From PL9.0 release notes:
The calibration process and the sharpening algorithm were reworked from the ground up to ensure homogenous results across all camera gear and shooting conditions (feature will roll out progressively).
Anyone knows more details?
Any practical differences seen, compared to PL8?
I haven’t had time to play around with it yet, but I hope I understand it correctly in that the new sharpness optimization is normalized so the same value tries to get the same sharpness for the same value, whether the lens is soft or sharp. Meaning it optimizes a little for already sharp lenses and a lot for soft lenses.
The problem with the old system was that some specific value such as 0 was good for old, soft lenses but oversharpened horribly for modern, already sharp lenses. Their default value of 1 always oversharpened, though. Often to the point of making pro gear photos look like smartphone photos.
Sounds reasonable, but there could be something more under the hood.
I find this cryptic: “feature will roll out progressively”, plus talk about calibration and algorithm.
It seems that DxO has realized that and in PL9 the default LSO Intensity of 100 corresponds to PL8 Intensity 0.0, while PL8 default +1.0 corresponds to PL9 value 133.3. For portraits you may like still lower values, but for very noisy DP photos taken with soft lenses the old default might work. I had few cases where PL8 LSO +2.0 was just right. On the other hand, with Plena lens (Nikkor Z 135/1.8) I quite often disable LSO.
Here is how PL9 maps LSO Intensity when dealing with PL8 dop files (based on my test with PL8.8 and PL9.0.1 trial) :
PL8 PL9
-3.00 15
-2.50 29
-2.00 43
-1.50 57
-1.00 72
-0.50 86
0.00 100
+0.50 117
+1.00 133
+1.50 150
+2.00 167
+2.50 183
+3.00 200
It’s linear in the [-3.0..0.0] and [0.0..+3.0] ranges, with slightly different angles.
Know? Guess…: Changing a very core part could necessitate other changes like e.g. reworked optical modules. If that were the case and the currently listed almost 125’000 modules, it seems highly probable that the rollout will take time to accomplish.
Practical differences? Depends on what one’s aspirations are. My guess (again) is that the effect might be more or less visible, depending on the gear that was used to capture the scene in the first place. Going from 70% sharp to 90% sharp might be fairly visible and going further to 95% could probably be detected too. Whether that difference is practical / useful / needed / … remains to be seen.
