The question is, what is noise reduction and why is it bad. DxO looks for noise reduction by looking for inter pixel correlation. If there is no inter pixel correlation, there is no smearing of detail between pixels, and it is not something to be worried about. If some clever manufacturers have found a way of suppressing noise without introducing smearing between pixels, then why is that anything but a good thing?
True, but the V1/J1 still got caught by DPReview:
If their results are complex, then why reduce the information into a single score? There are many critics (not just about cameras, but lenses, audio gear etc etc) out there who refuse to reduce their findings into a single number; they instead choose to lay bare all their findings. To simplify something that cannot be simplified is tantamount to deliberate distortion of information.
Actually Bob...it's DxO that ranks the cameras...not us. It's DxO that ranks the Nikon D3100 not only above the Canon 5D2 and Canon 7D....but ranks it higher than some Medium Format Digital Backs. Not if you're claiming that "those who know better" can go through the numbers to pass judgement, I'm all ears as to why the Nikon D3100 is better than the 5D2, 7D or MFDBs.The silliness is in the presentation of a single score to appeal to the simpletons who can only assess things on the basis of a single number. The DxO data provides quite enough information for those who know what their own priorities are to pass their own judgement.Indeed. But they paid NO heed to it.
They continued to score and rank the V1/J1 sensor when they should have instead discarded all the INVALID low light ISO results and not rank the sensor at all. They jolly well know those are PROCESSED results which do not in any way reflect the sensors true performance.
If that is not poor (I must say, highly biased) reporting, I do not know what is.
--
Bob
To appeal to simpletons, as I said.
It's an unwise choice. The simpletons still argue and it discredits the rest amongst the others.
Then, DXOMark which supposedly tries to evaluate bare SENSOR performance (NOT post-processing abilities) should have discarded all these high ISO results. At the very least, no low light ISO score should be awarded in such instances. To award a score is to deliberately mislead their readers especially since DXOMark is aware they are no longer evaluating bare sensor performance but manufacturers' post-processing abilities.
I never said anything else. i think DxO is very silly to try to appeal to those incapable of understanding their measurements by condensing them to a single score. However, how that score is arrived at is made plain, so it is easy to see. As to why the D3100 scores higher (which isn't the same thing as 'is better', that's straight forward:Actually Bob...it's DxO that ranks the cameras...not us. It's DxO that ranks the Nikon D3100 not only above the Canon 5D2 and Canon 7D....but ranks it higher than some Medium Format Digital Backs. Not if you're claiming that "those who know better" can go through the numbers to pass judgement, I'm all ears as to why the Nikon D3100 is better than the 5D2, 7D or MFDBs.The silliness is in the presentation of a single score to appeal to the simpletons who can only assess things on the basis of a single number. The DxO data provides quite enough information for those who know what their own priorities are to pass their own judgement.Indeed. But they paid NO heed to it.
They continued to score and rank the V1/J1 sensor when they should have instead discarded all the INVALID low light ISO results and not rank the sensor at all. They jolly well know those are PROCESSED results which do not in any way reflect the sensors true performance.
If that is not poor (I must say, highly biased) reporting, I do not know what is.
--
Bob
Probably the ones from the 5DII and the Hassy, as DxO scores suggest that they would.
Except that they don't say what you say they say, so who's the joker, the site that gives the 5DII and MFDB a higher score than the D3100, or the person who says that they don't?
its e/v from FF is -1.722 stops, and Canons APSC is -1.37 stops, so in making high ISO comparisons if theyre on the same technology there should be 0.344 stops between them.
well Ive had my own issues with DxO, and found it lacking. One of my own cameras, an E5 is reported there to have a DR of 10.5 stops, yet in review it will pull 10.5 in jpeg. Under IMATEST similar 4/3 sensors bring 11.5 stops of DR. Thats a stop over the so called engineering limit
Well, that website is there to make money. It is not silly to cash in on stupidity. Smarter people would understand the reasons and would excuse DXO. Examples like this can be found everywhere around us.
Figured out that 79 > 67?
The average is not relevant for all photographers as you have to know what is most important for you- colour depth, dynamic range or low light performance
DxO is making an effort to provide "Spec sheets for dummies". Well, I've seen ISBN 978-0471915904
ok lets suppose they do not
That says it all. In Imatest there is an option to use dcraw or LibRaw to get it work right on raw data without any noise reduction or other processing applied in conversion. That, of course, does not take care of lighting, focusing, and exposure skills.
yet I have to say again
