The real story behind the SONY ALPHA noise

[DapperSquirell]

Cogent, nonhysterical case you make.

It seems to me the real triumph of the A100 is it's capacity to produce professional results when handled with a professional's skills and yet remain accessible to the tyro, producing very good pictures for the beginner as well. Some may call that "entry level". Some may object because no Druid's license is required. So be it.

At the end of the day the success or failure of the Alpha line will be judged by the quality of photographs taken by those from the entire population; good photographers and bad. That's the nature of a mass market. And I believe that's a good thing for this Art and Science we call photography.

May not be such a good thing for internet trolls, though.

--
Daryl

 

[LeeR]

Tangible
A adjective
1 real, tangible

possible to be treated as fact; "tangible evidence"; "his brief time as > Prime Minister brought few real benefits to the poor"

2 physical, tangible, touchable

having substance or material existence; perceptible to the senses; "a > physical manifestation"; "surrounded by tangible objects"

3 palpable, tangible

capable of being perceived by the senses or the mind; especially capable > of being handled or touched or felt; "a barely palpable dust"; "felt sudden > anger in a palpable wave"; "the air was warm and close--palpable as > cotton"

4 tangible

perceptible by the senses especially the sense of touch; "skin with a > tangible roughness"

5 tangible

of especially business assets) having physical substance and intrinsic > monetary value ; "tangible property like real estate"; "tangible assets such > as machinery"

I hope this helps...

So, which of these did you mean? Do you mean Sony shots are more capable of being touched? More perceptible than other brands? Can you explain?

Lee

 

[_Al]

I used the word tangible to discribe that the pictures out of the Sony a100 have a percieved feeling to me that you can touch the subject in the photograph sort of like almost 3D..

--
Sony Alpha A100
Sigma 28-70mm EX DG
Sony DT 18-70mm
Sony Hvl-F56AM

Check YoSelf Before Ya Wreck YoSelf

 

[Mark K]

David

I do not aware you shoot SD10 as well. Sigma has sent out test units of SD14 for some time. You may have a trial. Heard jpg quality is less acceptible

For a short window I will have our loan test D80, my own A100, and
my daughter's new 400D all side by side - though I'm off to
photokina on Monday. The 400D is supposed to arrive tomorrow.

I would not swap my A100 for the D80. Wonder how the Canon will
answer that possibility... not that I will have a chance, it has a
home already.

And last week I had to buy another Sigma SD10. I broke our SD10 by
doing a wrong procedure for sensor cleaning, and having the shutter
close at 30 seconds. The rigid vacuum wand of the Green Clean kit
has a kind of moulded lip (like a fly's mouthpart) and the damn
thing caught the shutter as it closed - a brush or even a wipe
would have caused no damage, but this is a hard plastic probe and
it flicked a leaf out of place. Has to go back to Japan to be fixed.

When it comes back, we'll use one of the bodies for infra-red and
mono and remove the filter.

David

--
Mark K
http://www.pbase.com/herbridgemo

 

[junMel]

Why are we so worried about digital noise? If anyone here still continue using film cameras (like I do with my Dynax 9) you would be convinced that digital camera noise and film based camera noise (grain) are both in parallel due to ISO standards. If digital camera noise is much lower than film camera noise (grain) basing on the same ISO setting then I would send my Dynax 9 to my private museum.

The noise on pictures printed 8"x10" Kodak professional paper are inherent in my A2 and Alpha 100 set at ISO 400 and are comparable, side by side, with the pictures printed 8" x 10" with my Dynax 9 using Kodak ISO 400 professional film. So what's the difference? I still use ISO 400 setting for reason that my uncles love Kodak Tri-X's "grain" for adding some "romance" to their B&W portraits. They are those ludites who never fell in love with very sharp pictures and very vivid colored portraits.

One day, my uncle showed me a very sharp picture of a basketball match and he told me to go inside his processing room (lights on of course) and to have a look at the picture. From outside he yelled at me, "where is the hurrah?" I agreed with him instantly, the determination on the face of the player to go past the guards and dunk the ball was so realistic! But where is the hurrah? The sharp picture fell dead silent.

Come on, let's get used to it... LOVE THAT NOISE!!!

 

[David Kilpatrick]

David
I do not aware you shoot SD10 as well. Sigma has sent out test
units of SD14 for some time. You may have a trial. Heard jpg
quality is less acceptible

Have they? Not in the UK. We will be the first to get an SD14. No samples have reached the UK so far.

