brainiac wrote:
All cameras are magic. If you showed any one of 'em to Leonardo Da Vinci his eyes would pop out. I never want to forget how surreal is this voodoo that we call photography. You press a button and a view of a moment in time is recorded for posterity. Magick.
RDKirk wrote:
Dan, I disagree with your blog comment about Live View not being useful for portraits. Portraits are what I do, and I now use Live View nearly 100% of the time--any time the camera is on a tripod, and with Live View I make that any time I can possibly put the camera on the tripod. Live View works so well, I'll tripod the camera even if it's not necessary just to use Live View.
Even your point about getting from behind the camera to look at the subject directly is more applicable to Live View than to the viewfinder. It's certainly easier to glance briefly down at the LCD while otherwise interacting directly with the subject.
Perhaps counterintuitively--something you have to try to understand--putting the camera on a tripod with Live View is actually more liberating than being the tripod and thus being attached to the back of the camera.
I've just gotten a radio trigger for the camera so that I don't even have to stay in physical contact with the camera after I've set up the shot. I can move toward the subject, staying just outside the frame, and fully interact. This is working wonders for photography of children. ...Show more →
Interesting points, and I actually have thought more about my statement regarding portrait photography (which isn't my main thing, by the way) since I wrote that. I keep coming back to some photos I saw of Avedon doing the "American West" photographs - he seemed to be doing pretty much exactly what you describe: stepping out from behind from the camera and communicating directly with his subject.
The more I think about it the more I agree that your description of how to use Live View for certain types of portrait photography is exactly right. The only qualification I'd make is to point out that once you are away from the camera and making eye contact with the subject it really doesn't matter what mode you are in since you are not looking at the display... (However, I just added a footnote to my portrait comment at the blog acknowledging the point you make.)
(For anyone who wonders about my feelings about Live View on the 5D II, I like it a great deal and I rely on it a lot in my own work.)
BTW, what wireless controller are you using? Any upsides/downsides to this solution for you?
BTW, what wireless controller are you using? Any upsides/downsides to this solution for you?
Right now I'm using a cheapo from eBay. If this comes to be something I really depend on, I'll upgrade. I got the idea from a friend, the back of whose head you may see in recent Pocket Wizard ads photographing a little girl using a PW as a camera trigger.
After watching him work that way during a session--standing frequently within arm's reach of the child (his specialty)--I was amazed. Being in close proximity to the child without the camera in the way makes all the difference in the world.
The MK II is really a great camera, at the same time I think it could be so so much better if Canon decided to throw in another 100$ to better a couple of things.
pjbishop wrote:
RD, true, the pixels are closer together - there are more of them on the same size sensor, but the size of the photosites is also reduced in the 5D II. This is from DPreview's review of the II.
No, don't misunderstand the terms. But I was inaccurate with the terms myself.
Pixel pitch is the distance between pixel centers, and is the specification Canon usually gives.
But there is the overall pixel size that includes the "photo diode area" and the total "pixel size" that includes the transistors that surround each "photo diode area."
When people colloquially speak of "pixel size" in these "noise versus pixel size" discussions they're usually referring to the photo diode area.
The size of the photo diode area has not changed between the 5D and the 5D2. See the text and diagrams on pages 21 and 22 of the 5D2 white paper:
mh2000 wrote:
anyone who thinks the II is more noisy willing to post an example from both normalized for area?
No longer have 5D so side by side shooting can't be archieved. Well, if i have time, we can compare 1ds2 vs 5d2 at low iso. Any way, this is the shadow noise issue i am talking about. Just taking any 5D/5D2 low ISO headshot and zoom in person's hair to see the differences.
RD- I have the white paper. Canon says on pages 22 and 23,
"Compared to the original 5D, the gap between the on-chip microlenses on the EOS 5D Mark II has been narrowed while the photodiode area ratio (photodiode area divided by the pixel size) has been increased. The distance between the microlens and photodiode has further been reduced to improve light gathering efficiency, and new color filter materials have been added to increase light transmissions while retaining accurate color reproduction. These optical advances and technology improvements to the sensor, plus countermeasures that help suppress noise from other camera circuits, are responsible for the sensor's high signal to noise ratio, enabling higher ISO speeds and improved dynamic range at low ISOs. Those are impressive sensor improvements, considering that the original EOS 5D had larger 8.2 µm pixels compared to 6.4 µm pixels on the EOS 5D Mark II, and that pixel noise usually increases as size decreases."
I understood, and read the reviewers as understanding, this to mean that the modifications and new color filter materials on the 5D II sensor, in effect a net increase in the photo-diode area, were introduced for the purpose of bringing the noise ratio of the II into parity with that of the 5D's larger pixels. I was speaking of the size of the actual pixels but I think I understand the distinction you're making, that 'photo-diode area', meaning the pixels with their attendant transistors, is a more practical description in application when speaking of 'pixel size'.
mttran wrote:
No longer have 5D so side by side shooting can't be archieved. Well, if i have time, we can compare 1ds2 vs 5d2 at low iso. Any way, this is the shadow noise issue i am talking about. Just taking any 5D/5D2 low ISO headshot and zoom in person's hair to see the differences.
I was speaking of the size of the actual pixels but I think I understand the distinction you're making, that 'photo-diode area', meaning the pixels with their attendant transistors, is a more practical description in application when speaking of 'pixel size'.
By "photo diode area," Canon is talking about the light-responsive area of the pixel--which is what we are actually talking about when we colloquially banter about "pixel size" versus noise.
