Why 4K?

[Martin Stevens]

Sorry to disagree on this one. What exactly did you see that looked essentially identical?

Stop by RED and we would be happy to show you 4K footage that doesn't look like 2K anything.

Jim

OUT-OF-FOCUS projection is to blame here perhaps.

I hate to say this, but I actually prefer watching films on my 58" plasma with Blu-ray, because far too
often the local theatre projects the print so out-of-focus I could scream......and if its out-of-focus,
then even 4K will look rather poor at best.

 

[Tom Lowe]

All of this talk is going to be forgotten once people really get a chance to shoot on 5K+ RAW cinema cameras and see that footage projected at 4K RGB. Do you really think you are going to want to go back to film scans after that??? I absolutely guarantee that you will not. It's one thing to sit around talking about this and that... it's another to go out and shoot 5K+ RAW and see how ridiculously amazing -- dare I say jaw dropping -- the quality truly is. You guys should be listening to Jim very carefully on this subject. He's dead right about 4K.

Overall, digital is going to blow a hole through film, just like it did with stills photography.

Yesterday I was testing out the look of various Bluray rips on my 50" screen. I wanted to compare film scans vs digital acquisition. For film, I chose some serious Bluray contenders -- Baraka (65mm scanned at 8K), IMAX Space Station (65mm 15-perf), and The New World (recent 35mm Panavision anamorphic). To be honest, none of those films could hold a candle to Revenge of the Sith (F950R 4:4:4) in terms of detail, clarity, color, and cleanness of the image in 1080p digital display. And my Sith 1080p copy was an HDTV rip, probably much lower bitrate than true Bluray. All of the scanned films, including 65mm and IMAX films, also got blown out of the water by the insanely clean and beautiful Bluray image of WALL-E.

I don't care what anyone says. Scanned film looks worse to me than digitally acquired material. I remember shooting on 35mm Kodak film on my old film Canon film camera, and then scanning it. The scans always looked like crap. Then I bought a $250 digital Canon camera and all of a sudden, my photos looked amazing in digital format!

Digital 4K display and 3D are the way of the future.

 

[Julian Gerretsen]

OUT-OF-FOCUS projection is to blame here perhaps.

I hate to say this, but I actually prefer watching films on my 58" plasma with Blu-ray, because far too
often the local theatre projects the print so out-of-focus I could scream......and if its out-of-focus,
then even 4K will look rather poor at best.

With prints, it is often the practice that the projectionist will loosen up the gate slightly so as to avoid wear and stress on the film and reduce consequential breakage. This is an economic choice designed to safeguard the print for as long as possible. Unfortunately, a looser gate will almost certainly lead to a softer projected image. Theatres might be under pressure to meet certain breakage "targets".

A theatre that projects with a well locked down gate will have a sharper image but more stress on it's films/breakage ratio, thus getting a bad name from distributors. So perversely, striving for projection excellence and the best possible image for cinema-goers will lead to penalisation on the distribution side.

Of course, this shouldn't be an issue on digital projection though, unless they haven't got the focus set right. Maybe you and others should have words with your local theatre manager....

 

[Julian Gerretsen]

I don't care what anyone says. Scanned film looks worse to me than digitally acquired material. I remember shooting on 35mm Kodak film on my old film Canon film camera, and then scanning it. The scans always looked like crap. Then I bought a $250 digital Canon camera and all of a sudden, my photos looked amazing in digital format!

I totally agree. I've been scanning older 100 ISO negs shot with my F3 on quality Nikkor glass. The scans are of sufficiently high enough resolution/quality to zoom in on the individual film grain structure.

No matter how much you play around with these in Photoshop, they don't in any way match the (properly exposed) images I am getting off even my older 6.1MP D70s. And that's the problem with the film scans, at a certain point you are just getting higher resolution images of film grain.

 

[Andrew M.]

It does not really increase the perception of resolution, or at least not by much in a noticeable way. Our eyes can only be fooled into seeing so much resolution that isn't there....

Jeff, if I would not be a bit nearsighted I would almost took your logic for granted.

