the present standard, although it comes from another derivative (that had less industry support) didnt come into play until after Oct 20 2006
http://www.dpreview.com/reviews/olympuse330/
http://www.cipa.jp/english/hyoujunka/kikaku/pdf/DC-004_EN.pdf
this document precedes that but I can prove it if doubted
So what do you make of the e-410 then, which has "hot" ISO 100 and lower than expected ISO 1600?
http://www.dpreview.com/reviews/olympuse410/page16.asp
Even the more recent e-420 is off at ISO 100:
http://www.dpreview.com/reviews/olympuse420/page16.asp
If what you were trying to prove is that, that was because the e-330 came before this document, surely the e-410 proves that logic wrong. I have corroborated what dpreview pointed out too, I have one.
ok, how many cameras do DxO make, and how many photographs do they take?
Not sure why DxO has to make any cameras to understand the engineering issues involved in dynamic range- at least far better than preview. As for photos they have to take a bunch if they want their RAW converter be worth anything.
are you telling me that Andy is wrong about the the nature of DxO testing as it pertains to the standard on this matter, b/se thats an argument you best take up with him.
I don't have to take it up with him at all. I don't have to do anything. Andy is not an authority in this subject and DXo is. It's that simple. But for the record, I have taken it up to preview before to Phil and Simon.
Irrespective of that, your camera has an ISO measurement that conforms to the standard, while nobody argues how well a fit that is between jpeg images, its a fact that most popular RAW converters are amenable to that same specification although this is relatively recent.
The fact that RAW converters when they don't support cameras show the truth of what is going on- that the ISO's are pushed is very telling. Now, you can comply with the standard and get that exposure at the respective ISO, but what will happen is that cameras that are actually more sensitive to light/better sensors will show less noise. This is why the 620/e-30 and forward have so much more noise than the resolution increase from 10-12 megapixels would have suggested. An E-3 has noticeable more shadow range than the E-30 and forward. The real reason is that ISO push on the other cameras, which comply with the ISO standard as you say but then you get that noise.
This is why Dxo normalizes the sensitivity and talks in the graph about signal to noise ratio. This is important, because better sensors will have a better signal.
I think arguing that DxO are the only ones right and everyone else is wrong is quite silly. DxO are right for themselves, everyone else conforms with the standard
I don't argue that DXo are the only ones right. There are other reviews that mention the noise in the E-5 and some that don't but then they go ahead and compare JPEGS between cameras. At that point a better JPEG engine will do a better result and it's well known the Olympus JPEG engine is at the top (Fuji's the other). But as far as the sensitivity of the sensors, you can even forget about conforming to an ISO then- it's the measured signal to noise ratio with standardized light. I don't know how much more accurate it can get for a comparison and the other things they mention they do.
I dont buy that at all, from yourself its simply user derived opinion, its not testing.
You are more than welcome to say that, though I provided full instructions on how to reproduce some of the testing I do myself, only to be called names, which quite frankly, doesn't seem very logical to me.
As to the difference between DxO values and those of actual camera manufacturers on review, one is data, the other is photographic derived data, which is more important?
Normalized sensitivity data is important because that's all the data you get at a given intensity to do any derived data. The derived data varies greatly in JPEG but not in RAW as far as being exclusive to each brand.
if they dont, why are the data points different. You already know the answer,
Please refrain from making suppositions as to "what I supposedly know" that you want to prove and allegedly implicitly agree with because that's simply nonsense.
its ot a matter of the sensor and pipeline, its a matter of how the tests are constructed and only that. IMO its a huge blunder on DxO's part for the usefulness of their data, and the correlation to other available data.
It's actually quite useful. Like I said, I have been able to corroborate their expected noise/iso's/DR relative to the cameras I have. They do not cover banding situations though.
Of course I said several replies back that one easy way to see all of this is for yourself, just download RAWS of the E-5 at ISO 3200 and A900, play with them, push pull, see how has much more shadow control and upon resizing down, which looks better.
likewise fitting DR to a curve optimised to the highlight end will fit some cameras better than others, hence this leads to our previous discussion of IMATEST data and the proximity of DxO data
Not really. This is all pretty linear and it's arbitrary where the manufacturers set it to.
obfuscation, you were given the opportunity twice, 3x if you include this. I gave you fair warning and its a fair call on my part, and your concluding section here gives the answer to this anyway
I am sorry but I don't believe you are in any shape or form to call it this way, more so when you can't even show due process yourself with a camera like I did with the 620 and draw.
there you go with this assumption that DxO are right, and the rest of the planet is wrong. But FWIW there is some easy math on the noise value of downsampled image sizes. If in this event you went from 24Mp to 12Mp that works out to the root of (24/12) or 1.41
I don't see how that changes anything, or has to do with much
- Raist