Photoclubalpha

Hi All

Does anyone here have a A57, thinking of upgrading my A550 to one, was impressed with DPR review, but did they get it right?, looking at their RAW image test, the A57 seem to be miles better than the A65 at iso1600 and higher, is there really that much difference in real world photos?.

Cheers, Steve

I'd take a look at the real world images to decide on that. The 24mp sensor in the A65 and A77 is pretty noisy at high ISO lots of chroma noise. So be in no doubt for low light work the 16mp CMOS is a better bet all around. Saying that some say the A77's raw files clean up pretty well. Again I'd download a few from various sources and play around.

Right now I'd say everything beats the A77 at high ISO, even the old 12mp CMOS I've used on various models was IMO better in low light.

Yes, there is. That's why I sold my NEX-7 and bought another NEX-5n. The 1600 sensor is versatile and robust in low light, the 24 megapixel is tricky to get right and does not like difficult contrast and light.

David

Dear David
Is there anything wrong about Sony's image processor?
Anyway a57's IQ makes my decision to go for Nex6 more justified.

David Kilpatrick wrote:Yes, there is. That's why I sold my NEX-7 and bought another NEX-5n. The 1600 sensor is versatile and robust in low light, the 24 megapixel is tricky to get right and does not like difficult contrast and light.

David

Nothing wrong with Sony's processor really - it's all the laws of physics. The 16 megapixel sensor is also a later generation than the 24, in the NEX system.

See the current flood of firmware updates coming from Sony.

http://www.photoclubalpha.com/2012/10/1 ... -77-nex-7/

Everything is likely to be getting better. At the moment, we've only seen new firmware for 24 megapixels and that has changed the balance slightly, but we can expect new firmware for some of the 16 megapixel models too which may put them ahead in many ways (again).

David

I wonder if the firmware has helped the high ISO noise of the A77. Not to be silly but I felt the 24mp D3200 was not quite as good as the 16mp CMOS..though not "loads" worse either. A77 surprised at how much of a difference there is. I have to say even being willing to try the SLT thing..I just can't really see myself being the high ISO shooter I am even looking at the 24mp Sony models (either of them) the 16mp sensor looks loads better.

Even taking into accout half a stop of light loss SLT wise (it might be a bit less just a ball park number) it does not really explain why these models are so noisy. Bit of a buyer turn off, well it is if you do a lot of low light shooting.

I use a NEX-7 and 5N, In most circumstances I shoot at nothing higher than IS0 400, but found ISO 1600 quite acceptable with the NEX-7. In film days when I rarely went higher than 400ASA and I never thought I was missing anything - that's what tripods are for. Just using the cameras with small NEX primes, Samyang 8mm, Sigma 19 and 30mm and Sony 50mm OSS (I like a purist approach) Must say that I have been really pleased with both cameras and the quality obtained. If you go by DPR reviews (and they seem fairly thorough) you will see that the NEX 5N has higher quality scores for image quality RAW, Image quality JPEG and much better high ISO performance than the SLT-A57. Also the NEX-7 has much better high ISO performance than the SLT-A77. So for me I now feel I've reached a plateau with my cameras including the A900 and I'll now wait and see what happens and enjoy using them. Some wider angle NEX primes would be nice. I was hoping for something more radical from Sony with its new full frame camera I must admit.

I'd caution putting too much into DPR they have questionable testing IMO, though you can look at the studio comparisons at high ISO if you want to. DxO also for some reason rate the A77 higher than the A57 for high ISO, yet the images just don't show this in the real world.

That 16mp sensor can give you a reasonably decent ISO 6400 if you don't goof the exposure up. In practical field use I would say there is not a lot in it between the D7000 I had and the A57 for high ISO work, yet again DxO say differently. One thing is certain Sony's 24mp sensor is just not up to scratch for low light shooters in the A65/77 it's consistently beaten by just about everything out there. I would say both models need new sensors asap.