I agree that if the frame is flashed twice you would not get perception of higher resolution unless it was acquired twice by the camera in two slightly different intervals. But I want to attract your attention to the fact that in 3D picture acquisition, you have two images of 2K acquired by two cameras with slightly different view on the objects.
It is almost like pixel shifting technology but somehow in our brains it works much much better.

I should were glasses (-1.5 dioptres) when driving but I don’t. When I look on the street name signs with my one eye I can’t read it at dusk time. When I try to read this sign with my other eye the same thing, can read it. But when I open both my eyes no problem at all. How come? And the difference is more like 2K to 4K believe me. How does this works?

Also recently I am fascinated with window like experience with 4K images and new color science.

Especially the progress in perfect reproduction of colors is promising.
How much resolution and color fidelity we need to fool the viewer that he is actually looking out of the window when watching the still picture.

Kind of Turing test but for images instead of intelligence.
http://en.wikipedia.org/wiki/Turing_test

Anybody wants to be Mr. Alan Turing of image technology?
Define the test than.

For the beginning I would propose the test for one eye only with limited field of view for window and the image, no moving objects. And we will use teenagers as a judges to start with


Andrew

 

[Gary Paul Walker]

There’s lots of talk here and in some of the other threads of ‘future-proofing’ and can’t help thinking that it shouldn’t really be that applicable here.
You future-proof a computer on the understanding that the editing software you intend to use will require a higher spec, but what are you ‘proofing’ against in terms of the actual resolution?

 

[Scott Crawley]

Yesterday I was testing out the look of various Bluray rips on my 50" screen. I wanted to compare film scans vs digital acquisition. For film, I chose some serious Bluray contenders -- Baraka (65mm scanned at 8K), IMAX Space Station (65mm 15-perf), and The New World (recent 35mm Panavision anamorphic). To be honest, none of those films could hold a candle to Revenge of the Sith (F950R 4:4:4) in terms of detail, clarity, color, and cleanness of the image in 1080p digital display. And my Sith 1080p copy was an HDTV rip, probably much lower bitrate than true Bluray. All of the scanned films, including 65mm and IMAX films, also got blown out of the water by the insanely clean and beautiful Bluray image of WALL-E.

Digital 4K display and 3D are the way of the future.

And on 3/23 you will be able to view Days Of Heaven on Blu. How excited you must be. ;-)

 

[David Rasberry]

Either way, the Scarlet-fixed is my next step up from an HDV-level Canon HV-20. Having shot a lot of self-projects and small films, I really desire something that looks better (slow-motion is also a key seller for me). I'm not sure how many others are in the same boat as me.... as most people here seem to be a few $$ steps up from me.

-cc

I'm looking forward to the Scarlet Cinema model. Been following Red's development for nearly two years now, and nothing else under $10k seems very interesting anymore.
For me it is a personal toy more than a pro tool, so I am content to wait for it.

 

[David Rasberry]

Yesterday I was testing out the look of various Bluray rips on my 50" screen. I wanted to compare film scans vs digital acquisition. For film, I chose some serious Bluray contenders -- Baraka (65mm scanned at 8K), IMAX Space Station (65mm 15-perf), and The New World (recent 35mm Panavision anamorphic). To be honest, none of those films could hold a candle to Revenge of the Sith (F950R 4:4:4) in terms of detail, clarity, color, and cleanness of the image in 1080p digital display. And my Sith 1080p copy was an HDTV rip, probably much lower bitrate than true Bluray. All of the scanned films, including 65mm and IMAX films, also got blown out of the water by the insanely clean and beautiful Bluray image of WALL-E.

Comparing live footage film to digital animation is perhaps not a definitive test. With HDTV we have become so used to seeing the over enhanced, over edge sharpened, and oversaturated color that is typical of today's broadcast practice and most out of the box consumer TV's that haven't been well calibrated. It's kind of like running the stereo with the bass and treble cranked wide open. Doesn't look natural at all.

Not to say I don't agree with you that current gen digital is surpassing the capabilities of film as an acquisition medium. Just that they are two different aesthetics and each stands on its own.

 

[Michael Brennan]

5k sensor x 1 or or 7k sensor x 3?

5k sensor x 1 or or 7k sensor x 3?

