolor16 >> >> BTW, the JP2 tests for gray images pass successfully. > > That is very interesting. Perhaps JasPer is "working" but the computed error/difference is higher than the limit? > > If you use GraphicsMagick to open and look at a JP2 file that was written, does it visually look similar to the input image? Yes, you are right. All JP2 test images seem to be created correctly. When I convert images using the gm tool, JP2 images get generated and they look flawless. I double checked that the resulting files are really in JPEG 2000 format and not another format. May be this explains why when compiling PerlMagick some of my tests fail too. > Test Summary Report > ------------------- > t/jng/write.t (Wstat: 0 Tests: 11 Failed: 1) > Failed test: 11 To get more information about the error I executed the test manually. > perl -I/MY_PATH_TO/sources/GraphicsMagick-1.3.42/PerlMagick t/jng/write.t > 1..11 > ok 1 > ok 2 > ok 3 > ok 4 > ok 5 > ok 6 > ok 7 > ok 8 > ok 9 > ok 10 > mean-error=0.0120059817766502, maximum-error=0.143998661477254 > not ok 11 same here > t/jpeg/read.t (Wstat: 0 Tests: 5 Failed: 1) > Failed test: 5 > perl -I/MY_PATH_TO/sources/GraphicsMagick-1.3.42/PerlMagick t/jpeg/read.t > 1..5 > Non-interlaced JPEG ... > ok 1 > Plane-interlaced JPEG ... > ok 2 > Seattle FilmWorks image file ... > ok 3 > Garmin JNX image file ... > ok 4 > JPEG-compressed BMP image file ... > mean-error=0.00555143524702116, maximum-error=0.065385615688365 > not ok 5 As you document in Measuring Color Reduction Error (http://www.graphicsmagick.org/quantize.html) this seems to be more a relative error than an absolute. Best regards Matthias magic moving pixel s.a. 23, Avenue Grande-Duchesse Charlotte L-3441 Dudelange Luxembourg Phone: +352 54 75 75 http://www.mmp.lu