The noise at the pixel level may close, but given the large # of pixels in a camera like the 7D your not going to notice the noise as much in prints vs say a 40D.
Also, the 7D appears to have vastly improved or even eliminated the high ISO banding that can be seen in 40D and 50D files, which is a huge improvement.
the 40D still has he cleanest deep ISO100 shadows of that triple in terms of annoying looking artifacts like banding of various types
but yeah the 7D has by far the least banding at very high ISOs
theSuede wrote:
The Y scale is in Ev, or "stops". And the metrics are, from top to bottom:
Dynamic range at Base ISO (FWC compared to read-noise)
Dynamic range at true ISO800, not the "ISO800" that the camera states
Base ISO signal to noise ratio at 18% of FWC true ISO200 - not camera "ISO200" - signal to noise ratio at 18% of FWC true ISO800 - not camera "ISO800" - signal to noise ratio at 18% of FWC
And lastly, signal to noise ratio at 1% of FWC, at base ISO. This is deep shadow noise at base ISO-
I see that I clipped the "%" sign in some of the legends in my diagram when copying/pasting... Sorry. 7D to be added soon, but I'm waiting for DxO to verify my "true" ISO-sensitivity measurements....Show more →
here is what i got, normalized for true ISO as best as i could and normalized for print comparison (not 100% view), don't know how 7D isos rate yet so gave three guesses, the 20D tested here is worse than most at low ISO and the 40D is a bit better copy than most, the 7D release copies seem to do worse than the pre-production model did:
all 7D seem to have a type of vertical banding at ISO100-800 or 1600 depending, some have it one stop worse than others, more realistically noticeable (when doing high DR and wanting lots of shadow detail) only from ISO100-400 (or 800). when it does show it is the ugliest looking type of banding yet.
at high iso the 7D is ultra free of banding of any sort
jorkata wrote:
I understand your point. It has been discussed many times.
But to me, higher resolution = bigger prints.
If there's more noise at the pixel level on the hi res sensor, then this noise will be more visible in the bigger print as well.
In summary, noisier hi-density sensors give you the following options:
- same size prints with less visible noise
- bigger prints with more noise
That's a dubious proposition, as far as I'm concerned.
Ideally, you'd want bigger prints with same/less noise.
That's why comparing pixel-level noise is important. It tells you how much noise there will be in the bigger print.
it's not dubious since at low iso the noise is usually not an issue so it means bigger prints with noise not being an issue or at high issue it is even at worst no matter what you do really
jorkata wrote:
Hi skibum5,
How does mRAW look like in terms of noise - say, compared to the 40D?
i didn't get to look at mRAW yet
but carefully shooting some dogwood berries and leaves from a fixed position using the same lens with a 5D2 and 7D at ISO1600 I noticed in this reach limited image comparison:
7D has more reach but more noise and artifacts and CA and so on evident
if you downsample it to 8.2MP (which gives same reach as the 5D2) then you get a tiny bit better noise and detail and the same size looking berries as from the 5D2 so the 7D seems a tiny bit better for reach limited stuff even at high ISO
because of noise and small photosites being more sensitive to motion blur and af errors and so on at high iso it can be tricky to make images look nice across all 18MP, much easier to get them nice across as 21MP of a 5D2 under those conditions so it can be tough to get all that much of a reach advantage over the 5D2, but you can always shoot mRAW and do no worse and get a lot more fps
with care and good lighting you can reasonably get much better reach from the 7D, you simply can see a lot more tiny little details in far off stuff, even if using a 70-200 PLUS TC on both cameras.
if you are not reach limited the 5D2 universally blows away the 7D image quality though and it is easier to get the best out of the 5D2 than out of the 7D, sometimes quite noticeably easier
cameron12x wrote:
That's should not be unexpected. We're comparing apples to oranges here, in some respects.
This camera has a lot of things going for it when compared against other cameras at the same price point.
Cheer up!
Yeah, it's only to be expected. I mean the 5D2 collects light over a 2.6x greater area! That is ridiculously hard to make up for.
And you could always say that the 8fps, fast trigger, short mirror blackout totally blow the 5D2 out of the water too.
You wouldn't expect the new G11 to not get universally blow away by the 7D would you? (and that difference G11 to 7D is noticeably large than 7D to 5D2)
It's not like I said the 50D univerally blows the 7D out of the water when not reach limited or even said the 50D universally does a tiny bit better.
And one persons blown out of the water is another less dramatic persons does somewhat better.
skibum5 wrote:
if you are not reach limited the 5D2 universally blows away the 7D image quality
At what print size does the "blows away" difference matter? I'd bet it is at fairly large print sizes - but I'm just guessing as I haven't printed out samples from both cameras.