Is 5k is future proof?

Depends on your definition of future....

Super Hi Vision has been off the radar for a little while but it hasn't gone away.

Some history to put it into context.
After first beginning to study HD (they called it Hi Vision) in 1964, NHK Japan were producing HD programs in 1982.

Then they started research into the next big thing, Super Hi Vision in 1995 and demo'ed it in 2005.

Super Hi vision is 7680 x 4320 pixels x R x G x B, or 99 million pixels total, versus 6.2 million for current HD.

1.5 million people saw it in action in 2005 at Japanese World Expo.
It was demonstrated at NAB and IBC in 2006 and it was/is amazing.

NHK formed a consortium of broadcasters to aide development and at IBC 2008 they transmitted live for the first time.

Their road map circa 2009 indicates experimental broadcasting now and broadcasting proper in 2020.
The cameras used up to now comprised 4 x 8 million pixel sensors. 2 were green (and pixel shifted to get the res) the other two blue and red.
So with "only" 32 million pixels it was not full res.

BUT last year they said;
"We will develop a prototype full-resolution camera with three 33-
million-pixel image sensors by early 2011. We expect that display manufacturers will develop a flat screen display with the full pixel count by 2011, as well."

Given the accelerated pace of the planned introduction of HD 3D one wonders if this timeline will compress.

The ITU standard being discussed is REC.ITU-R BT.1769
60 frames progressive at 7680 x 4320 pixels or 32x more information than 60i HDTV.

They have always said that 3D experience goes hand in hand with LSDI (Large screen Digital imagery)

No doubt they have their heads down testing Super Hi Vision 3D at this very moment.


It seems likely that RED can scale to this spec, easy for me to say! but probably not to the complete satisfaction of NHK engineers who love 3 chips and a prism!

Uncompressed SHV data rate is around 24 Gbit/s!
On todays exchange rate, thats one frame per flash card



Mike Brennan

 

[Graeme Nattress]

The NHK demos are indeed very high resolution, but they lack a certain imaging aesthetic that makes for engrossing, dramatic imagery of the kind we like to view. As I said before, resolution is necessary, but not sufficient. NHK proves the "not sufficient" aspect admirably.

Graeme

 

[Elsie N]

Is 5k is future proof?

Depends on your definition of future....

Super Hi Vision has been off the radar for a little while but it hasn't gone away.

Some history to put it into context.
After first beginning to study HD (they called it Hi Vision) in 1964, NHK Japan were producing HD programs in 1982.

Then they started research into the next big thing, Super Hi Vision in 1995 and demo'ed it in 2005.

Super Hi vision is 7680 x 4320 pixels x R x G x B, or 99 million pixels total, versus 6.2 million for current HD.

1.5 million people saw it in action in 2005 at Japanese World Expo.
It was demonstrated at NAB and IBC in 2006 and it was/is amazing.

NHK formed a consortium of broadcasters to aide development and at IBC 2008 they transmitted live for the first time.

Their road map circa 2009 indicates experimental broadcasting now and broadcasting proper in 2020.
The cameras used up to now comprised 4 x 8 million pixel sensors. 2 were green (and pixel shifted to get the res) the other two blue and red.
So with "only" 32 million pixels it was not full res.

BUT last year they said;
"We will develop a prototype full-resolution camera with three 33-
million-pixel image sensors by early 2011. We expect that display manufacturers will develop a flat screen display with the full pixel count by 2011, as well."

Given the accelerated pace of the planned introduction of HD 3D one wonders if this timeline will compress.

The ITU standard being discussed is REC.ITU-R BT.1769
60 frames progressive at 7680 x 4320 pixels or 32x more information than 60i HDTV.

They have always said that 3D experience goes hand in hand with LSDI (Large screen Digital imagery)

No doubt they have their heads down testing Super Hi Vision 3D at this very moment.


It seems likely that RED can scale to this spec, easy for me to say! but probably not to the complete satisfaction of NHK engineers who love 3 chips and a prism!

