Rodney Charters, ASC - RED ONE, Canon 7D and 5D MkII side by side

[Tom Lowe]

I downloaded the 1080p clip: http://vimeo.com/7559839

Just my opinion, but here's what I saw.

1. The long-lens shot of LA on the 7D at :40 was disastrous.

2. The R1 footage starting around 1:20 looks incredibly soft. Was this missed focus? Shooting at 2K?

3. 5D2 aliasing at 2:15 = very bad.

4. 2:54... missed focus again on the Red?

5. The 5D2/R1 wipe at 3:04 = total destruction.

 

[M Most]

Just my opinion, but here's what I saw.

1. The long-lens shot of LA on the 7D at :40 was disastrous.

Describe "disastrous." Quite frankly, I don't see anything wrong with it. If you don't like the color correction, that's one thing. But I don't really see anything technically wrong with the shot.

2. The R1 footage starting around 1:20 looks incredibly soft. Was this missed focus? Shooting at 2K?

Focus is on the FG, and yes, I believe it was shot at 2K (Neil?). Looking at the color corrected version (at about 1:34) it doesn't really look that bad to me.

3. 5D2 aliasing at 2:15 = very bad.

I've seen this piece both on a 50 inch plasma and on a 2K projector, and I don't recall seeing what you're seeing here (I assume it's primarily the stairstep edge aliasing on the hair dryer helmets). This could easily have happened when recompressing for Vimeo and not in the camera. I can't guarantee that, but to my recollection those artifacts were not nearly as obvious in the original.

4. 2:54... missed focus again on the Red?

Look at the sparkle in his eyes. As far as I can tell, this is not that far off. Once again, recompression issues are possible here..

5. The 5D2/R1 wipe at 3:04 = total destruction.

Well, the Red version is considerably sharper. It should be for 10x the price and having technology developed for the purpose.

I think you're looking at this in an unintended way. The purpose of the piece is not to compare these devices 1:1. It's to show that the Canons can generate images that are surprisingly good for the cost, and that under the right circumstances can be intercut with images from a much more capable camera without undue embarrassment, thus giving indie filmmakers another tool for circumstances that warrant it. Personally, I think the direct A/B comparison on the last shot you mention is something I might have left out as the reaction might be the one you had, thus diluting the purpose of the exercise.

I also would say that since some of what you're bringing up is possibly attributable to things other than the camera, it's another very graphic example of why I mentioned previously in this thread the need to see things in proper venues rather than via Internet clips if one is to judge them. Unfortunately, far too few people are willing to do that these days.

 

[Tom Lowe]

All I know, Mr. Most, is that I viewed the 700MB+ 1080p video and reported exactly what my eyes saw. Take it for what you will. It's only my opinion.

The 5D2/R1 wipe at 3:04 looked like what happens when I put my glasses or contacts on.

 

[Sanjin Jukic]

All I know, Mr. Most, is that I viewed the 700MB+ 1080p video and reported exactly what my eyes saw. Take it for what you will. It's only my opinion.

The 5D2/R1 wipe at 3:04 looked like what happens when I put my glasses or contacts on.

Tom,

I can agree with you.

For my point of view all types of evaluation, tests, comparisons, etc,...are sort of competition.

In this case RED1 is a winner, the second to come is Canon 7D and 5DMK2 is the third.

As you're said the footage time code between 3:00 and 3:04 told us everything.

Here we go:

 

[michael zaletel]

IMHO Color differences in comparison footage too significant to evaluate other factors. But the T-Shirt texture in the above tells it all!

Oh, and I have both cameras and have put same lenses on both the RED and 5D2.

-michael zaletel

 

[Tom Lowe]

GAME. OVER.

 

[Brad Webb]

Tom,

Can you post a still and/ or explain what you mean about this:
1. The long-lens shot of LA on the 7D at :40 was disastrous.

I respect your opinion and I'd like to know exactly what it is you see and dislike in this shot. I'd post the still myself but I can't seem to find where the download link is on vimeo.

No doubt the Red is far superior to both of the Canons. Although I think that 5D wipe shot looks a little soft focus wise.

 

[Patrick Tresch]

There are definitely people out there shooting with the Canon cameras as their A camera.

Like Jean-Luc Godard.

Pat

 

[Tom Lowe]

Brad, it could be out of focus a bit, but I cannot really see how on earth that would even be possible given the 10x zoom focus assist on the 5D2! Anything is possible, though. It would a huge, huge error on the part of the 5D2 operator.

The long-lens shot at around :38 is just terrible at 1080p. It's jittery, it's got massive jello shake, the resolution looks worse than SD... it's just BAD.