David

 

[Two Truths]

Why?

Since when was it that every dSLR camera produced perfectly clean ISO 1600 images?

The Sony Alpha is no worse than the Nikon D70 or D200.
It isn't a Canon, so what?

--
Stuart / the Two Truths
http://www.flickr.com/photos/two_truths/
http://two-truths.deviantart.com/gallery/

 

[Hellashot]

Digital is far better than film for grain/noise. My brother who is a long time film shooter hates to use even 400 speed film. While I won't hesistate to use iso1600, 3200, or 3200 -1 on my 5D. Oh, and no NR needed from my shots unless I want a larger print (beyond the 300dpi native image resolution)

 

[Mark K]

KM 7D and 5D are capable of doing so

Why?

Since when was it that every dSLR camera produced perfectly clean
ISO 1600 images?

The Sony Alpha is no worse than the Nikon D70 or D200.
It isn't a Canon, so what?

--
Stuart / the Two Truths
http://www.flickr.com/photos/two_truths/
http://two-truths.deviantart.com/gallery/

--
Mark K
http://www.pbase.com/herbridgemo

 

[Ken_5D]

The 5d/7D are 6mp sensors... there is in this generation of sensor a 10mp noise penalty.

What came out in the D80 review is that the A-100 has amazing detail with its extinction resolution being 25% higher than the D80 or 400D.

And that level of detail you can't get on a 5D or 7D.

There are trade offs.. for some the 6MP strenghts are still best for others the 10mp is best.

------------
Ken - KM 5D
Sigma Trinity 10-20, 24-135 f2.8-4.5, 70-300 APO DG
http://www.cascadephotoworks.com

 

[AAK]

The superior resolving power of the A100?
I don't understand this.
Same chip as D80
Same pixels, same resolution.
Resolving power is identical. Period.

Now as to the things that "blur" the final image? Such as less sharpening or more NR?

Turn down the NR on the D80 (or turn it off - it is controllable, it's up the user) and you'll get the identical level of detail.

Most important, you're talking .JPG. In fact, this entire discussion is not about the resolving power of the camera (the sensor), but the difference in how the cameras generate .JPGs.

How about comparing RAW in the same converter for both the Nikon D80 and the A100?

The RAW file has no extra sharpening and no noise reduction at all.

Bottom line: You'll get the same picture from both, assuming that the lenses used are comparable.

The A100 is an excellent camera. But there is no magic in this camera that gives it more "resolving power" than any other camera with the same chip.

That's just wishful thinking!

Ok I know this has been discussed many a times over but I gotta
tell you that from my observation of the Nikon d80 canon 400d and
the Sony A100 Photographs
the a100 is the master of detail in this class of camera, if you
think otherwise go and have a look at the canon 1DS mark II review
in this website and you will see that canon have used noise
reduction far more conservativly in their profesional line just
like sony have on their A100 camera the camera might not have
weather sealing and all metal body but the photographs are
professional quality..

Also if you notice in most nikon and canon reviews including
dpreview review of these cameras they are fitted with the 50mm
prime lens when being reviewd however the SONY A100 does not get
this treatment so its no way a fair comparison.. I say do a review
of the sony A100 with the 50mm 1.4 prime lens on then you will see
the true resolving power of the A100 and the benifit of the
conservative NR and lets not forget if the camera decides the nr
level than the detail is gone for ever but if you decide than for
example in five years time software like noise ninja will have
advanced so much more that the in camera NR of 2006 will be
inferior when compared to software based nr of 2011 on the exact
picture! The sony a100 Photographs look more tangible then the
competition IMHO!
--
Sony Alpha A100
Sigma 28-70mm EX DG
Sony DT 18-70mm
Sony Hvl-F56AM

Check YoSelf Before Ya Wreck YoSelf

--
=~ AAK - http://www.aakatz.com
=~ Author of the H-Series White Paper
=~ http://www.aakatz.com/h1whitepaper [/U]

 

[AAK]

The 5d/7D are 6mp sensors... there is in this generation of sensor
a 10mp noise penalty.

What came out in the D80 review is that the A-100 has amazing
detail with its extinction resolution being 25% higher than the D80
or 400D.

Wishful thinking. I think the A100 is a wonderful camera, and I'm a big Sony fan. But I don't believe that test or your 25% claim mean anything at all unless you're shooting .jpg at default settings.

I never shoot .jpg, only RAW.

I never shoot at default settings - point-and-shoots are real good for that, not DSLRs.

