If the embedded video isn’t working, or if you only see the small thumbnail, it’s available at http://vimeo.com/2437826
If you want to cut to the chase, the test description is at 5:20, and the minute-long comparison footage starts at 6:24.
all the Canon 5D Mark II samples I’ve seen circulating the web seems to have crushed blacks and exagerated saturation (and they are all untouched by in post apparently).
Comparing the Canon 5D Mark II to the D90 (web-samples, don’t own any myself) makes me think anout EX3 vs VX200 where EX3 is so much better on paper but I still prefer the picture of the micro-pseudo-HD vx200
What do you guys think?
Posted by .(JavaScript must be enabled to view this email address) on 12/12 at 03:55 PM
I just got a 5D Mark II and I’m immediately noticing the crushed blacks that are so apparent in the Zacuto test (yet none of the testers bothered to mention it). What a disappointment. I see little if any subtle detail in dark shaded areas. It is all pushed to black. Yuck. So many of the initial video samples on the internet were done at night, where black areas were expected and therefore accepted without much comment. But now I see how misleading that was. This is a terrible disappointment to me. Crushed blacks would only be an asset when shooting a drab, low-contrast scene.
Note that ALL the other cameras tested had much better shadow detail.
Posted by .(JavaScript must be enabled to view this email address) on 12/12 at 04:09 PM
Indeed, I’m surprised both by the crushed blacks and the flattened, lifeless highlights in the 5D footage. Canons usually make much nicer pix than that.
If you read the comments at http://vimeo.com/2437826 they did fiddle with the 5D images a bit in post to even it out, but they would prefer to reshoot it with less contrast in the camera. It’s always the stuff you learn while reviewing the test that tells you what you should have done in the first place…
Sadly the specs page at http://zacuto.zenfolio.com/p82280286/h1429ae6c#h1429ae6c doesn’t indicate what image-processing settings were used for any of the cameras (gamma, knee, matrix, etc.). I see stuff in both the EX3 and HVX footage I think could be better tweaked; probably the HPX170 and HPX3000 as well. It looks to me like the RED ONE footage was processed in Rec.709, too; I know I can get better colors out of it than that.
Of course these sorts of tests always raise as many questions as they answer; it’s the nature of the beast. Still, it’s helpful to see a whole mess o’ cameras set up on a real scene. It’s a huge amount of work (especially with the 35mm and DaVinci color-correction session included), and I commend Zacuto for having made the effort.
Posted by Adam Wilt on 12/12 at 05:34 PM
I have seen a lot of 5D footage shot in a lot different lighting situations, and this was the worst looking by far. But the crushed blacks is common to everything I have seen. For me, the clean, almost noiseless low light performance outweighs that issue by a big margin.
On another topic, who the heck produced this video? Bad framing, worse editing, and that dolly cam got to be really annoying.
Posted by chucksav on 12/13 at 10:39 AM
yeah - 5D looked horrible for that - and I’m AMAZED that not once did they mention the fact that you’re stuck at 30p!
I’m glad they did these tests, but a bit irked at how “even handed” they claimed the results were - they make accessories, so probably didn’t want to badmount any piece of kit they support.
Also notice all locked off - if you pan around with 5D, hello rolling shutter! Not as bad as D90 (which is wretched for hand held), but still an issue.
There are definite qualitative differences between the results - it seems they are showing them pretty much in order of quality - so one to the next, not much difference, but compare film to D90/A1/5DMk2…yeeeks!
-mike, still a-ranting…
Posted by Mike Curtis on 12/13 at 04:40 PM
I want to know what all the various cameras were that shot the interview roundtable! Some of them looked horrible.
Color me (pun!) surprised about the Canon as well. Of course, there might be some nice detail down there in the blacks, if you CC’ed it well.
BAJ
Posted by .(JavaScript must be enabled to view this email address) on 12/13 at 08:28 PM
Roundtable was shot with HVX200s not 200As. 5 of them.
If you check out a video on my vimeo page… you can see a lesser compressed version of RED vs 5D2. http://vimeo.com/2356055
As the spunky/youthful producer of this shoot… I thought I’d comment on some of your remarks.
First of all, Mike, you’re right we didn’t mention much about rolling shutter and 30P. We wanted to get the ball rolling and get some people out there talking about these cameras. We could talk all day about the various workflow/camera details. And that’s why I’m glad all of that’s being mentioned on the blogs and in the forums.
Chuck, I produced it. Although I didn’t direct it. I’m sorry you didn’t like the dolly moving and what not. Not much I can do for you there. Take it or leave it! Thanks for watching though. I’m pretty proud of this work being only 18 and all.
