7D vs 5D2 noise

[pknight]

I think it's pretty obvious that the 5D2 versions are more highly sharpened.

Thanks. I was hoping somebody would point this out.

 

[Jose Rocha]

He has lost all the credibility by the his comment on 10-22.

I have never seen such faith outside of religion: It's blasphemy for anyone to comment on 10-22!

Have you ever considered the possibility that the reason for that might be that the 10-22 is a lens with great saturation? And you are simply wrong?

I agree, the 10-22 is the best WA lens I have ever played with and produces pleasant saturated colors with very nice contrast, a little better than my Sigma 10-20 EX which is very good. It's strange how some complain about the lack of saturation and contrast...

--
Jose Rocha

http://olhares.aeiou.pt/jplacebo

 

[Image Obsessed]

looking at this thread, it seems the 5dII noise is quite a bit better?

Check the color patch!

http://luminous-landscape.com/forum/index.php?act=attach&type=post&id=16696

 

[RedFox88]

50D was introduced at $1399 and the price dropped quickly to $1200. 7D is only $200 more expensive.

It was $1299 body only when it came out, as the 7D is being offered body-only, so a $400 difference. Even with the numbers you stated above, $1399 is *$300* less than $1699 (7D price) not $200.

The 40D was $1299 at introduction. The 50D was indeed announced at $1399.
I remembed this very well.

The info I was able to find about its initial price says differently as I stated before.

 

[GeorgeML]

The info I was able to find about its initial price says differently as I stated before.

From the Canon 50D press release:

The Canon EOS 50D Digital SLR camera is scheduled for October delivery and will be sold in a body-only configuration at an estimated selling price of $1,399.00 .

http://www.usa.canon.com/templatedata/pressrelease/20080826_eos50d.html

 

[John Sheehy]

I agree, the 10-22 is the best WA lens I have ever played with and produces pleasant saturated colors with very nice contrast, a little better than my Sigma 10-20 EX which is very good. It's strange how some complain about the lack of saturation and contrast...

My Sigma 12-24 is pretty good. I bought it because I have the 5D2, and it is the widest FF zoom avaialble at a reasonable price. It is a bit slow, and needs to be stopped down further for the corners of full-frame, but is pretty sharp once you do. I miss using the Canon 10-22, though, because it is so light and easy to work with, especially on the 400D, which is also very light.

When I get my 7D, I will probably start using the 10-22 again, when I leave the 5D2 home.

--
John

 

[John Sheehy]

The IQ of the 7D up to ISO 3200 is stunning and sets a new benchmark to beat. I'd like to hear about AF performance from new users.

But Roger Clark says that current cameras are shot-noise limited, and that sensor and especially pixel size determine IQ!

I have been saying for a long time that read noise is the real obstacle to low-light DSLR shooting, and the 7D continues to support my belief.

It is easy enough to simulate a camera with only shot noise; just take an almost noise-free image, and apply poisson noise to it, scaled to the number of photons you would expect - even P&S sensor sizes are marginally useful, without read noise, at ISOs into the hundreds of thousands, and good enough for newspaper images up into the millions...

BTW, the 7D should be better than any APS-C pushed to any ISO, well above 3200 (until something with less lead noise comes along).

--
John

 

[Rickard Hansson]

Otherwise i must say the results are quite good.

--
Rickard Hansson
Sweden

 

[Biggis]

It is easy enough to simulate a camera with only shot noise; just take an almost noise-free image, and apply poisson noise to it, scaled to the number of photons you would expect - even P&S sensor sizes are marginally useful, without read noise, at ISOs into the hundreds of thousands, and good enough for newspaper images up into the millions...

ISO > 1'000'000 ??? Nice
But wouldn't that be like shooting in the dark (i.e. no colors)?

I'm not an expert (pretty far from) but is there enough visible light there, at those ISO values (given that they are taken at reasonable shutter speeds...if that matters?!?), that we would recognize as 'natural' if it was ever recorded?

Other wavelengths are not so easily translated to the visible spectra (e.g. NIR/SWIR)
--
BigGis

 

[rwbaron]

Otherwise i must say the results are quite good.