Taken in Photo Warehouse, Christchurch, ISO 3200 f5,6 on the 18-135 kit lens IS, exif should be in. This has had RAW conversion in CS4/acr5.5 (I think?) Noise ninja auto, smart sharpen then save for web with bicubic sharper set.
Good grief, just start shooting. There is absolutely no way of determining what this camera is capable of or not. Time will tell. Most of us haven't flipped the start switch to the on position yet. I have followed this thread and 95% of the content is worthless at this juncture.
It's a 7d, not meant to be the top of the line professional gear, nor is it a Rebel series. Let me tell you about noise. If you're going for a big print, chances are is this will serve you well. If you're exposure is good, then the print will turn out well. If you're going for a big print, you won't be shooting high ISO anyways.
Just keep in mind this is a 7D, not a Mark 7. If this camera suits your needs and budget, then enjoy.
jorkata wrote:
But to me, higher resolution = bigger prints.
Yes - because it is illegal to print at 25 ppi with a well-interpolated file. If anyone found out that you had made a big print that way you would go straight to jail.
That's why comparing pixel-level noise is important. It tells you how much noise there will be in the bigger print.
How much bigger? It's crazy to have to specify print sizes before considering relative camera performance. It's even crazier to define different print sizes for the two cameras. By that reckoning a Minox makes better prints than a Hasselblad. That's an absurd way of considering and comparing camera performance. When I take a picture, I really don't know at that stage how big the image will be printed. So the only sensible way to compare two cameras is to consider how the images will look when they print to the same size.
I believe high MP sensors are great if they can provide absolutely clean low-isos and good high isos. I would really like to see some raw converted low iso shots.
skibum5 wrote:
all 7D seem to have a type of vertical banding at ISO100-800 or 1600 depending, some have it one stop worse than others, more realistically noticeable (when doing high DR and wanting lots of shadow detail) only from ISO100-400 (or 800). when it does show it is the ugliest looking type of banding yet
I do not know why you have been going from one forum to another talking about YOUR banding problems. The reality is that it's hardly universal and is mostly a QC (so what else is new) issue. Some copies have it bad and others are OK. Mine happens to be quite good (also verified by Gabor).
Regarding dynamic range, I think the difference is minuscule and hardly worth discussing. It's not like 0.5 eV (poorer on the XSi/450D) or 1 eV (that MUCH better on the D90/D5000) difference. Of course, what's more saddening is the relatively poor DR of the 5D2 (despite all that hype) as compared to the D90/D5000.
Edit: One thing I've noticed from your charts is the poor DR on the 20D across all ISO. Ya, so much for the benefits of low MP count.
Finally some shots..
All shots are with EF800 ISL Handheld. About 50% crop. I have to say that this camera beats 1DMK2N to get the initial focus acquisition specially when I set the lens's minimum focus limit at 6mm.It has been a struggle for me with 1DMK2N and EF800 combo (to get a quick response I have to set the focus limit at 20m: and many times it is too long).
These are easy shots with blue skies but very impressed with the tracking(way better than 5DII or 50D). I used center point with AF expansion.
ruhikant wrote:
Finally some shots..
All shots are with EF800 ISL Handheld. About 50% crop. I have to say that this camera beats 1DMK2N to get the initial focus acquisition specially when I set the lens's minimum focus limit at 6mm.It has been a struggle for me with 1DMK2N and EF800 combo (to get a quick response I have to set the focus limit at 20m: and many times it is too long).
These are easy shots with blue skies but very impressed with the tracking(way better than 5DII or 50D). I used center point with AF expansion.
Nice shots to illustrate AF tracking.
No offense here, but the images seem a little plasticky to me. Is that because of jpeg settings? I will set the NR down to low.
Seemed the forum was only arguments, graphs and data with no source listed, and other academic arguments that tell us nothing useful. Just people arguing their belief. Useless dribble.
Images tell us something useful.
When I get my 7D I will post actual images. Something useful.
ruhikant wrote:
Finally some shots..
All shots are with EF800 ISL Handheld. About 50% crop. I have to say that this camera beats 1DMK2N to get the initial focus acquisition specially when I set the lens's minimum focus limit at 6mm.It has been a struggle for me with 1DMK2N and EF800 combo (to get a quick response I have to set the focus limit at 20m: and many times it is too long).
These are easy shots with blue skies but very impressed with the tracking(way better than 5DII or 50D). I used center point with AF expansion.
A bit concerned that the pictures are not pin-point sharp. Or is my monitor screen bad? Apologies if that's the case.
. 
