What a difference a converter makes

[ryc]

I have been playing with RPP today and I am simply amazed with the detail it brings out compared to ACR. Maybe I am doing something wrong in ACR. If I am someone should tell me but look t these images.

You need to see the large images to really see what is going on. Look at the details in the green leaf on the main red bud and the detail on the purple flowers bottom left.

First one is RPP

Second is ACR

And full Size files can be seen here:

FULL SIZE RPP

Full Size ACR

 

[edwardkaraa]

Raw converter comparison:

http://www.alpha-numerique.fr/index...aw&catid=70:developpement-des-raws&Itemid=321

 

[fotografz]

Raw converter comparison:

http://www.alpha-numerique.fr/index...aw&catid=70:developpement-des-raws&Itemid=321

It's in French. In a nut shell, what does it say?

 

[carstenw]

Loosely translating the conclusion:

- The ranking of the best software for the A900 is very different to the A700.
- DxO beats IDC and ACR handily.
- Bibble Pro 5, Beta 1.1 is very good too, with better detail, but worse noise handling.
- Lightroom is far from being the best for high ISO.
- With all the software, ISO up to 800 yields good A2 prints.
- At ISO 1600, A3 or A4 prints are good with all software.
- The author had some NoiseNinja (built into Bibble) licence problems, and feels that if he could have used it fully, it would have been much better.
- With some software, noise management reduces colour saturation.

Finally, there are some links to mosaics of crops like this for ISO 1600:

http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/A900_1600ISO.jpg

And two links to raw images taken at ISO 1600 and 3200:

http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/ISO3200.ARW
http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/ISO6400.ARW

 

[edwardkaraa]

It's in French. In a nut shell, what does it say?

Marc,

You have to click on the buttons for each converter and see the difference for yourself. I do not necessarily agree with the author's conclusions, hence each should decide for himself depending on how he likes the different interpretations.

 

[edwardkaraa]

In the introduction, the author claims that DXO results bring the A900 to an almost comparable level to the 5D2 and D3X. He believes that the origin of the high-iso noise of the A900 lies essentially in the software, hence the abilty of DXO to solve it.

 

[edwardkaraa]

- The author had some NoiseNinja (built into Bibble) licence problems, and feels that if he could have used it fully, it would have been much better.

Actually he says that since he has a full licence for Noise Ninja, he was able to benefit from the full support of this software. He did not try the basic outfit (which doesn't require a licence), so he's not able to say if the results would be comparable or not.

 

[fotografz]

I have been playing with RPP today and I am simply amazed with the detail it brings out compared to ACR. Maybe I am doing something wrong in ACR. If I am someone should tell me but look t these images.

You need to see the large images to really see what is going on. Look at the details in the green leaf on the main red bud and the detail on the purple flowers bottom left.

First one is RPP

Second is ACR

And full Size files can be seen here:

FULL SIZE RPP

Full Size ACR

If you open both larger files provided and extract the Exif information (RPP doesn't provide initial camera info, but all the details of both files are available under the Advanced tab). It appears that the ACR file has +25 color noise reduction applied and the PPR file has none.

I wonder how C1 does in this mix?

 

[Georg Baumann]

Exactly, I am missing C1 in that comparison.

 

[fotografz]

Loosely translating the conclusion:

- The ranking of the best software for the A900 is very different to the A700.
- DxO beats IDC and ACR handily.
- Bibble Pro 5, Beta 1.1 is very good too, with better detail, but worse noise handling.
- Lightroom is far from being the best for high ISO.
- With all the software, ISO up to 800 yields good A2 prints.
- At ISO 1600, A3 or A4 prints are good with all software.
- The author had some NoiseNinja (built into Bibble) licence problems, and feels that if he could have used it fully, it would have been much better.
- With some software, noise management reduces colour saturation.

Finally, there are some links to mosaics of crops like this for ISO 1600:

http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/A900_1600ISO.jpg

And two links to raw images taken at ISO 1600 and 3200:

http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/ISO3200.ARW
http://www.alpha-numerique.fr/images/stories/logiciels/ComparA900/ISO6400.ARW

Thanks!

