odklizec
PK
Unfortunately Jeremy, I don't have the luxury of time required for detailed tests and analysis to confirm or deny your findings.
I'm sure DxO has many strengths as well as weaknesses, as any other RAW editor (the same is true for digital cameras). DxO, for example, have RGB Curves, which are still missing in LR3. This is a serious show-stopper for me and my image processing style. Therefore, I see no way LR3 would replace my Silkypix Pro anytime soon
But I see much better chances for LR3 than for DxO, because DxO will most probably never support the same range of cameras as LR. And even if I will most probably never own all new cameras, I like to try RAW from this or that new model. And this is something I'm unable to do with programs like DxO or Bibble (another very good RAW processor).
However, up to now, I saw no evidence that DxO is doing significantly better job in sharpening and color noise reduction than LR3. Definitely not in case of small sensor cameras like LX3 or G11, which I'm most interested in. In contrary, in today's example (please don't take it personally) I clearly see the disadvantage of your chosen processing. Maybe it can be improved, maybe not? I just see what I see. And I'm afraid, I'm very sensitive to sharpening and noise reduction artifacts. And I see serious oversharpening in today's comparison.
As for the G11 comparison you did some time ago and to which you are referring today, I see not only oversharpnening, but also strong noise reduction artifacts that are not present in the LR3 output. Once again, a side by side comparison...
Yes, the LR3 output is definitely noisier and less colorful/contrast. But I have no doubt that at least the color/contrast can be further improved. As for the color noise, this can be most probably improved too. If I'm not mistaken, you set both DxO and LR3 color noise to 8%? Try something around 30-40% and the color noise will be gone, with a lot of preserved details. As for the Luminance noise, DxO seems do nice job in eliminating this kind of noise. From what I saw, LR3 is not that good in case of Luminance noise reduction. So LR3 sample will most probably be still more grainier than DxO export. For me personally, it's not a problem, because I prefer grainy images. I'm even used to add some artificial grain to my photos
To be honest, I see no point setting both LR3 and DxO in the same way (e.g. 8% for Chroma noise reduction or for example tone-curve to linear). Both processors use different processing algorithms and default input values and setting them both in the same way is, in my humble opinion, meaningless. It's the same like setting two different cameras to standard or even zero values and expecting the same results? The only thing you will learn from this approach is that both editors/cameras use different defaults and completely different processing algorithms, as you very well know anyway. So why setting them both with the same values?
I, on the other hand, always try to achieve the best results with this or that editor. I don't really care about "default" settings and even less about the same settings, because I know they are always different in both programs. There is no, or very little chance you can produce exactly the same results, no matter how much time you spend with both programs. So instead of trying the same settings, try for example to achieve the best result with one program and then, without looking at the result from the first program, try the same with other one. Of course, the results will always differ in this or that way.
I'm sure DxO has many strengths as well as weaknesses, as any other RAW editor (the same is true for digital cameras). DxO, for example, have RGB Curves, which are still missing in LR3. This is a serious show-stopper for me and my image processing style. Therefore, I see no way LR3 would replace my Silkypix Pro anytime soon
But I see much better chances for LR3 than for DxO, because DxO will most probably never support the same range of cameras as LR. And even if I will most probably never own all new cameras, I like to try RAW from this or that new model. And this is something I'm unable to do with programs like DxO or Bibble (another very good RAW processor). However, up to now, I saw no evidence that DxO is doing significantly better job in sharpening and color noise reduction than LR3. Definitely not in case of small sensor cameras like LX3 or G11, which I'm most interested in. In contrary, in today's example (please don't take it personally) I clearly see the disadvantage of your chosen processing. Maybe it can be improved, maybe not? I just see what I see. And I'm afraid, I'm very sensitive to sharpening and noise reduction artifacts. And I see serious oversharpening in today's comparison.
As for the G11 comparison you did some time ago and to which you are referring today, I see not only oversharpnening, but also strong noise reduction artifacts that are not present in the LR3 output. Once again, a side by side comparison...
Yes, the LR3 output is definitely noisier and less colorful/contrast. But I have no doubt that at least the color/contrast can be further improved. As for the color noise, this can be most probably improved too. If I'm not mistaken, you set both DxO and LR3 color noise to 8%? Try something around 30-40% and the color noise will be gone, with a lot of preserved details. As for the Luminance noise, DxO seems do nice job in eliminating this kind of noise. From what I saw, LR3 is not that good in case of Luminance noise reduction. So LR3 sample will most probably be still more grainier than DxO export. For me personally, it's not a problem, because I prefer grainy images. I'm even used to add some artificial grain to my photos
To be honest, I see no point setting both LR3 and DxO in the same way (e.g. 8% for Chroma noise reduction or for example tone-curve to linear). Both processors use different processing algorithms and default input values and setting them both in the same way is, in my humble opinion, meaningless. It's the same like setting two different cameras to standard or even zero values and expecting the same results? The only thing you will learn from this approach is that both editors/cameras use different defaults and completely different processing algorithms, as you very well know anyway. So why setting them both with the same values?
I, on the other hand, always try to achieve the best results with this or that editor. I don't really care about "default" settings and even less about the same settings, because I know they are always different in both programs. There is no, or very little chance you can produce exactly the same results, no matter how much time you spend with both programs. So instead of trying the same settings, try for example to achieve the best result with one program and then, without looking at the result from the first program, try the same with other one. Of course, the results will always differ in this or that way.
Attachments
-
EXIFcomparison2.jpg
