| || |
Re: VerifyBitmap fails with identical bitmaps
User Costigan, Polly (firstname.lastname@example.org) posted:
Your first sentence is pretty much correct but not totally accurate. The
problem with bitmap comparisons is that little differences between machines
such as resolution, selected color scheme, even "brightness" of the colors
can impact the comparison. If you have identical hardware (video cards) and
setup (i.e., resolution) on all machines then these should not be issues so
a bitmap captured on one machine should work fine on another machine. If
there are only very minor differences then the tolerance levels within the
bitmap comparison functions might solve that problem. But if your
machine/setup differences are more than just minor ones then you may very
well need separate baseline bitmaps for each machine.
Due to the degree of sensitivity, my project does not automate bitmap
verifications. This works for us because (a) we still have a lot of manual
testing occurring anyway, (b) we don't have very many bitmaps to be
verified, and (c) it's simply more reliable.
Re: Re: VerifyBitmap fails with identical bitmaps
User Whitehouse, David (email@example.com) posted:
Bitmap verifications do compare selected blocks of pixels (desktop, client,
or user defined and their respective color tables between a stored file and
a screen image. Some variables can be programmed for timing and tolerance.
Comparisons that involve the desktop are more prone to error if the
resolutions of the recorded image and test machines are different.
Comparisons of the client window may work with different resolutions
providing the client is not full screen and the client images being compared
are not resized from some default that is screen resolution independent.
User defined bitmaps are relative to the window frame and basically follow
the same rules as the client window comparison. Depending on your needs, I
would recommend that you not use bitmaps. However, if you must, you should
minimize the size of the map to reduce the size of the comparison file. This
will help speed up the test and minimize the possibly of errors due to image
size, resolution and the proximity of other image variables that may appear
near your test zone.
Comparisons that involve different color tables or other color variations in
the compared image will always be error prone. Variances the bit and color
table comparisons may be compensated using the system options mentioned
above. You should consult the language manual for VerifyBitmap and the user
guide Agent Bitmap Options for elements that you can change.
Image brightness will not, in itself, cause an error in the comparison
unless the color table is also affected. Brightness is generally considered
to be a display control which has an analog or digital impact on the display
device and does not affect the image as it constructed and presented to the
display device. In other words, the bitmap is captured from the display RAM.
The display RAM does not contain brightness information. Brightness is a
characteristic which affects the entire image and is applied to the RAM data
as it is presented to the screen.