Adam, I’d love to talk about camera settings. I want to get things as good as possible before my adapter shootout next month with 7 different adapters.
Keep the comment comming,
Joseph Stunzi
Posted by jstunzi on 12/13 at 11:37 PM
Oh and one more thing, sorry I spelling coming wrong!
Posted by jstunzi on 12/13 at 11:38 PM
Mike: “probably didn’t want to badmount any piece of kit they support”—if you made camera support gear, you wouldn’t want to badmount it either, because folks would instead go to a company that made goodmounts, grin.
Joseph: Good on ya; getting just one shot in the can with so many different cameras is a Big Deal. As Mike can tell you, doing even a sketchy shootout with multiple cameras takes a lot of effort (Mike and I and Chris Hurd spent four days in ‘06 doing just that, plus prep and post time: http://www.adamwilt.com/HD/4cams-part2.html).
That the motion of the musicians didn’t reveal any obvious rolling-shutter jellocam indicates that it just wasn’t a factor for this particular shot. Indeed, I find more often than not that jellocam simply isn’t much of an issue. Yes, another setup might show it off more, but in between static talking heads and dynamic action shots there’s a whole universe of scenes, and for many of ‘em even the D90 is Just Fine.
As to camera settings, on the HVX200, 200A, and HPX170, I really like Cine-like V gamma for the way it rolls off highlights on skin tones (see http://www.dv.com/columns/columns_item.php?articleId=193001363); on the EX1/EX3 I opt for Cine Gamma 4 and either turn detail off or dial it down considerably.
Of course, your mileage may vary…
Posted by Adam Wilt on 12/14 at 12:24 AM
i say
GOOD ONYA MATE for doing a comparision…
a couple of notes
when you shoot something, have more than red and black and brown in the frame…
put in
green, blue, yellow, etc
hell, throw a mexican blanket in the corner
and
we have no idea the contrast range on the lighting…
6 f-stops?
8?
10 f-stops?
i like the outside on a sunny day for lighting contrast, since we all know what THAT contrast range is…
(although we all know its not the easiest to make shifting daylight consistent between cameras..)
also, from lookin at the Canon manual, you can adjust the contrast, saturation, etc and save as a ” picture style”—
and when the movie mode is activated, it uses that “style”—
can lowering the contrast bring the crushed blacks into range?
i know the still photo settings (RAW) have around a 10 f-stop range ( people say)
so the chip can do it,
we just need to ( hopefully ) set it correctly BEFORE the compression hits it
GOOD run at the test
bill s
Posted by billS on 12/14 at 03:27 AM
Re: my new 5D Mark II
In order to bring out more shadow detail (to uncrush the blacks), I tried all the different picture style settings, but I only noticed very small changes in the shadows. I was able to get a slight increase in shadow detail if I created a custom setting with lowered contrast, but this of course lowered the contrast of the whole scene, and the gain in shadow detail was not that great.
Posted by .(JavaScript must be enabled to view this email address) on 12/14 at 09:20 AM
We played with contrast and some of the other picture settings afterwards and to my eye… I didn’t see a huge difference. But we didn’t try loading images into the Canon software and creating a custom style… does anyone know if that allows more adjustment than doing adjustments in camera?
Posted by jstunzi on 12/14 at 10:45 AM
RE: Crushed blacks and whites
There is a great discussion at dvinfo.net about how to recover shadow detail using Apple Color or a custom Picture Style. Be sure to read both pages of posts:
http://www.dvinfo.net/conf/canon-eos-5d-mk-ii-hd/139602-my-fcp-workflow.html
Posted by .(JavaScript must be enabled to view this email address) on 12/17 at 07:45 PM
Shoot out? Are you kidding me?
This little comparison video is a zero sum gain, sure they have a nice set of cameras but they actually say nothing about any of them, and I quote:
“Any one of these cameras in skilled hands I think can do amazing work.”
This says nothing, null, statements like these should have been removed from the edit, we will not get started on the edit.
Moving on, they may have tried to get the same framing in their conform but if you don’t have the same DOF why even bother. Also Vimeo is a good and fine way to watch a cat on a screen but is no platform to compare video quality.
In conclusion:
If you do have a shoot out and end up with no opinion about the best camera of the bunch you should really give up and do everything the director says.
P.S.
shooting a documentary in the Amazon with a camera that does not record audio? Good luck.
Posted by .(JavaScript must be enabled to view this email address) on 12/25 at 03:17 AM
_thumb.jpg)
