A noisy world on the A77

[David Kilpatrick]

At ISO 200, with some bumped up exposure in pp, I thought - oh no, that's much too noisy, the A77 really is not much good even at low ISO.

willfuldamageall.jpg (160.16 KiB) Viewed 6678 times

Then I took a closer look at the image, all over. I was not looking at noise at all in the red tones of the notice. The A77 with CZ 16-80mm was actually resolving the inkjet dots of the printed sign, as the depth of field proved, at f/8 and a hand-held 1/125th, 80mm, with SSS. In fact there was no significant image noise and the level of sharpness was just that high.

willfuldamage-clip.jpg (179.75 KiB) Viewed 6678 times

100% clip. I'm beginning to appreciate that even though very few shots need 24 megapixels when I get out for travel work seriously with this camera and lens combo, I'll have some of the sharpest and largest files available.

David

[Dr. Harout]

You got me worried from the heading...

[redsim74]

You got me worried from the heading...

Me too!

Very few print methods involve continuous tone. We generally don't jam our eyeballs up to the newspaper (if we can even focus that closely) and complain about the halftone dots.

[bossel]

I think 6 megapixels is more than enough for me. But if I get 24, why not

[bossel]

Are you sure its not noise? A inkjet printer should use CYMK and there are no red dots! But then, I did not see the original scene...

[David Kilpatrick]

The 'noise' goes out of focus a bit further up. I don't know what process would be used to make the sign, but our local signmaker has inkjet printers which use a much larger inkset than CMYK - I think eleven waterproof inks which can match any Pantone shade.

David

[stevecim]

I bet you someone at DPR will pickup the second shot and try and use it to prove how bad the A77 is

[stevecim]

I get the feeling there are quite a few people over at DPR complaining about noise in photos , even at iso100, which is not really sensor produced noise.

[artington]

AP has just published an extremely positive review of the A77. However the resolution / noise comparison shots at ISOs up to 12,800 clearly show that resolution and noise at any given ISO are mostly better for JPEG files than for Raw. It would be interesting to know whether this reflects some pretty hefty image manipulation going on inside the camera or whether it's down to poor raw conversion. The article does not say which raw converter is used so I presume it's Sony's own.

[David Kilpatrick]

I think it could be down to measuring the noise. The JPEG noise suppression involves some pretty fierce smearing of colour values and smoothing of luminance, so RMS noise might appear low but actually the image looks very soft. Sharp edges and smoothed out textures. From raw, you get visible grain which can be controlled carefully without losing detail or a clean look. But it measures with worse RMS noise values.

David

[Mal]

Yes. I was taking a few days' break and got the A77 the day before, so shot Raw + JPG to be safe. The jpgs have a great tonal quality, but some of the fine detail appears smeared, for want of a better word. Any detail which has contrast (twigs etc. in a landscape) are beautifully sharp, but some shots on a sports pitch with not very good grass on it (astroturf?) are just a mush. The raws (ISO 800) show some noise but incredible sharpness. I'm more than happy - I can choose the trade-off between noise and detail. I remember when the A700 was released the fuss that was made over "cooked raw" because noise reduction was applied to them. It seems Sony can't win!

[twm47099]

Yes. I was taking a few days' break and got the A77 the day before, so shot Raw + JPG to be safe. The jpgs have a great tonal quality, but some of the fine detail appears smeared, for want of a better word. Any detail which has contrast (twigs etc. in a landscape) are beautifully sharp, but some shots on a sports pitch with not very good grass on it (astroturf?) are just a mush. The raws (ISO 800) show some noise but incredible sharpness. I'm more than happy - I can choose the trade-off between noise and detail. I remember when the A700 was released the fuss that was made over "cooked raw" because noise reduction was applied to them. It seems Sony can't win!

But the question is why didn't Sony offer jpgs with NR off as an option in the A77. One concern with the A700 cooked raw was could obvious NR even be eliminated by a firmware patch (answer was yes). Since raw on the A77 doesn't suffer from "cooked raw", it seems like the way to prevent the 'mushy jpgs' comments would have been to have a NR off. For a camera at the 7 level, that should have been a no brainer. Sometimes Sony seems to be their own worse enemy.

Related to Sony decision making not NR. On the flicker Sony club, a Sony representative had asked for questions to take back to the design team. She got a lot and after a while Sony responded. One question and answer struck me:

Question: "Why is there only one memory slot on the A77."
Answer: "Dual card slots are not usually found in cameras below $2K."

Now the A700 cost less than $2k (essentially the same as A77,) and it had dual slots - although not with all the features desired. So instead of designing a camera to a level of user (entry, mid, advanced/semi pro, or whatever your categories are,) they design it to the level of the competition at a price point. Maybe that means there will be an advanced user APS-C camera at the $2k level??

tom

[Birma]

Related to Sony decision making not NR. On the flicker Sony club, a Sony representative had asked for questions to take back to the design team. She got a lot and after a while Sony responded. One question and answer struck me:

Question: "Why is there only one memory slot on the A77."
Answer: "Dual card slots are not usually found in cameras below $2K."

Now the A700 cost less than $2k (essentially the same as A77,) and it had dual slots - although not with all the features desired. So instead of designing a camera to a level of user (entry, mid, advanced/semi pro, or whatever your categories are,) they design it to the level of the competition at a price point. Maybe that means there will be an advanced user APS-C camera at the $2k level??

tom

Hi Tom - I agree that's a disappointing answer. The top plate LCD, mag alloy body, weather seal, etc, etc, suggest this is the class of camera that would expect dual slots. I would respect an answer of "we had to make choice on features for a price point and this one didn't make the cut" more.

[David Kilpatrick]

I think the reason is slightly more significant. The A77/65 are the first cameras able to use the SDXC UHS-1 card specification with up to 95MB/s data transfer. I've been doing card tests this weekend, I am waiting for a 45MB/s card and hoping SanDisk will send me a new generation double that speed.

Card writing speed often depends on the physical slot mechanism used - it incorporates controllers. Most of these assemblies are separate replaceable parts and are bought in from third party makers, or other divisions of the company. It's more than likely Sony was not able to source a twin slot device, as used in the A580 (SD and MSPro) which offered the required SDXC compatibility and UHS-1 compliance.

So they used one which did what they needed for data throughput.

David

[Birma]

That makes a lot of sense David. I hope that means we can look forward to dual slots again in the future. The A9xx class will I suppose needs even faster data capture rates with those massive 36Mb files to heft about .