 

[Andrew Walker]

Wow, after seeing the 1080p stuff the difference between the RED One and the Canon cameras has been made clear to everyone...or at least it should be. Again the 5D2 and the 7D aren't A cameras. Sure if your video or whatever is going out to the web only then the Canon camera can be your A team. But for anything outside of the realm of the internet the difference between RED and Canon is huge.

 

[Tom Lowe]

Owned.

 

[M Most]

GAME. OVER.

I give up. I should have known that trying to reasonably discuss the point of the piece and prevent it from turning into what it has now obviously become - namely, another rather pointless A/B comparison between one device that costs less than $2K and another that costs more than 10 times that amount - was a bit of a stretch. I'm sorry for trying.

 

[Tom Lowe]

So it's useless to compare whether Canon footage can be cut against Red One footage at 1080p?

 

[Sanjin Jukic]

So it's useless to compare whether Canon footage can be cut against Red One footage at 1080p?

Let's make a conclusion:

- Canon 5DMK2 full FF35 video footage is a pure illusion and it would never come.

- Canon 7D 1080p H264 footage can be useful at its best conditions but it is still not 100% possible to control.

- Canon 7D and 5D,... are both still have a problem to nail the focus but this can be fixed later.

- Obviously that R1 is still inferior at high ISO but we still have to see M-X performance @ high ISO.

- Also any other digital cine camera can't compete with the latest Video DSLR's high ISO performance.

 

[M Most]

So it's useless to compare whether Canon footage can be cut against Red One footage at 1080p?

No, not under sensible conditions. It is, however, rather useless to make A/B comparisons and then conclude what's obvious before the comparison is even made, namely that a camera that costs $20K+ that records a raw record of a 4K sensor is going to be significantly sharper on fine detail than a camera that costs less than $2K and records a binned 1920x1080 record from a sensor and then compresses it using H.264. The usefulness of the intercut sequence is that when both are brought to 1920x1080 size, shots from different angles can be intercut without people jumping out of their seats, and more importantly, shots that can be made under certain production conditions - like available light at night - with the cheaper device (but not the more expensive one) can be used in the same piece without embarrassment. One is clearly "better" than the other (as it should be), but that doesn't mean that the "lesser" one is unacceptable or has no usefulness.

 

[Tom Lowe]

No, not under sensible conditions. It is, however, rather useless to make A/B comparisons and then conclude what's obvious before the comparison is even made, namely that a camera that costs $20K+ that records a raw record of a 4K sensor is going to be significantly sharper on fine detail than a camera that costs less than $2K and records a binned 1920x1080 record from a sensor and then compresses it using H.264. The usefulness of the intercut sequence is that when both are brought to 1920x1080 size, shots from different angles can be intercut without people jumping out of their seats, and more importantly, shots that can be made under certain production conditions - like available light at night - with the cheaper device (but not the more expensive one) can be used in the same piece without embarrassment. One is clearly "better" than the other (as it should be), but that doesn't mean that the "lesser" one is unacceptable or has no usefulness.

I thought everybody already knew that, though, Michael? What I took from this video is that toe-to-toe, under anything approaching normal shooting conditions, people WILL jump out of their seats at 1080p during R1/5D2 cuts. The wipes prove that.

That is a lesson I learned the first week the 5D2 came out, when I tried to cut the 1080p video with 5.6K RAW timelapse.

 

[Kwan Khan]

Thanks for sharing...

 

[M Most]

I thought everybody already knew that, though, Michael? What I took from this video is that toe-to-toe, under anything approaching normal shooting conditions, people WILL jump out of their seats at 1080p during R1/5D2 cuts. The wipes prove that.

That is a lesson I learned the first week the 5D2 came out, when I tried to cut the 1080p video with 5.6K RAW timelapse.

The wipe was essentially the same angle A/B'd across both cameras. As I said earlier, I probably wouldn't have included that. Having seen this piece (and others) under proper, non-super-compressed-Internet conditions, I would say that for the most part, intercuts between the cameras were OK as long as they were either different angles or different scenes. The difference in sharpness is considerably less apparent than I would have thought, provided there is movement in the shot, and one isn't looking at the same shot on both cameras. This is what is allowing a show like 24 (or Southland, for that matter, who used quite a bit of Canon footage) to get away with using it.

 

[Tom Lowe]

The wipe was essentially the same angle A/B'd across both cameras. As I said earlier, I probably wouldn't have included that.

Sheesh. Maybe I am a little fanatical about image quality or something, but I would never get within a thousand miles of intercutting those shots.

 

[Ace]

The difference in sharpness is considerably less apparent than I would have thought, provided there is movement in the shot, and one isn't looking at the same shot on both cameras. This is what is allowing a show like 24 (or Southland, for that matter, who used quite a bit of Canon footage) to get away with using it.

Theres a million dollars of advice right there. Thanks Mike for that nugget! This (high motion) usage is clearly where they can excel in being used in concert with film/red/genesis.