[Magick-bugs] Image Comparing (via "compare" or "convert") returnsdifferent results on ImageMagick 6.3.0 and 6.3.3

[Keith Bernstein]

That looks like it... thanks a lot, Glenn, that solves the mystery (well, except for the one headless machine we have at our co-lo
which always gives back "65278"

So, is there a way I can set this Q-value, or account for it, as part of the call to "compare"?

Thanks.

- Keith

-----Original Message-----
From: Glenn Randers-Pehrson [mailto:glennrp at comcast.net] 
Sent: Tuesday, May 01, 2007 4:33 AM
To: keithbernstein at mail.com
Cc: magick-bugs at imagemagick.org
Subject: Re: [Magick-bugs] Image Comparing (via "compare" or "convert") returnsdifferent results on ImageMagick 6.3.0 and 6.3.3

At 08:32 PM 4/30/2007 -0700, Keith Bernstein wrote:

>Is there an explanation for this? Can I just take (e.g.) (8117.97 / 31.5874) 
>and use that to normalize threshold values between
>these two systems?

Interestingly, that ratio works out to almost exactly 257, which would
be the ratop in sample values on a 16-bit build versus an 8-bit build.

Run "convert | head -2" and look at the Q values on the two systems.  One
is probably Q8 and the other Q16.

Glenn