Francis ,
Here is your (first) posted (1024x576 pixel-size) JPG with LX3 Exposure Compensation =0.0 EV (which caused "Flashing Highlight" indications on your LX3). The DxO Optics Pro 6.60 (and the PaintShop Pro 9.01) RGB histograms of the original image posted does not show more a very tiny (virtually insignificant) amount of color-channel "clipping" taking place:
Here is that posted image processed using DxO Optics Pro 6.60 (DxO LX3 JPG Optical Corrections Module used) with the
(DxO) Exposure Compensation set to -0.60 EV :
.
Here is your (second) posted (1024x576 pixel-size) JPG with LX3 Exposure Compensation = -1.0 EV (causing no Flashing Highlight indications on your LX3). The DxO Optics Pro 6.60 (and the PaintShop Pro 9.01) RGB histograms do not show
any amount of color-channel "clipping":
Here is that posted image processed using DxO Optics Pro 6.60 (DxO LX3 JPG Optical Corrections Module used) with the
(DxO) Exposure Compensation set to +0.40 EV .
.
All DxO Optics Pro 6.60 processing parameter settings were identical - except that the (DxO) Exposure Compensation was set (a net) 1.0 EV higher than it was in the case of the (DxO processed) first image (above).
Note that the difference in the LX3 Exposure Compensation between the two shots was exactly compensated for the by the difference in DxO Exposure Compensation settings during processing - yet
there appears (to me) to be clearly visible differences in the amount of image-details in the two processed images .
Since your first-posted (higher exposure-level) JPG image has virtually no actual color-channel "clipping", the visible differences are not due to any significant image-sensor non-linearity that the LX3 is able to alert you to the existence of (via the "Flashing Highlights" indicator in Review, or in Playback, mode viewing).
It is my experience that the "flashing highlights" indication does not exist when the Live Histogram does not (itself) report (composite RGB color-channel luminance) "clipping".
I believe that the differences in visible image-details demonstrates non-linearity of the CCD image-sensor photo-sites of the LX3 (which occurs
prior to a 100% reading on the Live Histogram, and the corresponding "Flashing Highlights" indication in Review, or in Playback, mode viewing).
The existence of such suspected image-sensor photo-site non-linearities is precisely why I recommend restricting the Live Histogram reading to between 50% and 70% of full-scale reading (when using center-weighted metering) - which corresponds to
reducing the LX3 exposure-level (from the Live Histogram and Flashing Highlights maximum levels) by between 0.5 EV and 1.0 EV.
DM