In these discussions, nobody is considering the "dead space" of transistors surrounding the light-responsive area. But of course, it is there and Canon is dealing with it.
Canon says: the photodiode area ratio (photodiode area divided by the pixel size) has been increased.
To discuss in round numbers, let's say the total 5D pixel--that is the photodiode area (the light-responsive area) plus the transistor area (dead space)--is 64 square microns and the photodiode area within that is 16 square microns, giving a photodiode area ratio of 0.25.
Then in the 5D2 (in round numbers), Canon reduces the area occupied by transistors, reducing the total 5D2 pixel area to 36 square microns but maintaining a photodiode area of 16 square microns, now having a photodiode area ratio of 0.44.
We've been bandying "pixel size" around pretty carelessly all this time as though "pixel size" was a concrete indication of the light-responsive area--but it is not. Moreover, we have been treating "pixel size" and "pixel pitch" as though they were the same thing--but they are not.
The photodiode area is the "photon well," and between the 5D and the 5D2, that has not been changed.
Sorry, but the number on angels dancing on the heads of these two pins seems to be very similar...
Dan
(And at 200% magnification no less! Where you must also keep in mind the fact that you are looking at different areas of the two images - 100 x 100 is a smaller portion of the 21MP 5d2 image than 100 x 100 on the 12MP 5D. Oh, and exposure is different in the comparisons.)
Someone wrote:
anyone who thinks the II is more noisy willing to post an example from both normalized for area?
Will Patterson wrote:
My old 5D does the same thing when you crank up the exposure. The 1DS does it as well.
All canon bodies have shadow noise at low iso, but not as bad as 5d2. Here is some (not side-by-side shooting) samples from 5D & 1Ds2. They do look cleaner than 5D2 previous samples.
let's me know you still want 5D2 & 1Ds2 low iso side-by-side shooting. Well, it won't make any difference from the ones that you have already seen. The image speaks for itself. I don't think we need to debate anymore on this. I went through the old 5d/1ds2 files and their shadow noise look much better than 5D2. We can say as much as we want but i am sure our eyes won't lie to us in this case.
Okay, here is a side-by-side example of 5D and 5D2, same subject shot at the same session. A 70-200 was on the tripod, I just switched the cameras. I also have an example with quick noise reduction.
There is a banding pattern to be seen in the 5D2 image, however it's not as ugly to my eyes as the reticulated look of the 5D noise. When noise reduction was applied, the pattern of the 5D2 smoothed out sooner than that of the 5D...in other words, the 5D2 noise could be handled with less detail loss than the 5D image.
Added: The noise in the 5D image also extends farther into middle and highlighted tones than on the 5D2 image, indicating a slightly lower noise level overall for the 5D2--more tones of the 5D2 are above its noise level than in the 5D image.
Great comparo, Dan, and one that validates my intention to get a 5DII-in a few years, or when my MKI dies.
Michael, don't know how many print, but I've started printing more of my favorite shots, and with the Canon 9000 I got a great deal on, I absolutely love the prints, more than the screen. Just a joy to behold, easy to do, and not too expensive if discriminating.
Braniac, the camera is truly a magic box, has always been my view of it as well. If anyone had told me 20 years ago how photography would be now, I'd have never believed it. It was fun before, but work, now it's 99% fun (for amateurs anyway).
Good thread.
RDKirk wrote:
Okay, here is a side-by-side example of 5D and 5D2, same subject shot at the same session. A 70-200 was on the tripod, I just switched the cameras. I also have an example with quick noise reduction.
There is a banding pattern to be seen in the 5D2 image, however it's not as ugly to my eyes as the reticulated look of the 5D noise. When noise reduction was applied, the pattern of the 5D2 smoothed out sooner than that of the 5D...in other words, the 5D2 noise could be handled with less detail loss than the 5D image.
Added: The noise in the 5D image also extends farther into middle and highlighted tones than on the 5D2 image, indicating a slightly lower noise level overall for the 5D2--more tones of the 5D2 are above its noise level than in the 5D image....Show more → http://farm3.static.flickr.com/2603/3816820992_ee32082c8b.jpg
I don't agree with this conclusion. The reticulation-style noise is present in both files, and will not be intrusive in prints since it is very small scale and random in both the x and y axes. You would need to make very large prints before it stopped looking like film grain and became intrusive, and it responds well to noise reduction.
However, if you look in the shadows under the chin and the dark area in the bottom right corner (hair?) you can see the beginnings of vertical and horizontal stripes in the 5D2 file. Unfortunately this criss-cross pattern isn't random in both axes, and forms stripes across the whole image which can easily become visible in smaller prints where you are pushing the limits of the camera. Here's an example of how the 5D2 criss-cross effect can be intrusive where the more random noise of the 5D wouldn't be: http://cyberphotographer.com/5d2/speech_criss_cross.jpg
I'm not saying that the 5D has more latitude in the shadows than the 5D2 at iso 100, but the 5D2 has a shadow noise problem that the 5D doesn't really have. Consequently one is unable to dig into the shadows in the same way without noise becoming a problem on the 5D2. To me the linked comparison actually looks like a win for the 5D because I can imagine the criss-cross texture harming the 5D2 image more than it harms the crop. Noise reduction doesn't work well on the criss-cross pattern, because the stripes are precisely the kind of large scale structure that the human eye is good at picking out of random data, and because any NR software capable of removing the stripes will also be excellent at removing the contents of the picture itself.
