F3 vs. AF100 vs. 5D

[Barry_Green]

Judging based off screen shots, extracted from a 720p vimeo, of the doc, isn't going to give you the true story. Best to see it from the SCCE blu-ray whenever it becomes available. That said, I can testify that at the screening, on a big-screen, from a 4K projector, the differences between the AF100 and F3 were fairly minimal, and the gulf between both cams and the DSLRs was pretty significant (enough to cause the audience members to laugh out loud when the DSLR shots came up). As such, bumkicho's screenshots here are reasonably accurate representations of the differences.

 

[bumkicho]

I shoot with D7000 and GH2 and they serve me good for what I do and I also like many online videos shot with 5D, but when it comes to resolution, there is no comparison. Just look at the mirror frame and the details on white shirt the girl in front of mirror is holding. F3 and AF100 resolve details far better than 5D and D7000.

And yes I would love to see the entire 2011 shootout on Bluray!!

 

[Everts]

Judging based off screen shots, extracted from a 720p vimeo, of the doc, isn't going to give you the true story

actually the file that is streaming is a 1080p file you can save it from the IE cache.
Didnt find an original download though.

 

[G.P.]

eh I give up haha.

 

[Duke M.]

You deleted the rant just as I was about to respond. LOL

GP I think you'll always be dissatisfied with any test like this because you're geared to trying to produce beautiful images. These test shots are about making ugly images with stress tests.

I also know you're worried that producers will cite this out of context to justify what they want to justify. However, the F3 does very well overall and with S-Log will do even better so relax. If you're worried they'll justify a DSLR on a big shoot, just show them Postmaster's comparison of the dresses. End of discussion.

BTW I did find the under and over exposure test the most helpful too. Of course that's only so far. This was only the first installment.

 

[G.P.]

You deleted the rant just as I was about to respond. LOL

I've decided to enforce self censorship haha. I feel like I'm beating a dead horse with what I've been trying to explain to people, so I deleted everything. I'm just gonna make some movies.

 

[J Michael]

It was not. That's what was asked at the Q&A of the SCCE, people insisted that the DSLRs must not have been properly focused. And you have to consider -- they had 12 cameras there, and every single one of them used the exact same lens, and consistently they got focus exactly correct on nine of the cameras, but somehow managed to mess up focus every single time on the DSLRs? With dedicated camera ops for every camera? Nope. They were extremely meticulous. They didn't miss focus.

Weill I said it looks OOF. And I seem to get non-soft results with a wide on my 7D so perhaps I'm missing the point. For example see the detail in the gentleman's hat to the right of the Koran burning gentleman in the center. What I do see in analyzing my footage is I have to be very careful about the frame I select due to the motion. This is with a Tokina 16-50 zoom down around the 16mm end.

 

[Bassman2003]

I think the point of these comparisions is to show the relativity of detail and sharpness. The DSLR will look sharp against some cameras and soft next to others. It really is a sliding scale. Motion did not play a part in these tests as the cameras were stationary and a lot of the scene was not not moving. The fact is the DSLRs took it on the chin in this test for detail compared to dedicated video cameras. They still can make nice images, but not as nice at 1080p to the other cameras.

The whole key here is that the footage was viewed at real resolution. No internet small frames which hide all imperfections.

 

[Duke M.]

Weill I said it looks OOF. And I seem to get non-soft results with a wide on my 7D so perhaps I'm missing the point. For example see the detail in the gentleman's hat

The hat is much closer so the pattern is bigger. That's why when you use a resolution chart it has to be frame properly.

Look at this close up of the hat. The outer big pattern is clear (low resolution), but the middle of the top of the hat has turned to mud because the pattern got smaller and it couldn't be resolved. You may have a higher resolution image, but that only affects the frequencies at which the artifact occurs.



Don't get me wrong. I do have a DSLR, but its not the first choice for videos.

 

[J Michael]

Don't get me wrong. I do have a DSLR, but its not the first choice for videos.

Ha ha no kidding. Shooting this type of stuff with a DSLR is a PITA and I agree there are a lot of tradeoffs image quality-wise. Here's the actual crop of the hat from the original frame -

 

[Duke M.]

Right. Still loses detail in the crown. Then move the hat 15' farther from the camera and you'd have the situation with the dress.

 

[Danespina]

I still prefer the 5D over the AF100, looks more cinematic to me

 

[roei z]

i never understood what looks so 'cinematic' on the 5D...
unless the waxy look and pink faces are the definitions.

 

[Danespina]

i never understood what looks so 'cinematic' on the 5D...
unless the waxy look and pink faces are the definitions.

Wish I knew what the definition of the cinema look was. Alot of my opinion just comes through heart and gut reaction rather then scientific analysis. I found the 5D gave things a David Fincher sort of feature film look, where with the AF100 I kept getting vibes of "The Office" or a corporate video. This especially became true once the day got darker. If the AF100 wasn’t staring down a city scape or neon lit signs, to me the shots would start feeling like a made-for-tv movie where as the 5D would retain its “Cinema”.

 

[roei z]

Wish I knew what the definition of the cinema look was. Alot of my opinion just comes through heart and gut reaction rather then scientific analysis. I found the 5D gave things a David Fincher sort of feature film look, where with the AF100 I kept getting vibes of "The Office" or a corporate video. This especially became true once the day got darker. If the AF100 wasn’t staring down a city scape or neon lit signs, to me the shots would start feeling like a made-for-tv movie where as the 5D would retain its “Cinema”.

ok, that sounds more interesting. sorry to kinda hijack the thread.
i myself can't see the 'fincher' look on all 5D footage i saw but i guess it's a personal thing.
however i totally agree with you on most of the AF100 footage. something looks videoish.
THAT's exactly why i find my GH2 so good. i personally like my footage better than all AF100 stuff i saw.
i shot a feature with it, getting nice flat results with two FD's.
if that's the budget limit before choosing to go with F3 or RED,
i'de still pass any Canons, and even pick the GH2 over AF100 (of course the pro-level hardware is a winner but that's not the issue).

 

[Danespina]

I’ve never shot with the GH2, but I’ve seen some footage and would probably prefer that look over the AF100 as well. Ultimately I think I’m going to settle on the FS100 for myself.

 

[roei z]

a nice settlement

 

[Robert Ruffo]

This picture is the answer for DSLR shooters who, when you show them a resolution chart and their favorite DSLR isn't resolving much more than standard definition says "I don't shoot resolution charts, I only shoot real people."

Point to this picture and say "That's why they invented resolution charts, so you'll know how real people will look."

+ 100

 

[tom.wong]

I still prefer the 5D over the AF100, looks more cinematic to me

always interesting to see what people think is the "cinema" look. considering s35 3 perf has no where close to the same field of view or depth of field as full frame. my idea of cinematic is high dynamic range, where the highlights just roll off like butter, and skin tone rendition being natural... skin tones can look nice on a 5d, which is why i think you can get that initial wow factor on the aesthetic. everything else kind of screams digital to me though...

 

[dcloud]

always interesting to see what people think is the "cinema" look. considering s35 3 perf has no where close to the same field of view or depth of field as full frame. my idea of cinematic is high dynamic range, where the highlights just roll off like butter, and skin tone rendition being natural... skin tones can look nice on a 5d, which is why i think you can get that initial wow factor on the aesthetic. everything else kind of screams digital to me though...

same sentiments. its all in the dynamic range for me. color can be fixed in post.