Dxomark has A700 v4 up now

[bakubo]

http://www.dxomark.com/index.php/eng/Im ... /Alpha-700

[Dusty]

"When we first analyzed the Sony A700, our preliminary tests showed that it made use of green-channel RAW NR (Noise Reduction). To ensure fair ranking on dxomark.com, we only score cameras that do not exhibit any sort of spatial signal filtering, taking care to test every camera we report on......"

-Sonolta

In other words, we only test cameras that use processes that we approve, and if they achieve good (or bad) results using another process, we discount that.

THAT'S what we mean by fair, balanced and objective testing!

Dusty

[Yagil Henkin]

not all that surprising, with V.4 - but there is still some low-light gap.

[Javelin]

Don't even Canon and Nikon do some stuff to their raw files? like Nikons clipping the shadows for example? or is that something else?

[Dusty]

You must be kidding! LOL! What a joke!! ----

It did not take me to long to remember why I do not have the time on my schedule for this forum. Too many uninformed users making too many IDIOTIC posts on too many subjects that they do not have a clue on...

-Sonolta

Again you miss the point, Don. I wasn't complaining about v4 firmware, I was remarking on DXOLabs practice of not testing aspects of the camera if they don't use the methods that THEY think best. If someone would come up with a new process that DXO doesn't approve of, even if it's results were 6x superior to everyone else's, they wouldn't include it in the tests. That's not objective, so again, they fail the objectivity test.

Faulty testing methods means that you have to always take their results with a grain (or teaspoon) of salt.

Dusty

[Javelin]

Wow the A350 come out better as low light camera than the A700. doesn't that go against what people generally say? and the A700 ties the 5DII in dynamic range

[01af]

Who cares about DxOMark anyway?

-- Olaf

[Dusty]

@ The folks that missed it the first time.

"To ensure fair ranking on dxomark.com, we only score cameras that do not exhibit any sort of spatial signal filtering, taking care to test every camera we report on."

FTR...Obviously Olaf cares about DXOMark, otherwise he would not bother to bash DXOMark at every oppurtunity.

-Sonolta

To ensure fair ranking, they should deal with whatever the camera outputs as RAW. If someone were to embed an automated version of Photoshop in their cameras, and it tweaked the RAW pre-write, DXO should use that output to test against because that's what the user has to work with. Anything else is injecting their version of expected results into the testing.

Dusty

[Argonaut]

"To ensure fair ranking, they should deal with whatever the camera outputs as RAW."

That sounds reasonable to me. If DXO wants to be even more fair than they already are [g] then perhaps they should test the complete camera, not just one mode of operation. If the a700's original NR puts it behind (or ahead of) the competition, then let's hear about it.

I mean, DXO says the site is a "... database of objective digital camera image quality measurements ...", and then they say they will test only unprocessed RAW data. That means they are testing the sensor, not the 'digital camera image'. IOW, their database is (1) incomplete and (2) subjective, as far as image quality measurements go.

Hey - guess what! DXO itself makes "The best RAW converter on the market". And you can buy it from them! Only 124 euros and up! I'll bet it can make those RAW a700 shots sit up and beg. No need for any lousy in camera processing.

No conflict of interest here, no sir.

[stevecim]

But they don't match real world results, if you go by the Dxomark, it shows that there is around 10% differance between the A350 and a700, D300 and since the a700 and D300 cost double a350.
Makes the a350 look like a great camera, even at high iso the D300 is only around 12% beter and the a350 is better than the a700 at high iso, it leads both cameras in color depth and is less than 5% behind in Dynamic Range. With such close results you would expect that most humans would not be able to tell which photo came from what camera.

I have an a350 and have down loaded a lot of a700 raws and above iso400 the cameras are miles apart. sure the a350 can look ok at iso800+ at 6x4 prints but it really depends on the subject matter.
there is no way the output of the a350 is with in 10% of the a700/d300, and the differance becomes visable at iso400 up. Clearly the dxomark results give no indication on what a final photo will look like.
it gives no clues on how the FW will process the images. With real world results, I "think" the a350 has great colors and "seems" to have a good dynamic range but it really comes down to subject matter and lighting.

I also use DxO raw converter, and it's great except for DRO, IDC and PS CS2 are lights years ahead of dxo when it comes to DRO. I can recover lots of details from shadows with IDC or PS CS2 (after converting the sony RAW to DNG raws) but can not get any sort of shadow details using dxo ( maybe it's just me).

From what I can see, above iso400 the a700 and d300 produce photos that need less PP than the a350 and at iso800 and above the a350 is not with in 10% of the a700/D300, it's not even close, don't forget the glass in front of the sensor has a big impact of the final result has do subject matter and lighting.

cheers, Stephen

[bfitzgerald]

I believe the DxO numbers, as much as I do listening to a Joseph Goebbels broadcast..

I see little to back up their claims, and no measure of IQ either for DR, which is frankly astounding...looks like Don is their propaganda king! Hope they pay well!

[bfitzgerald]

There are no facts, just charts and numbers!

