Canon 5D MKII Discussion--New Thread.

[dwdallam]

This is a thread for the discussion of the Canon 5D MKII.

Please follow these guidelines:

*Keep the discussion focused on the performance and usability of the 5D MKII

*Personal related discussions or "digressive" technical or speculative discussions should be relegated to PMs or another thread, so as not to clog and derail the discussion of the 5D MKII.

This is a good link, which investigates the video mode of the 5D2.
http://blog.vincentlaforet.com/2008/09/20/...o-our-industry/

The images are screen grabs from the video mode, which look like still images. Very amazing, and shot at 3200 ISO.

[Ray]

The problem is, DWDALLAM, there are no RAW image comparisons of same scenes shot under the same conditions with any other camera.

All we've got at present are mostly jpeg shots. We can only speculate and make educated guesses about ultimate performance compared with rival cameras.

I've taken shots at the equivalent of ISO 12,800 with my 5D which look usable. The results depend very much on the eveness of lighting and the darkness of shadows. Where there are dark shadows or black surfaces, noise and banding will be very apparent at such a high ISO. If there are no deep shadows and dynamic range (or subject brightness range) is low, results can be impressive.

The only true test is same exposure shots at appropriate ISOs for correct (ETTR) exposure, compared at equal physical image sizes.

We simply don't have any such comparisons yet. It's all a bit premature.

[dwdallam]

The problem is, DWDALLAM, there are no RAW image comparisons of same scenes shot under the same conditions with any other camera.

All we've got at present are mostly jpeg shots. We can only speculate and make educated guesses about ultimate performance compared with rival cameras.

I've taken shots at the equivalent of ISO 12,800 with my 5D which look usable. The results depend very much on the eveness of lighting and the darkness of shadows. Where there are dark shadows or black surfaces, noise and banding will be very apparent at such a high ISO. If there are no deep shadows and dynamic range (or subject brightness range) is low, results can be impressive.

The only true test is same exposure shots at appropriate ISOs for correct (ETTR) exposure, compared at equal physical image sizes.

We simply don't have any such comparisons yet. It's all a bit premature.
[{POST_SNAPBACK}][/a]

All true, but talking pictures in such low light situations at 1/125th was previously impossible with a Canon camera, nevertheless. And--those images are video grabs. I think it's amazing.

But you are right. We'll need to see how it performs side by side. From what I've read--those who have actually tested the jpg function--I think 6400 ISO is going to be very, very usable, even in low light conditions that you present.

Another thing to remember, as you and other know, when you have noise in an image visible at 70-100% on screen, even w/o noise reduction you won't see it in an 8x12 print, or even a 12 x 18 (although I haven't confirmed that--and I'm talking noise at ISO 1600-3200 with the 1DS3 because at an 8x12 the resolution is so compacted as to render noise unseen.

But that's all elementary because all things being equal, the same can be said for the 5D2, since its resolution is the same. So at 6400 ISO it's noise will be more usable than the 1DS3's. All this is of course given that the jpg's are not better than the RAW files.

I'm with the blogger on th4e low light capabilities however. The lower the light I can shoot in the more creative headroom I have. One of my best shots was shot under a San Franciso street light at 2AM with no other light. You can see it here:

(Server where linked image resides is down for maintenance right now but should be up soon)

[a href=\"http://www.fredmiranda.com/forum/topic/686418/chat.php]http://www.fredmiranda.com/forum/topic/686418/chat.php[/url]

And that's why I'm so interested in ISO and noise. It will allow us to shoot in light conditions that were previously either very hard, required secondary lighting, or were simply impossible. I mean with WB control, we can shoot in any color of light, provided it's bright enough.

[feppe]

I'm unclear how the "extended" ISO ranges in 5Dii (and others) actually work. Is, say, 6400, actually shot at 3200 and pushed to 6400 in-camera? And 12800 is actually still 3200?

If so, is there any other magic going on other than what could be done in Lightroom or PS with cameras which don't have extended ISO ranges?

