The 10mp sensors are lifeless

John Sheehy · Sep 28, 2006

aarif said:
I remember a while back looking at images from the canon 10D and
thought to my self these look better then the 20D.

Could exposure be part of the equation? The 10D meters for ISO 64 when set to "ISO 100", and the 20D meters for about 125, a difference of a stop. The 10D automatically exposes to the right for you, pushing the readout noise floor and the shot noise down, relative to the captured subject luminance.

--
John

John Sheehy · Sep 28, 2006

Edymagno said:
The reason I say this, is because of the "smoothness" of the entire
image and when "pushed" in ACR (more exposure) never showed
artifacts of any kind. In the 20 and the 5 the noise* becomes in
extreme cases patterns in the shape of a grid. Besides this, in the
5D sometimes I see micro banding when "pushed" slightly. Of course,
this is mainly seen in the shadows.

Edymagno said:
Is it noise ?

Yes. It's line-noise, instead of pixel noise. It is much more visible than pixel noise of the same statistical strength. Being that each line contains only blue or red pixels (but not both; always shared with green), it can be a bit chromatic.

Edymagno said:
Sometimes I think is like the same pixel pattern
position itself stamping its own on the capture.

It is different every frame, but the statistical strength is consistent, unless you get extra from electrical interference.

Edymagno said:
With the 10D never had any of these problems even when increasing
exposure more than 2, almost 3 steps. The noise always was
film-like, I mean random. Therefore more usable images and a lot
less retouching.

It maybe pixel count, sensor technology, internal interference or
external, I don´t know, but to my understanding, Canon delivered a
finer electronic image capture device 3 years ago.

The 10D actually has stronger line-banding, relative to max RAW exposure. You don't see it as much because at ISOs 200 and higher, the 10D has more random noise to mask it, and the 10D meters about a stop hotter than the 20D and other recent cameras, so that cuts the strength of all noises by a stop, relative to captured signal.

--
John

alanr0 · Sep 28, 2006

LVBrit said:
IMHO the higher pixel count lenses are better at finding flaws in
the optics. Olympus tell us that they had to go to extraordinary
lengths to get good quality from their high end lenses for the 4/3
system.

Agreed.

LVBrit said:
There is a potential knock on that may also be affecting picture
quaility. Many high megapixel count consumer cameras process the
image to remove blooming and CA. I wonder if the same type of
processing is creeping into DSLRs.

Possible, but CA correction in a DSLR would have to cope with lens to lens variations. If this was happening, I would expect the manufacturers to be shouting about how wonderful it is. They probably prefer you to become mildly unhappy with the kit zoom and then shell out for some more expensive glass.

Cheers.
--
Alan Robinson

Mike Birse · Sep 28, 2006

... such as different anti-aliasing filters?

John Sheehy · Oct 7, 2006

Sorry for the delay. I've been hibernating the computer for a couple months with some unfinished replies, and its time to get rid of some of these browser windows.

aarif said:
I remember a while back looking at images from the canon 10D and
thought to my self these look better then the 20D.

That's not surprising. The 10D meters for about ISO 64 when set to 100, while the 20D meters for about 125 at ISO 100. With blackframe noise at 1.9 ADUs in the 10D, and 2.07 in the 20D, that means slightly over a stop more usable shadows for the 10D, as metered by the camera. The 10D always exposed to the right for you. For JPEGs with normal contrast settings, though, that was bad news.

The 10D also had a CFA that allowed more light to be captured in the blue channel, by about 1/3 stop. That helps a bit in incandescent light.

As you get to the higher ISOs, though, the 20D gets way ahead of the 10D. The 20D's ISO 1600 has about the same level readout noise (relative to max signal; not to the metering) as the 10D's ISO 400.

--
John

Steven Hight · Oct 7, 2006

aarif said:
The thing is I have KM 5D and 7D and now I bought the Alpha 10mp
which is a good camera and captures excellent details but I still
prefer the KM IQ over the Sony. Also I have seen the images from
the new Nikon and was not impressed at all specially with the water
color look it has in higher ISO’s due to strong NR add that to what
I noticed with the Canon 10D and 20D images from the same guy
pushed me toward what I’ve said.

Now I know you do not agree but it’s just my observation and I
wanted to know if anyone shares that’s with me.

Thank you

I am interested in your observations regarding the quality differential between the 6mp and 10mp cameras. Once in a system the upgrade path is predetermined destiny, not undertaken without significant research on many purchasers part.

Nikon being the system, and a long planned for visit to New Zealand already locked in I struggled for over 3months to secure a D200 for the trip. Was it worth the hassle.

