You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At least for FLIR GiGE cameras, changing binning changes the apparent size
of the chip, such that the region of interest is configured using a smaller
number of pixels. This appears to be contrary to how INDI expects binning
to be configured. Following the CCD Simulator (hopefully it is a decent
example of implementing a INDI::CCD device), it rather appears that the
region of interest is still expected to be set using the entire available
Width/Height. It therefore appears necessary to mitigate this mismatch
since (at least for ekos) the client appears to mess up the apparent image
if binning is configured but a smaller region is selected, but in reality
the entire avaialble sensor was binned.
So, for now, even when binning is used, this patch makes the driver use the
full WidthMax and HeightMax manually divided by the binning to select the
region of interest.
This means that any use of get_x_offset().val(), get_y_offset().val(),
get_width().val(), or get_height().val() from the GigECCD::camera object must
be manually multiplied or divided by the values from get_bin_x().val() and
get_bin_y().val() as appropriate. Similarly, camera->set_geometry(...) must
be given values that are divided by the appropriate binning.
It remains to verify this behavior on GiGE cameras from other vendors, or
at least to check what the standard says, if anything useful on the matter.
0 commit comments