It is curious to a degree when you have a look at the comparison between the three as to just how the A77 manages too get a better overall score than the A57 when its sensor doesn’t better that of the latter in any of the individual measurements, (although it was apparently level pegging in a couple), the D7000 is just slightly better than both in the various measurement choices although I expected it would be slightly more than the results show due to the light loss in the SLT design.
http://www.dxomark.com/index.php/Camera ... nd3)/Nikon
Greg

Greg,

Those figures the DxO spits out are absolutely useless, as far as modern larger-sensor cameras are concerned. The fact is that they measure noise at 18% grey, and all decent sensors still capture enough light at that level at least up to ISO12800. As a result, the noise there is almost pure natural photon noise, and not the electronic noise added by the camera itself. All larger modern sensors have about the same quantum efficiency, the SLT mirrors lower that by about 0.3EV, and that's about it. So those DxO measurebators are putting them all pretty close together, with only exception of the incorrect interpretation of ISO sensitivity by DxO, which adds some uncertainty into the results and further removes them from reality.

The really objectionable noise is the very unnaturally looking electronic noise that actually may only become dominating in the image at levels below 5%. And that's exactly where the a77/65 look absolutely abominable.

Ah ok agorabasta, I didn’t research exactly how DXO calibrated the test and its scope, I just assumed that to be relevant the snr result would have to be global in nature and then averaged down to a final standard on a scale. The problem is with that ‘result’ of theirs thrown into the mix it then affects the other results as well.
But even making allowances for that incorrect testing procedure the A77 ‘result’ is looking to be incorrect by their own standards when I couldn’t find anywhere that the DXO findings on that camera were superior to the A57. (In order to justify the higher overall test score)
Greg
Ps ‘abominable’ already…ha ha! (snowman) you gotta laugh at the overall aptness of that one agorabasta

Greg,

There are two problems with DxO. First, measuring noise at 18% doesn't measure the camera-specific noise, in measures only the natural noise always present in the light as it is due to the photon quanta granularity. Second, they refuse to use the proper definition of ISO. Utterly useless they really are, the same as all the freckled punk geeks in this world.

Their higher rating for A77 vs 16Mp sensors is simply because 24Mp capture more detail, so there's more of an 'image' in the capture, and there's also that ISO50 that really helps with the best attainable quality.

Hmm it looks like that leaves us with a problem agorabasta, where is there a site that actually does scientific evaluations of sensor capabilities and camera system comparisons?
Greg

Greg,

There's none, but you knew that already. The more scientific types work for monies, and there's none for them in this business. They all work deep inside the paying corporations and they don't disclose too much in public.

I just scan through the samples available and normally have no problem to make up my mind off what I see.

DxO can't live up to the real world. And I've used the D7000 a fair bit in low light (and the D90) both rank higher than the A57 on that site. Now appreciate the light sucking mirror and all which will influence things to a point (though the sensor might be updated a bit over the D7000 one I have no idea)

I compared the D90 to the D7000 on DxO. There is no way the D90 has better high ISO/low light performance than the A57..not a hope. Not to say it's bad far from it but ISO 3200 with some care is it's limit, and then you're skating on thin ice with colour reproduction at times. ISO 6400 forget it (well forget it for decent print sizes) DR wise there isn't a hope in hell the D7000 has 1.5 stops more DR than the D90 shot both extensively and yes you can do that old 16mp CMOS pull the shadows up big time, highlight end I found the D90 actually was better even taking exposures into account.

I'd be a lot more interested in what goes on real shooting rather than some iffy French site that tries to "measure everything" and my conclusion is that the A57 is very similar in low light performance and DR to the D7000, if there are differences it's not enough to worry about for practical field use. I'd always look at the images and raw ones when looking at these cameras, and I've compared results and I can't really say how DxO can get such odd numbers. Onto the 24mp CMOS appreciate you can clean up the chroma noise a bit but only a blind man would say the low light performance is better than a 16mp CMOS camera.