Three three-letter cameras: EX1, F23, RED

[Tom Lowe]

Ah, I see. Well, I guess if you take into account all the stuff that can go wrong in the field shooting, and that we won't always be operating with the very best lenses or under optimal lighting conditions, etc, it's still amazing to know that we have a camera that can put out a stunning 2K image with 35mm cine lenses and DOF that can compete with - and in many ways beat - systems costing ten times as much as Red.

I'm still one of the people hoping for 2K scaled overcrankage, but I haven't heard much on that front lately. Have you guys ruled that out, or put it on the backburner basically?

 

[Joel Kaye]

I'm still one of the people hoping for 2K scaled overcrankage, but I haven't heard much on that front lately. Have you guys ruled that out, or put it on the backburner basically?

Well scaled 1080P is ruled out. I think that it's fair to say scaled 2K went with it. (don't shoot the messenger, I've been shot already)

This thread covers it - and here's the post that nails it.
http://www.reduser.net/forum/showpost.php?p=177161&postcount=52

 

[David Mullen ASC]

Resolution-wise, regardless of the numbers, all that matters is that RED is competitive with 35mm resolution at much lower cost.

If some of you want to say it exceeds 35mm resolution, that's fine, but the truth is that it all depends on the quality of the 35mm photography in question. I find it hard for anything (short of 65mm) to quite match good contact-printed 35mm anamorphic photography shot on good lenses in decent light levels... but in terms of matching Super-35 and standard 35mm, certainly it seems to me that digital has approached or matched that benchmark in practical terms. Certainly it's true nowadays with so much 35mm going through a 2K D.I. anyways. If 2K (unfortunately, IMHO) is becoming the post standard for theatrical features then certainly the RED at a usable 3.4K or 2.5K, whatever, is good enough, resolution-wise.

I think dynamic range is really the next big area to tackle. We don't need 6K or 8K digital cameras (except maybe for IMAX screens), what we need is two to four more stops of dynamic range at our disposal (and believe me, half the time we will be disposing of them...)

 

[Tom Lowe]

I think dynamic range is really the next big area to tackle. We don't need 6K or 8K digital cameras (except maybe for IMAX screens), what we need is two to four more stops of dynamic range at our disposal (and believe me, half the time we will be disposing of them...)

Yep. This is where things stand now.

I do think people should be careful, though, not to bottleneck major features at 2K (if it can be avoided), since 4K is very likely to be the new 1080p, five or ten years from now. Video games are already headed for 4K, and the LCDs are on their way soon enough.

In terms of cameras, the real holy grail, for me, would be a 65mm sensor with 6K resolution downsampled to a real and super clean 4K, with 13 stops.

 

[David Mullen ASC]

In terms of cameras, the real holy grail, for me, would be a 65mm sensor with 6K resolution downsampled to a real and super clean 4K, with 13 stops.

Yes, that would be really cool, though you'd need a great focus puller and the best lenses possible to take advantage of that resolution and deal with the depth of field issues. But imagine a return to "Lawrence of Arabia" level of photographic quality and presentation in movie theaters...

Every time I see a 70mm presentation of "2001" or "Lawrence of Arabia", like at the Aero, I think about what we've been missing for decades now. The closest since then was perhaps the 35mm anamorphic print I saw of "The Thin Red Line" which was quite good (I probably saw a show print at the Mann Village in Westwood when it opened.)

First, though, I'd love to see RED movies regularly projected in 4K but with another two stops of latitude in the camera.

Whereas just in the last two years, 2K D.I.'s have gone from being occasional to being commonplace, and scanning at 4K and then downsampling to 2K has become a little more common, though rare -- I don't see as much of a push for all-4K D.I. work yet. We need some new leap in scanning / processing speed and cheaper storage for that to happen --- as it is, most post houses are having a hard enough time with the increase in demand for 2K D.I.'s, so the time spent scanning and the amount of storage needed just for the larger number of projects are already causing most D.I. facilities to max out their capacity. I think the first priority will be to make 2K D.I.'s easier to achieve on a regular basis at more facilities before anyone seriously tackles jumping up to all-4K D.I. workflow except for one or two high-profile projects a year.

I mean, you can probably say that for the average movie released theatrically -- let's say a "Drillbit Taylor" -- a 2K D.I. is more than adequate. In the push to raise technical standards, we forget that a lot of stuff just gets dumped in the market every month that probably doesn't need more resolution, nor would they take advantage of it creatively if they got it. Just as it would be a waste to shoot a lot of those movies in 65mm.