It appears to me there was a difference in focus.

Bob

--
http://www.pbase.com/rwbaron

 

[carlk]

That’s a very good point. Shot noise is no mystery. It’s constant and easy to figure out. I wonder how many photons will fall on a 15sq-um area under, say, EV0 condition. It would be great if it is still far from shot noise limited as you said.

The IQ of the 7D up to ISO 3200 is stunning and sets a new benchmark to beat. I'd like to hear about AF performance from new users.

But Roger Clark says that current cameras are shot-noise limited, and that sensor and especially pixel size determine IQ!

I have been saying for a long time that read noise is the real obstacle to low-light DSLR shooting, and the 7D continues to support my belief.

It is easy enough to simulate a camera with only shot noise; just take an almost noise-free image, and apply poisson noise to it, scaled to the number of photons you would expect - even P&S sensor sizes are marginally useful, without read noise, at ISOs into the hundreds of thousands, and good enough for newspaper images up into the millions...

BTW, the 7D should be better than any APS-C pushed to any ISO, well above 3200 (until something with less lead noise comes along).

--
John

 

[Peter 13]

One thing nobody's talking about is the new metering system. Canon's previous Evaluative metering is embarrassingly dumb, fooled by virtually all and even minor backlighting, even if you're focused on a nearby subect (no distance information taken into account). I've always been impressed by how well the D300's Matrix Metering works, clearly using the subject much more in the metering calculation. I hope the 7D's entirely new system works as well.

I have exactly the opposite complaint. Since I tend to use one AF point most of the time (not only the center one), the exposure is too dependent on the subject that I focus on.

 

[John Sheehy]

It is easy enough to simulate a camera with only shot noise; just take an almost noise-free image, and apply poisson noise to it, scaled to the number of photons you would expect - even P&S sensor sizes are marginally useful, without read noise, at ISOs into the hundreds of thousands, and good enough for newspaper images up into the millions...

ISO > 1'000'000 ??? Nice

On a G11, ISO 1 million would mean that 0.9 photons would strike each green-filtered pixel, at the whitepoint, on average. That's a total of about 6 million photons for the entire image at the whitepoint (extra losses in the red and/or blue channel); ample for a mediocre small image.

But wouldn't that be like shooting in the dark (i.e. no colors)?

I don't understand the reference. There is no human-like special pixels for luminance in low light; a Canon sensor detects three images; three B&W images with different color filters. The empirical world that cameras operate in is not affected by human vision mechanisms.

I'm not an expert (pretty far from) but is there enough visible light there, at those ISO values (given that they are taken at reasonable shutter speeds...if that matters?!?), that we would recognize as 'natural' if it was ever recorded?

In the absence of read noise (and the little bit of dark current we get with normal exposure times), which was the context of my comment, we are just counting photons in dark areas, as we are in bright ones. Less photons = more noise, relative to signal.

Other wavelengths are not so easily translated to the visible spectra (e.g. NIR/SWIR)

???

--
John

 

[jrscls]

One of the biggest reasons I went FF (5D MK II) was so that I could use my 70-200 f2.8 IS lens for portraits the way it was inteneded. Right now, there is no 1.6 crop equivalent which would require a 44-125mm f1.8 IS lens to get the same FOV and DOF.

The 7D will be great for general purpose photography, especially wildlife and sports, but unless there are much better faster lenses in the portrait range (at a reasonable cost) then I will always keep a FF body.

 

[carlk]

That's how I felt about my 70-200 2.8IS too. On the other hand I have 85L and do like the crop when I shoot telephoto. Someday Canon may make a 70-200 equivalent mid-tele for crop. There is a good possibility now that it gets serious on APS-C.

One of the biggest reasons I went FF (5D MK II) was so that I could use my 70-200 f2.8 IS lens for portraits the way it was inteneded. Right now, there is no 1.6 crop equivalent which would require a 44-125mm f1.8 IS lens to get the same FOV and DOF.

The 7D will be great for general purpose photography, especially wildlife and sports, but unless there are much better faster lenses in the portrait range (at a reasonable cost) then I will always keep a FF body.