New A900 noise?

[David Kilpatrick]

I didn't shoot anything over 800 - partly because we were told clearly enough that the performance would not be correct. Even at 800, the JPEGs show the same sort of processing which applies to the A700 with NR, and I know that does not mean anything - until I can see the raw files properly, with a converter not copying the in-camera JPEG method (like IDC may do), I'm not drawing any conclusions.

First: the twin BIONZ is identical to the A700. Second: the on-chip process is identica, and the 24 megapixel sensor shares the architecture of the IMX021 according to Sony. Third: the pixels are physically larger.

They insist that in the final release, the noise at high ISOs will be better than the A700. It's clear to me that the firmware in the A900 I tried is not the same as V4 in the A700. Therefore, I suspect that v4 for the A700 is effectively later than the firmware in the pre-production A900s. They may even be trying it out on the A700 first.

I think all this discussion is meaningless since no-one has had a final version A900. I assume the body given to Duncan McEwan was the best they could possibly find and had the latest upgrades. I have seen his pictures and he is sending me both JPEGs and raw conversions on DVD soon. He had to leave to run a photo imaging workshop at Inversnaid and said he might not get the files to me before next week, but they are the shots shown off by Sony, and whether or not any other magazine or website has requested them, we get them. When I have a chance to examine Duncan's meticulous shooting first hand (not on a projection screen) I'll know a bit more. But it looks as if the bottom line is sometime in October, final production cameras hit the market, people start using them on real subjects hopefully with a choice of raw converters, and the image quality including real world noise levels will be seen.

David

[David Kilpatrick]

Duncan was shooting FOR Sony - I'm sure he got the pick of the crop. The review sites probably pestered Sony offices for an advance sample and got allowed to try whatever they had, on the condition they did not post high ISO etc noise samples.

They then went and did exactly that. Of course!

The Bionz is identical by the way - only a single Bionz is used at a time. The process pipeline is duplicated, and the CPU picks whichever 'side' has got furthest to queue the next image to. We were told there is no difference at all in the processing. You'll see this when Shirley eventually unravels the workshop session where the technical staff answered questions.

David

[Octupi]

I actually believe DK is correct in that the firmware was not set on the pre-production models that most received. It takes quite a bit of time to develop these firmware upgrades and test on the various country models and confirm they are correct. The disc of images that DK is waiting on most likely has the last phase of the firmware in place, just prior to release of the firmware.

The firmware we have for the a700 (v4) is what the production firmware for the a900 will be. This I believe to be true and am not shaking my head. So much judging is going on here and at Dyxum about the quality of the images, yet we have not seen any images from a RAW converter that supports a900 RAW yet. I know it is human nature and self control is not...but I think we all need to expand our horizons and practice a little.

Two things could happen.

1. Production models of the a900 show up and yield the results we expect and we say, told you they would be better.

2. Production models of the a900 show up and those that flamed the original images find the nearest hole to crawl in because they cried of unspilt mile.

In time, in time...we will have the actual results. Until then all we are doing is beating a dead horse.

[PhotoTraveler]

Ok, can we calm down on the mass of links.

So far, I'm impressed with the output. I went through the DPR JPEGs and thought they looked pretty good. The A900 is not a D3, don't know why people try to compare it. Further, it's not a camera for sport shooting, it's clearly a camera where the core users will put it on the lowest ISO all the time anyways.

For pre-production, non super tuned camera, it looks good to me, oh yeah and it's a 24MP FF camera for 3K with SSS and a huge view finder. It is what it is, and that's pretty darn nice. It's not the first camera to have some rough outputs at this point.

The big potential I see is that while some things could be noisy, it's very fine noise. I think some of these shots could look great in B&W. Now if Sony were to ever go for broke on luring people in, make a A900M (mono), rip that bayer pattern off. This thing would be insane.

I don't mind some noise when going to black and white. I have purposefully shot my 7D at high ISO just to "noise it up" for the shot to make a nice B&W later.

Anyrate, we need to wait till early october for a new ACR to come out (adobe will launch one after the Photokina camera wave has happened). Then we can get some much more solid understanding of what we have here.

[marktuma]

Don,

I thought all we'd seen so far were previews, not reviews? Has anyone done an actual review with the final version, or is that not possible yet?

Mark

[Dr. Harout]

It cannot be the final because the tested cameras were pre-production models. So let us wait a while.
As I said previously have patience.

[Octupi]

LOL...fanboy I am not. But to each his own.

Personally, don't care if you've had 300 sony/minolta cameras. The fact is, as Dr. puts it, we need patience. No sense continuing an argument on assumptions, we will see when RAW processors have the code in place and the camera has released firmware.

[KevinBarrett]

....it took Sony one year to give us a700 firmware 4. Let us prey that it does not take another year to bring the a900 up-to-speed.

I wonder if it means anything that you typed "prey" instead of "pray." Just teasing.

[Dr. Harout]

LOL...fixed.

Ha ha ha, you didn't need to fix, it sounded much better...