Well, this is interesting to say the least. ISO 1600 on the A900 is a very important ISO for me because of the occasional need to open up backgrounds when shooting flash.

Of note, in this specific test, C1 Pro does about as well as LR except the greens in LR are slightly less noisy.

DxO seems to retain good color and contrast with less noise although not quite as much micro detail as C1 shows.

If this is even somewhat correct, could this suggest selection of different converters for different applications?

It would appear that DxO would be a good candidate for higher ISO wedding/portrait work where cleaner skin tones trumps micro detail.

Your thoughts?

 

[Braeside]

There is a problem with Dx0 and the A900 at present, shadows are green tinted. Apparently DxO know about this, there was a recent discussion here

 

[jlm]

i am having trouble seeing a significant or even discernible difference, even in the green leaf by the bud...a slight color difference, mostly in the green, is all I see

 

[carstenw]

Actually he says that since he has a full licence for Noise Ninja, he was able to benefit from the full support of this software. He did not try the basic outfit (which doesn't require a licence), so he's not able to say if the results would be comparable or not.

Ah, whoops, speed-reading again. Sorry. I was not aware of the existence of a basic module, and thought he was saying something else.

 

[carstenw]

If this is even somewhat correct, could this suggest selection of different converters for different applications?

It would appear that DxO would be a good candidate for higher ISO wedding/portrait work where cleaner skin tones trumps micro detail.

Your thoughts?

Being no A900 owner, take my opinion with a heap of salt, but unless you are printing for exhibition, I cannot imagine that the small differences between the best converters in the 100% crops translates to anything more than negligible differences in print. Small tweaks might even be able to almost completely nullify any advantage one has over the others.

 

[picman]

Exactly, I am missing C1 in that comparison.

I have both C1 and DxO. At the moment I somewhat regret having got DxO as I really prefer C1. Mind you, I never shoot above 400 ISO and usually at 200 ISO. My problems with DxO are multiple: slow (up to 10 times slower than C1), you cannot see important changes unless you magnify at least 75%, the colors just do not seem OK to me when compared to Aperture and C1. This last point now seems to be confirmed by that other thread. One thing it certainly has got going for it is the distortion correction which especially for the CZ 16-35 can be nice. But as far as chromatic aberration and purple fringing goes I simply seem to have better results with C1. It is not perfect though: no lens-specific lens corrections, no dust-tool and it has a few annoying bugs.


Cheers, Bob.

 

[fotografz]

i am having trouble seeing a significant or even discernible difference, even in the green leaf by the bud...a slight color difference, mostly in the green, is all I see

There is more of the cloth pattern of the fake magenta flowers bottom left. However, I wonder if its due to Color Noise filtration on the ACR file where the the other one has no color noise filtration. There shouldn't be any need for color noise filters at that ISO.

 

[ryc]

Looks like I will be trying the DxO tonight. ZJust based on the results I saw on the site comparing the converters. The question is how will it handle the detail like in RPP which was not compared.

 

[douglasf13]

Looks like I will be trying the DxO tonight. ZJust based on the results I saw on the site comparing the converters. The question is how will it handle the detail like in RPP which was not compared.

I wouldn't waste too much time with DxO until they get it fixed for the A900:
http://forums.dpreview.com/forums/readflat.asp?forum=1037&thread=32090670&page=1

It seems odd to me that, as precise and thorough as DxO claims to be, that they can't figure out the A900 black point??

As far as the RAW test Edward posted, IMO, the DxO results are WAY too NR'ed, and one of the problems with DxO is that, like some more popular converters, there is always some baseline NR applied.

 

[Quino Terceño]

To my eyes the the noise structure (or its absence) feels nicer on the ACR file. RPP's midtones are turning a bit muddier but the file is holding more detail in general, overall color seems better also.

 

[dhsimmonds]

This thread takes me back to darkroom discussions and comparisons of the plethora of developers for use with whatever ASA(ISO) film!

At the end of the day whatever works for you is the very best one....believe me!