Uncompressed SHV data rate is around 24 Gbit/s!
On todays exchange rate, thats one frame per flash card



Mike Brennan

Wow! Sounds like buying a GPU stock would be one way to cash in on something like this. Wonder how big a flock/gang/pride/covey etc. of Rockets it would take to stream this? But don't you need a football stadium for this to matter?

(Maybe I should have said "wonder how big a BOX of Rock_'s it would take to stream this?")

 

[Steven Caesare]

I'm all for a raising of general standards and higher-quality options... as long as there isn't going to be some "resolution police" out there telling me I can't use diffusion filters or old lenses, etc.

...SNIP...
My point is that we make movies, not shoot resolution charts. Not every movie requires that every pore on an actor's face come into razor-sharp focus. We have a future ahead of us where we will -- and should -- see all types of photography, crisp & hard, soft & delicate, etc. And some of it won't measure-out to 4K.

I'd also point out that Red makes cameras that shoot 2K RAW for slow-motion shots, which would not meet the criteria for resolution that they are saying will necessary to future-proof their movies. They are also making a 3K RAW camera that will only measure 2-point-something-K in resolution.

The general idea I get from Jim's posts on the subject is that he (and we?) are looking for a digital successor to film. Somethng that at least equals it's characteristics, and perhaps even succeeds some of them, as even film release prints often measure far worse than the original acquisition specs.

There are a number of digital acquisition/display systems already out there, but they all suffer from at least one major flaw as compared to film, and resolution is almost always one of them, although DR, color reproduction etc... are also often issues.

Therefore Jim is suggesting that a worthy goal is a pipeline from acquisition to display that allows for fidelity as good as, or better than, film. In his estimation, 4K is the goal. In the same way that 35mm film prints are currently "standard".

Does the best material measure out to the maximum resolution that projection film stock is capable of? No. Does the best digital aquisition measure out to the limits of a 4K projection system? No. But they both are in the same neighborhood in terms of percentage of resolution they can muster.

And both of those values are FAR greater than the alternative proposals that center around 2K-ish acquisition/distribution. And the results are cleary discernable for many people.

Does this mean that for artistic purposes you can't cut in 16mm footage, a shot on an old lens that degrades the image in a pleasing way, a stack of filters, HDV footage, etc...? Of course not. It's done for film today all the time.

But by giving the largest digital "pallette of resolution" to work with, those lower fidelity shots can easily be accomodated. If the pallette is constrained to begin with, then the artisitic ability to present a wide desert vista in "Lawrence of Arabia II: Electric Boogaloo" in stunning 4K is NOT available to the artsit.

At least that's how I read it.

-sc

 

[Michael Brennan]

The NHK demos are indeed very high resolution, but they lack a certain imaging aesthetic that makes for engrossing, dramatic imagery of the kind we like to view. As I said before, resolution is necessary, but not sufficient. NHK proves the "not sufficient" aspect admirably.

Graeme

Yes, they are probably gunning for the highest res experience at the moment rather than a particular aesthetic. Like the early HD cameras.
Techniques to achieve various aesthetics in digital cameras is far better understood now than when HD was introduced in 1990s, so perhaps the question is if modern looks can be tweaked processed or otherwise incorporated into the image chain at 24Gbit/s!


It may be that life like resolution ie 100 million pixels per screen and "suspension of belief" will not be happy partners in any event.

On the other hand who would have thought that IMAX would be screening narrative movies shot for the smaller cinema screen?

Yesterdays technique of sometimes creating an unreal image to help suspend belief and so create a mood to help tell a story may be overtaken or at least ride in parallel by a wholly immersive crystal clear experience. The public' seem to have an appetite for "crystal clear immersive' if the success of 3D in the box office is a guide.

"Crystal clear immersive" itself, can have many moods and aesthetics of course.

Fun to dust off the crystal ball and ponder.


Mike Brennan

 

[Julian Gerretsen]

Either way, the Scarlet-fixed is my next step up from an HDV-level Canon HV-20. Having shot a lot of self-projects and small films, I really desire something that looks better (slow-motion is also a key seller for me). I'm not sure how many others are in the same boat as me.... as most people here seem to be a few $$ steps up from me.
-cc

Chris, I am in pretty much the same boat as you. I've been using a Canon XH-A1 for the past 2.5 years for commercial type work, Sony VX2000 and various others before that. Although HDV has been good, I've always been waiting expectantly for something like RED. I've been scraping the money together for a feature film which has been in the slow cooker for about 3 years.

