BIG NEWS - Hands on with CONTINUOUS raw recording on Canon 5D Mark III

[Sean Cunningham]

And yet, most cameras are now shooting 4K. Just wondering, why are you fighting progress so much man? If I didn't know any better I'd asume you would rather send the industry back to film.

 

Movement does not equal progress.

 

And here you have refuted none of my responses to what you've said.  If I didn't know any better I'd assume you didn't really understand what you were selling.

[Sean Cunningham]

I'm guessing you have bigger issues than "RAW" and "4K".

 

Yes.  I like things that are real.  I like to make intelligent choices.  I am a dangerous person.

[tomekk]

They posted one frame of DNG, dude, for both cameras.  Let the punters who wanted to declare the G6 a horribly, horribly soft camera compared to the GH2 based on compressed video uploads of no providence argue over detail and image clarity via streaming sources.

 

I was just indirectly asking what screen size + viewing distance you need to see the difference. Just curious, because I don't care about BMCC having more details if you can't see quite a bit of improvement in most standard watching conditions (tbh I don't care about it at all since I'm a run and gun amateur wannabe but that's a different story ;)).

[Sean Cunningham]

I was just indirectly asking what screen size + viewing distance you need to see the difference. Just curious, because I don't care about BMCC having more details if you can't see quite a bit of improvement in most standard watching conditions.

 

It was obvious at 100% magnification on my screen.  I blew it up for the grabs from the 32bit CS6-AE viewport for illustrative purposes on a web page such as this.

 

I even said that if you had no basis for comparison there's nothing lacking in the 5D version.  Hyperbolic claims of this thing being a BMD killer, of any model, provoked me into having a look...putting aside a host of deal-breaker caveats that may or may not be overcome by the ML folks that currently prevent this achievement from transcending its current reality as no threat to any cinema camera.

[thomse]

Our first 5DII RAW test. For some reason some only see it in 480p.. 

[Payne]

I didn't say most of the content is 4K, I said the "Hardware" is 4K. To put it simply, if the movie theater in your area has shown a 3-D feature, then most likely it has a 4K projector. They foresee 4K standardization soon.

Not to nitpick but this is untrue. Very very few screens are showing in true 4K at this moment in time. There are hundreds or maybe even thousands of the Sony Dual lens 3D systems in theaters across america that say they are 4K in the Sony branded pre-roll in the theater. These are really projectors that have 2 x 2K projection systems in one projector body. They are usually projecting from 2K (2D) DCPs or 2K 3D DCPs that are actually split between each lens/sensor and projected interleaved (which does not actually give you anywhere near 4K projection). Additionally, the current Digital Imax system does the same thing with two completely separate interleaved 2K Christie Cinema projectors projecting from a 4K DCP. This translates to a very, very slight increase in detail over a single 2K system (and is done to give each eye more light). This translates to very few true 4K DCPs being projected on true 4K projectors. Even in specialty situations where we have a movie like Prometheus that was captured at 5K native. I believe I read that Prometheus was posted as a full 2K DI. The Visual effect sequences were definitely rendered in 2K. So this means that even when they had specialty 4K screenings at select locations those were partially or completely 2K content on a 4K system.

 

Even in a true 4K situation, such as when I saw Samsara projected at the Arclight Hollywood in 4K it was stunning but not nearly as much of an improvement over 2K (in a high end theater setting mind you) as you would hope. And this is coming from a guy with 20/10 vision who has overseen dozens of premieres and special screenings (many of them at the arclight). 4k is amazing but it's not as amazing as it should be or as prevalent as you are led to believe.

 

The single most stunning resolution experience I have seen was the Dark Night in Imax at Universal City. Those prints were supposedly struck from the negatives. This means the viewer was seeing something between 8k and 16k on the screen during the imax sequences. THAT was what I always hoped true 4K would look like. Someday.

[HurtinMinorKey]

Our first 5DII RAW test. For some reason some only see it in 480p.. 

 

Looks better than most of what I have seen so far. Although, there might be some dropped frames.  1080 works on embedded, but not on YouTube site. 

 

I've always thought that Canon handles grass in bright sulight so nicely (for stills) it's good to see it finally come through in the videos.

[Evan Butka]

Anyone else having issues with the CompputerBay 128GB cards? I'm using a MacBook Pro and a USB 2.0 CF card reader. 1 out of 20 tries it mounts... and then as its transferring it ejects..

[Sean Cunningham]

...The when I saw Samsara projected at the Arclight Hollywood in 4K it was stunning but not nearly as much of an improvement over 2K (in a high end theater setting) as you would hope...

 

Arclight Hollywood, aka The Cinerama Dome, is a premiere theater amongst premiere theaters.  There isn't a theater chain in existence, in the US, that maintains this venue's quality.  Most theaters are filled with employees who don't care, no projectionists and managers that might as well be working at Burger King...hence why I only see movies at The Alamo Drafthouse since I no longer live in LA.