[dwdallam]

I'm unclear how the "extended" ISO ranges in 5Dii (and others) actually work. Is, say, 6400, actually shot at 3200 and pushed to 6400 in-camera? And 12800 is actually still 3200?

If so, is there any other magic going on other than what could be done in Lightroom or PS with cameras which don't have extended ISO ranges?
[a href=\"index.php?act=findpost&pid=223184\"][{POST_SNAPBACK}][/a]

I don't know the technology behind it, but you can't duplicate the high ISO capabilities in LR or PS at all. You simply won't get an acceptable image. If you do a search on the Nikon D3 and look for some of it's examples shot at ISO 6400 and 12800, you'll get the idea.

[madmanchan]

6400 is a "normal" ISO setting on the 5D II. The two extended ranges (which are effectively pushed) are 12800 and 25600. In this sense, it is similar to the D3/D700.

[Panopeeper]

you can't duplicate the high ISO capabilities in LR or PS at all

Just the opposite is true: you are better off with underexposure and adjustment in LR/ACR than with the fake ISO (12800 and 25600 with the 5DII). The camera does the same adjustment (there is no ISO gain associated with the fake ISO), but you lose one respectively two full stops from the dynamic range. If you shoot with 6400 and increase the intensity in ACR, you can save 1-2 stops of highlights.

One could say that highlights are irrelevant in low-light situation, but that is not so. When using ISO 3200, the DR is (probably) two stop less than with ISO 200. The loss is betwen 3.5 and 4 stops with ISO 12800. Clipping of highlights is real.

[Ray]

It's interesting how people can get seduced by the ISO number. When the D3 came out, I don't believe I ever saw a single comparison on the internet between the D3 at ISO 6400 and any other camera which didn't have an ISO 6400 setting.

In fact, the 5D underexposed by just 2/3rds of a stop at its setting of ISO 3200, or underexposed 1&2/3rds stops at ISO 1600, represents ISO 6400, after push processing in the RAW converter. (ISO 3200 on the 5D being actually ISO 4000).

[Panopeeper]

Ray,

once you passed me a 5D raw file #9199, with ISO 3200, still much underexposed (at the front of Ratana's Kitchen).

That shot is a good example for why not to use a fake ISO: some highlights, for example the shaded lamps have been clipped. The noise would be the same with ISO 1600, one stop underexposed (I mean one more stop underexposed), but the clipping *may* have been avoided.

Of course, this clipping is nothing earth-moving, but it demonstrates the principle: don't give away the DR for nothing in return.

[mike.online]

The Laforet video just dropped (see first post). I've got to say, I'm impressed.

From what can bee seen in that video, and the sampled that Micheal has posted, this is certainly a capable camera.

EDIT: Direct Link. http://www.usa.canon.com/dlc/controller?ac...&articleID=2086

[Ray]

Ray,

once you passed me a 5D raw file #9199, with ISO 3200, still much underexposed (at the front of Ratana's Kitchen).

That shot is a good example for why not to use a fake ISO: some highlights, for example the shaded lamps have been clipped. The noise would be the same with ISO 1600, one stop underexposed (I mean one more stop underexposed), but the clipping *may* have been avoided.

Of course, this clipping is nothing earth-moving, but it demonstrates the principle: don't give away the DR for nothing in return.
[a href=\"index.php?act=findpost&pid=223484\"][{POST_SNAPBACK}][/a]

Quite right! I have been guilty myself of using ISO 3200 instead of the same exposure at ISO 1600. But it's sometimes useful to be able to see a review of the shot taken on the camera's LCD screen, which is not excessively dark. This is all part of the seduction of the artificially high ISO setting.

None of the sample 5D2 shots at ISO 6400 and above look particularly impressive to me. If I can get a good result at ISO 3200 with the 5D2 in circumstances where ISO 1600 with the 5D is not quite adequate, I'll be happy   .

[dkeyes]

The Laforet video just dropped (see first post). I've got to say, I'm impressed.

From what can bee seen in that video, and the sampled that Micheal has posted, this is certainly a capable camera.

