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
In getPhase(), stop using different weights for sumWc and sumWcp.
This should improve linearity e.g. in earlyTerminationByPhase().
Use AWB statistics, and optionally try to detect IR lighting and
suppress PDAF in that case, as IR interferes with PDAF on IMX708.
Make earlyTerminationByPhase() harder to trigger; extrapolate less.
Improve quadratic peak fitting in findPeak(). The old approximation
was good but only really valid when points were equally spaced and
the MAX was not at one end of the series.
For PDAF-based CAF, suppress lens movement unless "phase" has had
the same sign for 4 frames in a row. This suppresses noise/wobble.
Change the logic for falling back from PDAF to CDAF, and for re-
triggering a CDAF-based scan. It is now triggered by image change
(based on both Focus and AWB statistics) followed by stability.
Modify CDAF scan pattern so it doesn't always start at the lowest
setting, but can sometimes start from the current lens position.
(Though in the worst case this requires two coarse scans, if the
first one went in the wrong direction.)
Shorten fine scan and allow it to be skipped when step_fine == 0
Signed-off-by: Nick Hollinghurst <nick.hollinghurst@raspberrypi.com>
0 commit comments