 

[Steven-Marc C.]

In terms of cameras, the real holy grail, for me, would be a 65mm sensor with 6K resolution downsampled to a real and super clean 4K, with 13 stops.

Don't forget digital refocusing!
http://www.refocusimaging.com

 

[Dj Joofa]

I don't see as much of a push for all-4K D.I. work yet.

Please see the MTF graph in Glenn Kennel's paper of May 1991 issue of SMPTE journal regarding DI resolution on a film negative, which he further explains in his book, saying that at 25% drop in lines/mm, the relative MTF response for 2K was 40%, 3K 50% and 4K 53%. So going to 3K from 2K still has some merit, while going 4K from 3K was just law of diminishing returns. (All these numbers are based upon if my memory serves me correctly.)

 

[Policar]

Yes, that would be really cool, though you'd need a great focus puller and the best lenses possible to take advantage of that resolution and deal with the depth of field issues. But imagine a return to "Lawrence of Arabia" level of photographic quality and presentation in movie theaters...

Every time I see a 70mm presentation of "2001" or "Lawrence of Arabia", like at the Aero, I think about what we've been missing for decades now. The closest since then was perhaps the 35mm anamorphic print I saw of "The Thin Red Line" which was quite good (I probably saw a show print at the Mann Village in Westwood when it opened.)

First, though, I'd love to see RED movies regularly projected in 4K but with another two stops of latitude in the camera.

Whereas just in the last two years, 2K D.I.'s have gone from being occasional to being commonplace, and scanning at 4K and then downsampling to 2K has become a little more common, though rare -- I don't see as much of a push for all-4K D.I. work yet. We need some new leap in scanning / processing speed and cheaper storage for that to happen --- as it is, most post houses are having a hard enough time with the increase in demand for 2K D.I.'s, so the time spent scanning and the amount of storage needed just for the larger number of projects are already causing most D.I. facilities to max out their capacity. I think the first priority will be to make 2K D.I.'s easier to achieve on a regular basis at more facilities before anyone seriously tackles jumping up to all-4K D.I. workflow except for one or two high-profile projects a year.

I mean, you can probably say that for the average movie released theatrically -- let's say a "Drillbit Taylor" -- a 2K D.I. is more than adequate. In the push to raise technical standards, we forget that a lot of stuff just gets dumped in the market every month that probably doesn't need more resolution, nor would they take advantage of it creatively if they got it. Just as it would be a waste to shoot a lot of those movies in 65mm.

I've seen LoA in 70mm and it is pretty amazing--way better than what you see in theaters today. But I also saw the premier of a well shot action movie that was Super35 and had a DI (presumably 2K) and it looked amazing on a huge screen. The print they used for the premier was obviously much better than your average theatrical print.

So, really, I think the whole issue of format doesn't really matter to any extent if not seen in the context of projection. I saw No Country for Old Men in a theater with the focus so thrown off it was almost unwatchable at times, and that's a beautiful movie, too. The average print looks pretty bad, but if 4k digital projection catches on, that will be a pretty significant improvement and then format may matter much more significantly... Right now, no audience members in an average theater care about the difference between a 2k or 4k DI because in an average theater the image is too far degraded. (And, of course, some prefer the 2k DI: Bill Pope, for instance, likes how it reduces noise from faster stocks.)

But, ironically, the same transition to 4k projection, while making image quality more important again, may also encourage the use of an apparently inferior medium (digital). If at any point digital is significantly cheaper and easier to use than film, there will be a quick transition. The lack of archival ability, workflow issues, and fact that film is projected as film are the only things that really need to be overcome.

Image quality really isn't that important to the studios once you get to "good enough," besides. Technicolor was replaced by the inferior kodak color stocks; nitrate replaced by inferior celluloid; etc.

Lastly, re: dynamic range, while I think it makes a huge difference, it may not be as important to the end result as we all think. If you're shooting indoors, you can control your highlights, anyway, and even if you're shooting outdoors there are methods of control. And with a real time histogram, exposures can be made more precise. So it may be a bigger pain on set...but the resulting image can be as good, even without the extra few stops.

Plus, 3D seems conducive to a far deeper DoF and this is being pushed heavily by the studios right now. If 3D catches on, deep focus will become more prominent. If anything, I think a transition to 2/3'' 1080p chips is more likely than one to 65mm "8k" ones...at least in the short run.