The A100 looks better in that test, but the A100 achieves it through more sharpening and less NR in it's .jpg processing.

But the D80 has the same chip as the A100 and, in reality, identical resolution with comparable lenses.

Shoot RAW. Process both in ACR or Bibble, you'll get a surprise.
The images will look virtually identical.

All you're discussing here are the idocyncracies of the .jpg engines on the two cameras and the difference in their default settings.

--
=~ AAK - http://www.aakatz.com
=~ Author of the H-Series White Paper
=~ http://www.aakatz.com/h1whitepaper

 

[aarif]

There is no story behind it. Its simply a good camera but with few flaws that I hope will be ironed out in the future models.

Regards

--
You're welcome to visit my favorite Gallery
http://www.pbase.com/aarif/favorites

 

[Ken_5D]

The superior resolving power of the A100?
I don't understand this.
Same chip as D80
Same pixels, same resolution.
Resolving power is identical. Period.

Now as to the things that "blur" the final image? Such as less
sharpening or more NR?

Turn down the NR on the D80 (or turn it off - it is controllable,
it's up the user) and you'll get the identical level of detail.

Most important, you're talking .JPG. In fact, this entire
discussion is not about the resolving power of the camera (the
sensor), but the difference in how the cameras generate .JPGs.

How about comparing RAW in the same converter for both the Nikon
D80 and the A100?

The RAW file has no extra sharpening and no noise reduction at all.

Bottom line: You'll get the same picture from both, assuming that
the lenses used are comparable.

The A100 is an excellent camera. But there is no magic in this
camera that gives it more "resolving power" than any other camera
with the same chip.

That's just wishful thinking!

A) Has Sony or Nikon ever confirmed that they are the same sensor?

B) The number of sensors is the pixel resolution.. it is not the final measure of the sensors ability to RESOLVE Detail. Unlike film the entire surface of the Sensor is not light sensative.. there can be variations in sensor site size and the lenses used over the sensorsight to gather light.

C) I would assume that Phil isn't judging this with jpg output which can be manipulated..

If it doesn't matter why test it? Lets just assume all 10-MP sensors and cameras are the same.
------------
Ken - KM 5D
Sigma Trinity 10-20, 24-135 f2.8-4.5, 70-300 APO DG
http://www.cascadephotoworks.com

 

[David Kilpatrick]

But the D80 has the same chip as the A100 and, in reality,
identical resolution with comparable lenses.

Shoot RAW. Process both in ACR or Bibble, you'll get a surprise.
The images will look virtually identical.

Phil's resolution test uses raw not JPEG (I thought he was still using JPEGs but he's not, I was put right on that some time ago). ACR can not process D80 files so you can not have seen the result.

Using those programs which can process both types of file which I've got, Capture NX, Silkypix and Capture One Pro, I can assure you the higher extinction resolution of the A100 makes a substantial difference. It produces raw files with considerably more textural information and finer detail than the D80, at ISO 100.

The advantage is reversed at about ISO 320 - beyond this, the Nikon raw file is far better in the shadow areas. But the Sony still beats it in well-lit, fully exposed parts of the image even at ISO 1600. The effect is not pleasant in some ways, the Nikon image is much more even in detail and sharpess from shadow all the way to highlights. The Sony image suffers considerable loss of quality in the shadows and jumps into sharpness (and lower noise) in the brighter areas.

I've been comparing D80 and A100 for three weeks now. I would not swap the A100 for the D80 but I surely wish the A100 did a lot of things as well as the D80 does.

David

 

[David Kilpatrick]

The superior resolving power of the A100?
I don't understand this.
Same chip as D80
Same pixels, same resolution.
Resolving power is identical. Period.

Now as to the things that "blur" the final image? Such as less
sharpening or more NR?

Turn down the NR on the D80 (or turn it off - it is controllable,
it's up the user) and you'll get the identical level of detail.

RTFM - you can't disable NR. When set to OFF, it still operates at high ISOs, and there is a fixed level of NR constantly in operation even at low ISOs.

Most important, you're talking .JPG. In fact, this entire
discussion is not about the resolving power of the camera (the
sensor), but the difference in how the cameras generate .JPGs.

How about comparing RAW in the same converter for both the Nikon
D80 and the A100?

Done it. The Sony has a very different quality and type of detail. It's using a different AA filter and imposes different gain values on the Bayer matrix image. Hence the higher noise.

The RAW file has no extra sharpening and no noise reduction at all.

In both cases, it does.