What bothers me is this..
All the cameras show a decrease in DR when you go up the ISO scale. This is not shocking, I notice this on the 5d..with ease..the DR narrows a fair bit. But not so with the A200, the DR does decrease, but not to a huge degree.

So do I believe some oddball graph, or real world shooting? Do you really expect me to buy the dumb DR numbers banded about on there..that the G10 has neg colour film beating DR, when it blows out at the sight of a midtone, let alone a highlight. These numbers are worth almost nothing at all to any photographer. I cannot take them seriously at all, sorry Don..but what is up there..just does not seem to have anything backing it up at all.

[Vidgamer]

There are no facts, just charts and numbers!

What bothers me is this..
All the cameras show a decrease in DR when you go up the ISO scale. This is not shocking, I notice this on the 5d..with ease..the DR narrows a fair bit. But not so with the A200, the DR does decrease, but not to a huge degree.

Are you looking at the same chart I am? I took Don's link, switched the 5d to the A200, and the G10, A200, and A700 all follow the same diagonal line for DR, pretty much -- just with one camera better than the next. But the DR does decrease for the A200.

So do I believe some oddball graph, or real world shooting? Do you really expect me to buy the dumb DR numbers banded about on there..that the G10 has neg colour film beating DR, when it blows out at the sight of a midtone, let alone a highlight. These numbers are worth almost nothing at all to any photographer. I cannot take them seriously at all, sorry Don..but what is up there..just does not seem to have anything backing it up at all.

I picked the LX3 and it does better than the G10, as I would expect. I selected the A100, and that does better still (fortunately!). What's so wacky about it? (It is interesting to see how different the ISO settings are. I guess manufacturers can set ISO whatever to be anything they want it to be. Just adjust the amp gain...)

[stevecim]

[quote="Sonolta"][quote="stevecim"]But they don't match real world results, if you go by the Dxomark, it shows that there is around 10% differance between the A350 and a700, D300 and since the a700 and D300 cost double a350.

LOL...The a700 and D300 outperform the a350 in almost areas not directly related to IQ...Like frame rates, viewfinder size, write speeds, handling, etc, etc, etc..this is what factors into the cost.

Agreed, but none of those things should affect the final output, I'm not saying Sony cameras are better than they come up on dxomark site , what I'm tring to state is that the dxomark show little differance between the cameras, but real world there is a lot of differance between the final output. There is more than 1 stop between the high ISO preformance of the a350 and a700/D300. Yes there seems to be a relationship between the results on dxomark and real world results of the camera, but you can not state that camera a is better than camera b just on the results of dxomark. there are camera's that preform better than their dxomarks and camera's that don't preform has good. The bottom line is their results are normalized, and the fact is when you normalize results you hide really good and bad results and you do not get the true result of any thing, just a normalized results.

if dxo want to put togther a really usefull site, they could make up some test jigs where they can measure the preformance of the raw sensors then compare it to the results they get out of the camera, then we could see how bad/good each company is at getting the most out of a sensor.


I have an a350 and have down loaded a lot of a700 raws and above iso400 the cameras are miles apart. sure the a350 can look ok at iso800+ at 6x4 prints but it really depends on the subject matter.

Yes, look at the charts and that is what the numbers are reflecting...a350 losing by a stop or so at least. Same goes for native size KM5D stuff..it kills the native size a350/a700 stuff. Unless you are printing HUGE high ISO the KM5d is a steal.

the differance between them, is the type/size/colour/amount/levels of noise and how the noise is handled by the firmware, there are nikon and cannon camera's that have lower results than the a350 but they have better high ISO output.

there is no way the output of the a350 is with in 10% of the a700/d300, and the differance becomes visable at iso400 up. Clearly the dxomark results give no indication on what a final photo will look like.

They say as much! Look at how close the a700 and D300 are and they should be close. Like I said your 10% is a stop or more at high ISO.
good, then you agree that the dxomark results have nothing to do with the final photo

No kidding. DXO tests are spot on and the reviewers agree with the DXOMark numbers.

Are they, for every camera? going by dxomark the a350 should be miles ahead of the Nikon D2X, yet looking at the reviews I could find online , at iso800/iso1600 the D2X looks miles better.

I also use DxO raw converter, and it's great except for DRO, IDC and PS CS2 are lights years ahead of dxo when it comes to DRO.


Stop talking 10%.. it's a stop of DR here, worse color there, less tonal range here, and more IQ squashing noise and NR over there. 10% is a joke...You need to study all of the plots...and the native ISO numbers.

That's not what their results say , thats what you get when you combine their results with the real world facts you have of the cameras in your head. you knew how the cameras preformed before you looked at their site, it just happens that some of their results match with real world results, a lot don't.


Cheers, Stephen

[dynaxfl]

I agree with Don here. I think DXO gives you a very good idea on what each camera can do, even though the difference between two cameras can be very little. I think its the reviewer to add other fact in, such as price/quality ratio, etc. But overall it gives a very clear and fair data.