Bug #59785 for Imager: Clarification of hardinvert behaviour

Wed Jul 28 13:51:52 2010 dorian [...] cpan.org - Ticket created

Subject:

Clarification of hardinvert behaviour

According to the documentation in Imager::Filters, the 'hardinvert' filter is supposed to invert all channels of an image including the alpha channel. When I ran my test script (./hardinvert.pl bipartite-test.png > bipartite-out.png) it appears that the alpha channel was left alone. If the alpha channel was inverted as well, would it not create a knockout where the original shape was and fill in the erstwhile negative space with (in this case black) pixels, or am I misinterpreting things? Anyhow, it seems to me that leaving the alpha channel alone is certainly more common (if not outright more desirable) behaviour, as those interested in truly messing with the alpha channel would probably want to do it on its own. I just want to make sure that the behaviour stays the same across releases. Thanks

Subject:

hardinvert.pl

#!/usr/bin/perl use strict; use warnings FATAL => 'all'; use Imager; my $im = Imager->new; $im->read(file => $ARGV[0]) or die $im->errstr; $im->filter(type => 'hardinvert'); $im->write(fh => \*STDOUT, type => 'png') or die $im->errstr;

Subject:

bipartite-test.png

Download bipartite-test.png
image/png 15.2k

Subject:

bipartite-out.png

Download bipartite-out.png
image/png 8.3k

Wed Jul 28 19:14:43 2010 TONYC [...] cpan.org - Correspondence added

On Wed Jul 28 13:51:52 2010, DORIAN wrote: Show quoted text

> According to the documentation in Imager::Filters, the 'hardinvert' > filter is supposed to invert all channels of an image including the > alpha channel. When I ran my test script (./hardinvert.pl > bipartite-test.png > bipartite-out.png) it appears that the alpha > channel was left alone. > > If the alpha channel was inverted as well, would it not create a > knockout where the original shape was and fill in the erstwhile negative > space with (in this case black) pixels, or am I misinterpreting things? > > Anyhow, it seems to me that leaving the alpha channel alone is certainly > more common (if not outright more desirable) behaviour, as those > interested in truly messing with the alpha channel would probably want > to do it on its own. I just want to make sure that the behaviour stays > the same across releases. > > Thanks

Thanks for the report. I'll change the documentation to indicate that the alpha channel is left alone. I may add another filter that inverts all channels (though that's possible with the convert() method.) Tony

Wed Jul 28 19:14:43 2010 The RT System itself - Status changed from 'new' to 'open'

Wed Aug 11 21:53:49 2010 TONYC [...] cpan.org - Correspondence added

On Wed Jul 28 13:51:52 2010, DORIAN wrote: Show quoted text

> According to the documentation in Imager::Filters, the 'hardinvert' > filter is supposed to invert all channels of an image including the > alpha channel. When I ran my test script (./hardinvert.pl > bipartite-test.png > bipartite-out.png) it appears that the alpha > channel was left alone. > > If the alpha channel was inverted as well, would it not create a > knockout where the original shape was and fill in the erstwhile negative > space with (in this case black) pixels, or am I misinterpreting things? > > Anyhow, it seems to me that leaving the alpha channel alone is certainly > more common (if not outright more desirable) behaviour, as those > interested in truly messing with the alpha channel would probably want > to do it on its own. I just want to make sure that the behaviour stays > the same across releases.

This is fixed in Imager 0.77: - update the documentation of hardinvert to match the change in 0.62. https://rt.cpan.org/Ticket/Display.html?id=59785 - added hardinvertall filter which also inverts the alpha channel (sorry for the mess) which has just been released. Thank you for your report. Tony

Wed Aug 11 21:53:50 2010 TONYC [...] cpan.org - Status changed from 'open' to 'resolved'