 

If it's "meh" at the Cinerama Dome...that's saying something.  Roll-out to the AMCs and Tinseltowns of the world was likely very premature.

 

 

edit: PS, that film, Samsara, is a compelling example of why film is still important for a variety of types of films today.  They took an amazing hit to their image quality deciding to go with the best DI offered.  A 4K DI is proxy resolution compared to their o-neg.  4K is a proxy to what's on the neg of anything "Filmed in Panavision" even.

[Payne]

Arclight Hollywood, aka The Cinerama Dome, is a premiere theater amongst premiere theaters.  There isn't a theater chain in existence, in the US, that maintains this venue's quality.  Most theaters are filled with employees who don't care, no projectionists and managers that might as well be working at Burger King...hence why I only see movies at The Alamo Drafthouse since I no longer live in LA.

 

If it's "meh" at the Cinerama Dome...that's saying something.  Roll-out to the AMCs and Tinseltowns of the world was likely very premature.

To be fair it was much better than "meh" but not "blowing my mind" like a true shot in imax, projected in imax (with no effects) type film. Also to be fair it was not in the cinerama dome but instead in one of the main complex theaters. This means that the optical quality will be much better than the dome (no distortion) but also not as big (allowing one to easily scrutinize extra detail).

 

I actually think 4K is more impressive in medium to small sized theaters because that is where you get the complete saturation of detail in a small place in the sweet spot of your field of vision. This where it starts to look like a window. A large theater screen makes it look more like something better than 2K but not life changing. Kind of like how the jump from a 2mp to 4mp camera is a welcome improvement but not a whole new strata of visual fidelity. (NOTE: I realize that going from 2k (approx 2mp) to 4K (approx 8mp) is much bigger than a 2mp to 4mp jump. It's that it "felt" like a 2mp to 4mp type jump in still photography.

[Sean Cunningham]

To be fair it was much better than "meh" but not "blowing my mind" like a true shot in imax, projected in imax (with no effects) type film. Also to be fair it was not in the cinerama dome but instead in one of the main complex theaters. This means that the optical quality will be much better than the dome (no distortion) but also not as big (allowing one to easily scrutinize extra detail).

 

But my point is still, main room or side rooms, that venue maintains every aspect of the experience.  If it wasn't mind blowing there it wasn't going to be anywhere.  I've had as many disappointing experiences at both the Fox in Westwood and the Chinese in Hollywood as I have good.  I've never had a bad experience at the Dome which was (and I'm betting still is) the go-to place for cinefiles.

[Payne]

@ BurnetRhoades: Agreed. I was honestly expecting more detail considering the film was shot in 65mm and got a true end to end 4K DI. It was beautiful and I made sure to pay Arclight prices to see it but I was honestly hoping for more resolution than I got. Again I am lucky enough to have good eyes to take it all in as well. Many people, unfortunately, do not.

 

Final thought on Samsara in 4K: As you know this type of film does not have fast panning but when there was any moderately fast dollying or panning that immediately cut any real world resolution benefit over 2K. 4K was really best displayed on the slow moving aerial wide shots and close ups of intricate clothing or something that does not move like the giant sand mandala (sic?).  

[sugartown]

We can make stuff happen. If that doesn't suit your way of working, then fine.

 

Huh? Look, there were people animating DSLR stills in the early DV days, and making music videos shooting literally 1 frame at a time like animation back in the 90's...Smashing Pumpkins and REM videos did this with it's own stuttering aesthetic ....so sure, there will always people who will break there backs and make amazing things, no matter the tools.  You don't have to work how professionals work, it's true...and in many cases you shouldn't...but...

 

It's a potentially (we're still finding out the details) back breaking, tedious process, outside the norm of practical production, no matter the budget, or crew. And with a $3500 camera when there are other options in that range. At least know that's what you're talking about.

 

Whether slating, requiring playback for music, shooting "go agains" in one take, camera moves, or just framing, and hitting marks... 40 seconds is intrussive. It would make post a nightmare during a period where workflow is already having some growing pains.

[Sean Cunningham]

Why do vendors keep picking these stupid, old file systems in the first place?  NTFS isn't new or exotic. 

[peederj]

I'm sure you know we are discussing two different lenses feeding two very different size sensors. As a result, the focal plane position and depth of field for the two shots will not be comparable directly. The BMCC will have a much greater DoF and so even if the focal plane is at the same point in the image (I don't see that we've been given confirmation of that, typically these unscientific tests are completely unreproducible because the testers are too lazy to tell us exactly what was done in them) as we move away from that plane in either direction the smaller sensor will retain smaller circles of confusion optically at the sensor. This optical advantage can easily make up for significant disadvantages of the sensor and later processing (visible in the BMCC's false color artifacts on the pebbles for instance, which you had to butcher to tame).