EDIT: Direct Link. http://www.usa.canon.com/dlc/controller?ac...&articleID=2086
[a href=\"index.php?act=findpost&pid=223491\"][{POST_SNAPBACK}][/a]

Checked out the video as well, hard to believe it came from a dslr camera. Unless Nikon comes out with a top of the line version with the same specs, this will be in a league by itself.

[dwdallam]

Just the opposite is true: you are better off with underexposure and adjustment in LR/ACR than with the fake ISO (12800 and 25600 with the 5DII). The camera does the same adjustment (there is no ISO gain associated with the fake ISO), but you lose one respectively two full stops from the dynamic range. If you shoot with 6400 and increase the intensity in ACR, you can save 1-2 stops of highlights.

One could say that highlights are irrelevant in low-light situation, but that is not so. When using ISO 3200, the DR is (probably) two stop less than with ISO 200. The loss is betwen 3.5 and 4 stops with ISO 12800. Clipping of highlights is real.
[a href=\"index.php?act=findpost&pid=223467\"][{POST_SNAPBACK}][/a]

Yes, I meant "normal" ISO settings, never extended, which is an important aspect for sure. For the 1DS3 that's 1600.

[Panopeeper]

Yes, I meant "normal" ISO settings, never extended, which is an important aspect for sure. For the 1DS3 that's 1600.

The 1DMkIII and 1DsMkIII are rather the exception: ISO 3200 is real on the 1DsNkIII, and ISO 6400 is real on the 1DMkIII. The only fake ISO is 50.

The talk was about the 5D; that one does not have real 3200, and the 5DII will top at 6400, I guess.

[wilburdl]

I myself am interested in what it looks like at 100-400 iso as those are what I normally use. Are there any studio shots w/model available yet?

[dwdallam]

I myself am interested in what it looks like at 100-400 iso as those are what I normally use. Are there any studio shots w/model available yet?
[a href=\"index.php?act=findpost&pid=223844\"][{POST_SNAPBACK}][/a]

I really wouldn't be worried about that.

[dwdallam]

The 1DMkIII and 1DsMkIII are rather the exception: ISO 3200 is real on the 1DsNkIII, and ISO 6400 is real on the 1DMkIII. The only fake ISO is 50.

The talk was about the 5D; that one does not have real 3200, and the 5DII will top at 6400, I guess.
[a href=\"index.php?act=findpost&pid=223817\"][{POST_SNAPBACK}][/a]

I didn't know the 1DS3 had a normal ISO of 3200. I just never go over 1600 because I read afew places with examples that the noise was, well, not good let's say. However, on Micheal's review of teh D3 and 1DS3 he says: "Why ISO 1600? Because that's the Canon's [1DS3--edited] highest setting without going into a boost mode. The Nikon doesn't go into a boost mode until above ISO 6400."

But now that I think about it, that's right. Unless you "expand" the ISO setting in custom functions, it's 1600 for the 1DS3.

[jjj]

There have been some full sized samples posted here for a while now.
http://www.dpreview.com/previews/canoneos5dmarkII/page14.asp
I downloaded a couple and distinctly underwhelmed. Though hard to tell if it was user error, crappy lens or the camera that was the cause of the poor quality.

[dwdallam]

There have been some full sized samples posted here for a while now.
http://www.dpreview.com/previews/canoneos5dmarkII/page14.asp
I downloaded a couple and distinctly underwhelmed. Though hard to tell if it was user error, crappy lens or the camera that was the cause of the poor quality.
[a href=\"index.php?act=findpost&pid=224254\"][{POST_SNAPBACK}][/a]

Those are Beta model JPGs and they really show nothing about performance.

[daws]

Call me stupid, but I don't get it.

Art school guy; pro video producing/directing/editing for 25 years; amateur enthusiast landscape photography for 3 years. My video kit and my photography kit don't share anything -- not lighting, not DOF, not post suite, not tripods, not shooting technique, not storytelling, not even my brain.

I'm sitting here with my 5DI, looking at the spec sheet for the 5DII, and asking... why?

There's a pile of stuff in the 5DII that I'd love to have. Video isn't one of them.

What is it that I'm not getting?