 

[David Mullen ASC]

Please see the MTF graph in Glenn Kennel's paper of May 1991 issue of SMPTE journal regarding DI resolution on a film negative, which he further explains in his book, saying that at 25% drop in lines/mm, the relative MTF response for 2K was 40%, 3K 50% and 4K 53%. So going to 3K from 2K still has some merit, while going 4K from 3K was just law of diminishing returns. (All these numbers are based upon if my memory serves me correctly.)

Take a look at the examples here:
http://digitalcontentproducer.com/mag/video_digital_cinemas_special/index.html

It does suggest some loss is happening with the 2K scans, though it also suggests that scanning at 4K and then immediately downsampling to 2K is just as good as all-4K.

 

[David Mullen ASC]

Lastly, re: dynamic range, while I think it makes a huge difference, it may not be as important to the end result as we all think. If you're shooting indoors, you can control your highlights, anyway, and even if you're shooting outdoors there are methods of control. And with a real time histogram, exposures can be made more precise. So it may be a bigger pain on set...but the resulting image can be as good, even without the extra few stops.

Plus, 3D seems conducive to a far deeper DoF and this is being pushed heavily by the studios right now. If 3D catches on, deep focus will become more prominent. If anything, I think a transition to 2/3'' 1080p chips is more likely than one to 65mm "8k" ones...at least in the short run.

I don't see 3D becoming the norm for features, not the average dialogue-heavy mid-range movie like a drama or comedy. More for some genre pictures. Besides we already have 2/3" 3-CCD 1080P cameras. But I agree that a transition to 65mm sensor cameras is unlikely. I think 35mm-sized sensors will become more or less commonplace for digital cinema work.

Certainly a limited dynamic range can be compensated for, but it requires more work, which may require more time and money, which sort of works against the argument for shooting digitally. A wider dynamic range allows things to move faster on a shoot as you encounter a wide range of contrast in daily shooting. For example, on my current feature, we were in lobby of a building on Thursday with large skylights and the sun was moving in and out of fast-moving clouds, causing the light levels in the room to rise and fall by a couple of stops, sometimes during a take. With the latitude of film I know I can fix some of the shots where it got too bright momentarily. There are many movies that end up shooting in challenging lighting conditions where a wide latitude is the only thing that saves you.

But, yes, you can certainly create successful, beautiful images with the current latitude of digital cameras. I just want more flexibility, that's all.

 

[Policar]

What you're saying makes a ton of sense and that's a great example; sorry if I came off as contrary; I was just playing devil's advocate to some extent.

I do think that the transition from film to digital will have more to do with economic factors than how good digital looks (since it already looks "good enough"), but that's a whole other issue. As for 3D, who knows? I'm just looking forward to Avatar.

 

[Deanan]

The print they used for the premier was obviously much better than your average theatrical print.

So I wonder if that's why Vision Premier is called 'premier'... because it's only really used for premieres

 

[Uli Plank]

Besides we already have 2/3" 3-CCD 1080P cameras. But I agree that a transition to 65mm sensor cameras is unlikely. I think 35mm-sized sensors will become more or less commonplace for digital cinema work.

Talking about three letter cameras: Sony is just making the move to 35mm-sized single sensor with the F35…

Regards,

Uli

 

[Álex Montoya]

Responding to David, whose posts are always a pleasure to read, I hope that build 16 can scrap an extra stop from the noise floor.

That way we can shoot with the camera as if it was rated 600, so we can shoot with less light while gaining that extra stop for highlight protection.

BTW, I got to know that a well known DP that recently used the RED for and spot, underexposed everything by a stop and had all the monitors gain compensated to make it look good for the client.

 

[Dj Joofa]

Take a look at the examples here:
http://digitalcontentproducer.com/mag/video_digital_cinemas_special/index.html

It does suggest some loss is happening with the 2K scans, though it also suggests that scanning at 4K and then immediately downsampling to 2K is just as good as all-4K.

David, thanks for the link. Interesting read. The finding there are in agreement to Kennel's, for 2k vs. 4k DI scans. However, as Kennel suggesed that 3k and 4k are not really that different even to begin with, it would have been interesting to see if Matt Cowan had provided samples of 3k scans.

 

[David Mullen ASC]

I think 3K scans are a more recent phenomenon thanks to the Arriscanner I believe. Generally most D.I. work is in 2K or 4K, not 3K.