 

Which is why the first comparisons to be made should be test charts (properly focused of course, with lenses of similar scientifically validated performance even if not identical due to mount/sensor size differences). Why we can't get 100% reproducible test chart data first of all I have no explanation to provide you. We can still argue somewhat objectively with this anecdata but we have to be very careful in understanding everything we may be seeing before declaring something "proven."

 

 

This isn't "perceived detail", or false detail.  This is blades of grass:

 

...that's unaltered 5D on top, BMCC on the bottom, 1:1 to each other @ 200%

 

 

...if it were "false" detail then further enhancing wouldn't further resolve maintained blade structure, here still at 1:1 to each other, 5D on top and enhanced BMCC on the bottom (mild de-moire + my LCE scheme).

 

...and here we have 5D on top and BMCC on the bottom scaled down to 1080.  The comparison would be even worse for the 5D if you were to go the other direction. 

[Oscar M.]

I'm sure you know we are discussing two different lenses feeding two very different size sensors. As a result, the focal plane position and depth of field for the two shots will not be comparable directly. The BMCC will have a much greater DoF and so even if the focal plane is at the same point in the image (I don't see that we've been given confirmation of that, typically these unscientific tests are completely unreproducible because the testers are too lazy to tell us exactly what was done in them) as we move away from that plane in either direction the smaller sensor will retain smaller circles of confusion optically at the sensor. This optical advantage can easily make up for significant disadvantages of the sensor and later processing (visible in the BMCC's false color artifacts on the pebbles for instance, which you had to butcher to tame).

 

Which is why the first comparisons to be made should be test charts (properly focused of course, with lenses of similar scientifically validated performance even if not identical due to mount/sensor size differences). Why we can't get 100% reproducible test chart data first of all I have no explanation to provide you. We can still argue somewhat objectively with this anecdata but we have to be very careful in understanding everything we may be seeing before declaring something "proven."

 

Agreed.

 

Zacuto should include the 5D3 running ML in RAW on their next shoot out . . . that will be really interesting. Everything else is NOISE.

[blackrat]

THAT is actually a very good question !

Because all of this will be 90% useless if not long enough recording can be pulled on set for movies ... with actors waiting for the cam to cool down or something weird like that.

It doesn't appear to get hot. The liveview stream is put out there all the time anyway when in liveview or movie mode. All-i already sometimes writes out to the card jsut as fast as does holding the shutter down when stills mode.

 

OTOH it seems to stop every 4GB for now which means only 49 seconds continuous at 1920x1080. Once the hack is further along I imagine they will be able to have it immediately start up a new file, hopefully with no frames or maybe only 1 or 2 skipped (maybe this is already possible I've only fiddled with the hack a little bit so far). If you mess around with pressing info button or run it with full ML zebra/focus peaking and everything tryng to work at once sometimes it quits out. But basically it can keep recording to that for as long as it doesn't hit disk format file size limitations (4GB for FAT32) and I expect they will be able to seamlessly (or all but so) get around that as happens with regular video (and maybe there is already a setting that allows for that).

[blackrat]

Really depends what you are shooting.... For music videos and similar.... 40 seconds is fine.

Yeah or lots of natural world/wildlife/scenics. People always forget it's not all long format scripted stuff.

[blackrat]

I think the CF interface tops out at 150MB/s, could be wrong on that. SSDs would be nice for the capacity but I think 1000x or maybe 1100x is the max for the current CF spec.

 

I've tried the higher frame rates now.

 

960x540 at 60fps works fine, but some aliasing.

 

1920x540 is squashed looking.

 

1280 x 720 is actually 1280 x 672, and looks a bit squashed too.

 

But the card can handle it pretty reliably at 48fps.

 

Next on my list to try is anamorphic lens & 1920x1280 (3:2), that is going to be my preferred mode. From 3:2 with the Iscorama you will get roughly 2.35:1 :)

 

What setting gives the 1:1 zoomed mode? That looks great for wildlife and macro.

[blackrat]

Don't wanna burst bubbles around here, but while all of this is great... only 40 seconds of recording time is not gonna cut it for narrative. Hence, I don't think this will be the Black Magic killer everyone is saying it will be. If BlackMagic Pocket/Production can shoot for several minutes, WITH audio, TV and Cinema shooters like myself will need to stick with it.

 

Still, a nice little treat for people who only shoot demo videos or music videos, I guess. 

Once again the long form narrative folks always seem to think that nothing else exists. Not music video, not commercials, not short take pieces (and even very many long for narrative takes are 49 seconds or less, although certainly plenty are longer), not wildlife, not natural world/landscape/scenics. Long form narratives are just a part of a larger world.

 

I'm pretty sure there will file span soon anyway. Although for really serious stuff if it skips 1 or 3 frames every 49 seconds I supposed that may be one frame too many, but just one or two frames as repeats every 49 seconds might not really be terribly noticeable.