Check out the results and make up your own mind.

http://www.pbase.com/steven_hight/new_zealand_2006

Petteri Sulonen · Oct 7, 2006

The following are all actual-pixels crops converted with default settings from RAW files. One is shot with a 6MP dSLR, one with an 8MP one, and a third with a 12 MP one. Care to take a guess as to which is which?

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]

john farrar · Oct 7, 2006

Hard to tell but if I had to make a guess: 12, 6, 8.

I think there is what we might call in English a certain 'wishy-washy' quality to some images shot on the new larger sensors. 6mp appears to offer more 'punch'.

--
John.

John Doyle · Oct 7, 2006

I wonder what the answer to this little test is. I'll guess 6, 12, 10 but it's a blind stab in the dark.

I am curious to know as the first one looks the cleanest to me.

Petteri Sulonen · Oct 7, 2006

One out of three.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]

Jay Turberville · Oct 7, 2006

In order,

10M, 6M, 8M

But I'm totally unconvinced the difference is discernable. I figure any perceived differences are really from lens, focus and subject differences.
--
Jay Turberville
http://www.jayandwanda.com

Jay Turberville · Oct 7, 2006

I guess I meant...
12M, 6M, 8M

Jay Turberville said:
10M, 6M, 8M

--
Jay Turberville
http://www.jayandwanda.com

DavidMillier · Oct 7, 2006

It's a trick - they're all from the same camera and you are just trying to demonstrate that the mind games we play far outweight real world differences...

Petteri Sulonen said:
The following are all actual-pixels crops converted with default
settings from RAW files. One is shot with a 6MP dSLR, one with an
8MP one, and a third with a 12 MP one. Care to take a guess as to
which is which?

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]

--
Galleries and website: http://www.whisperingcat.co.uk/mainindex.htm

Petteri Sulonen · Oct 7, 2006

That would be lying. I don't lie, as a general rule.

I'll give the answers in due course...

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]

aarif · Oct 7, 2006

Just wanted to say that you cannot judge by looking a couple of snaps and come to a conclusion. I talking about personal experience over a few years, I feel with current technology Pixel size has a minimum to maintain image quality but in future of course things change

Sony DSLR-A100
1/250s f/4.0 at 150.0mm iso200
;

Konica-Minolta Maxxum 7 Digital
1/160s f/4.0 at 200.0mm iso200
;

Canon EOS 5D
1/800s f/2.5 at 135.0mm iso320
;

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

DietmarP · Oct 7, 2006

Petteri Sulonen said:
The following are all actual-pixels crops converted with default
settings from RAW files. One is shot with a 6MP dSLR, one with an
8MP one, and a third with a 12 MP one. Care to take a guess as to
which is which?

Yep.
6, 12, 8.

andrewD2 · Oct 7, 2006

If you had a similar crop to the 7D shot the A100 shot would have about 4MP left.

Andrew

Petteri Sulonen · Oct 7, 2006

The answer is 6, 12, 8. Two of you got it right. I'm a bit disappointed, though, that the initiator of the thread ducked the question instead, and simply claimed "years of experience" instead.

My personal take on the matter is that yes, I think there is a difference between the way the three cameras render stuff at the pixel level -- but the difference is very, very subtle. Certainly nothing worth starting threads about. But that's just my opinion, as usual.

Petteri
--
[ http://www.prime-junta.net/ ]
[ http://p-on-p.blogspot.com/ ]

gheil · Oct 7, 2006

Sony has no superchip, they cannot create the dynamic range of big sensels with much smaller ones! It is Sony that made all these (D80, A100 and K10D) 10MP sensors. The camera makers are being lead down the Megalpixel niny garden path by the Emperor, Sony. Canon with its 400D is just playing along with the game "see we can make 10MP too, hope you like it". Japan Inc is playing us all for fools. Someone has to say it: "The Emperor has no clothes".
--
http://public.xdi.org/=greg.heil

Kenstrain · Oct 7, 2006

My "emperor" has adequate clothing. It seems to have about the same light collection efficiency as previous models. So I can bin shadow pixels if I want to. No loss, and some gain for highlight resolution.
(It is a 400D.)
Ken

The 10mp sensors are lifeless

Forum Pro

Forum Pro

Senior Member

Active member

Forum Pro

Member

Forum Pro

Veteran Member

Leading Member

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Forum Pro

Veteran Member

Member

Veteran Member

Forum Pro

Senior Member

Forum Enthusiast

Keyboard shortcuts