Bottom line: You'll get the same picture from both, assuming that
the lenses used are comparable.

The A100 is an excellent camera. But there is no magic in this
camera that gives it more "resolving power" than any other camera
with the same chip.

That's just wishful thinking!

It isn't. The A100, used for raw and not JPEG capture, has a different implementation of raw image storage and processing. To start with, the Nikon uses RGGB raw data in order RGRGRG then a line of GBGBGB. This is then smoothed and compressed in the camera's processor, which always attempts to minimise noise. Sony uses RGGB data in order RRRRR, GGGGG (one line) then GGGGGG, BBBBBB (one line - the lengths are the total number of red pixels in a readout line etc) and compresses without any smoothing of values, though the compression is slightly lossy, which will mean that extremely small value changes are quantized.

The two cameras are completely different in the way they process and store raw files, let alone produce JPEGs. They uses different methods of NR and file compression. They have different AA filters.

'Same chip' means very little in this instance. If it did, then every camera using the same Sony chip would have been identical in performance in the past - and they never have been.

David

 

[David Kilpatrick]

There is no story behind it. Its simply a good camera but with few
flaws that I hope will be ironed out in the future models.

Actually, there is a story behind it, and it's got a lot to do with the requirements of the DRO+ processor built into the camera, as well as some decision Sony made about AA filtering, raw file format and speed of compression. They have the fastest 10 megapixel camera as far as buffer-write-sequence etc stuff goes (the specifications published are not accurate and are exceeded with decent cards - they seem to have assumed use of the CF card adaptor with Memory Stick!), and they have the highest residual detail resolving camera in its class. They also have the highest dynamic range of 8.7 stops, but as Phil's test removes noise threshold, his test shows 8.1 stops. The 8.7 stops is there if you are willing to accept the noise level. And it is used by DRO+, which needs access to extra headroom, and does some NR of its own while digging into darker tones.

Because they made these decisions, based on incorporating a feature which many of us do not or can not use (I for one, as I shoot raw files for everything) the A100 is stuck with a unique character. It could be seen as a mix of flaws, or a mix of benefits. It is certainly different.

If they read these forums, you can be sure they will think hard before going down the same route.

By 2008, A100s could be in big demand as rostrum artwork copy cameras or for extreme sharpness freaks - because maybe the next model won't be quite the same, and will be more like the the D80.

The Canon 400D did not arrive today. It will probably drop on my doorstep when I am flying to photokina. It may just (regrettably) have the best qualities of the A100 (sharpness) and the D80 (smooth noise-free images) combined.

David

 

[aarif]

Thanks David,

They really should have left the AA Filtering as it was with the KMs and not go for this extreme resolution which I think has more cons then pros, in portraits you can notice the heavily shaded areas showing some strange characteristics(over-saturated to dark) even at ISO 200 and that is exaggerated at 400, areas like near the hairline and around the eyes to me that is one of Alphas major flaws.

Just imagine a 10mp 7D image now thats what I wanted

Regards

--
You're welcome to visit my favorite Gallery
http://www.pbase.com/aarif/favorites

 

[AAK]

You're right, no ACR. Use Bibble Pro for the same test.

Can anyone link me to a RAW file that demonstrates this remarkable 25% extra detail?

Thanks!

--
=~ AAK - http://www.aakatz.com
=~ Author of the H-Series White Paper
=~ http://www.aakatz.com/h1whitepaper

 

[AAK]

I'll buy into the AA filters. Ever since the D70, Nikon has been paranoid about NR. They had a lot of complaints.

But sharpening and NR in RAW? There is a tiny bit of sharpening as part of the initial data processing stream. But no matter what the sharpening settings are in the camera, they are not carried through in the RAW when opened in non-Nikon converters.

But you're wrong about the sharpening and NR in RAW. Open a D80 file in Bibble Pro. You'll see tons of noise, and no noise reduction. The camera DOES apply chroma NR to .jpgs, and Nikon Capture does the same.

But it's not there in the RAW. Not in any Nikon RAW file. I know, I've post-processed probably 50,000 Nikon Raw files in non-Nikon converters. And believe me, there is no NR. I know, I have to deal with the noise.

Which, BTW, looks very much like the A100 noise!

BTW, don't get me wrong, I am very impressed with the results from the A100. But, like Phil, I am very skeptical that, ultimately, "there's much to it".

--
=~ AAK - http://www.aakatz.com
=~ Author of the H-Series White Paper
=~ http://www.aakatz.com/h1whitepaper