I didn't want to go down the 35mm-adaptor-on-HDV route, as although you get the depth of field, you are still at the end of the day dealing with a highly compressed 1440X1080 resolution. I am so used to fantastic image quality from my Nikon DSLR's, I wanted the same kind of resolution in moving pictures. I had at one stage thought that 16mm would be my only option for decent, reasonably economical acquisition, but of course, RED has changed all that. Yes $17,500 for an R1 is a major outlay (+ various essential accessories and storage), but it's an incredible jump to pro-levels of image quality. All this for not much more than the cost of stock and processing of 16mm... and of course with a far superior image quality.

I think many people on reduser are in the same boat... not particularly blessed with major finances, but knowing that RED can allow us to realise the dreams that before were pretty much out of reach.

 

[David Rasberry]

.

I didn't want to go down the 35mm-adaptor-on-HDV route, as although you get the depth of field, you are still at the end of the day dealing with a highly compressed 1440X1080 resolution. I am so used to fantastic image quality from my Nikon DSLR's, I wanted the same kind of resolution in moving pictures. I had at one stage thought that 16mm would be my only option for decent, reasonably economical acquisition, but of course, RED has changed all that. Yes $17,500 for an R1 is a major outlay (+ various essential accessories and storage), but it's an incredible jump to pro-levels of image quality. All this for not much more than the cost of stock and processing of 16mm... and of course with a far superior image quality.

I think many people on reduser are in the same boat... not particularly blessed with major finances, but knowing that RED can allow us to realise the dreams that before were pretty much out of reach.

I'm an old (60) 16mm shooter. Scarlet 2/3" will be the first digital camera that many of us can afford to own that should equal or better the quality of S16mm film.
Haven't been this excited about a new camera since my first Bolex. I'm planning on using the cinema version with vintage 16mm lenses, primarily a Cooke 9-50mm and an Angenieux 12-240.

 

[Rob Ruffo]

Jim & TED,

We also needs to keep in mind one thing that the current resize algorithm is not good enough for 1080p / SD finish and that is also the reason people choose those cameras.

Do we really need 4k except for 3D?

You can use any resize algorythm you want, including even running a batch in Photoshop using any number of plugins. One you have a series of 4K digital images, the choice is yours. There are at least 6 excellent algorithms for downsizing. I see no point here.

 

[Rob Ruffo]

Yes, they are probably gunning for the highest res experience at the moment rather than a particular aesthetic. Like the early HD cameras.
Techniques to achieve various aesthetics in digital cameras is far better understood now than when HD was introduced in 1990s, so perhaps the question is if modern looks can be tweaked processed or otherwise incorporated into the image chain at 24Gbit/s!


It may be that life like resolution ie 100 million pixels per screen and "suspension of belief" will not be happy partners in any event.

On the other hand who would have thought that IMAX would be screening narrative movies shot for the smaller cinema screen?

Yesterdays technique of sometimes creating an unreal image to help suspend belief and so create a mood to help tell a story may be overtaken or at least ride in parallel by a wholly immersive crystal clear experience. The public' seem to have an appetite for "crystal clear immersive' if the success of 3D in the box office is a guide.

"Crystal clear immersive" itself, can have many moods and aesthetics of course.

Fun to dust off the crystal ball and ponder.


Mike Brennan

Yet Avatar is far from "crystal clear" (lots of thin DOF), and it is highly expressionistic on many levels. Did pretty OK at the box office from what I hear

 

[Rob Ruffo]

All of this talk is going to be forgotten once people really get a chance to shoot on 5K+ RAW cinema cameras and see that footage projected at 4K RGB. Do you really think you are going to want to go back to film scans after that??? I absolutely guarantee that you will not. It's one thing to sit around talking about this and that... it's another to go out and shoot 5K+ RAW and see how ridiculously amazing -- dare I say jaw dropping -- the quality truly is. You guys should be listening to Jim very carefully on this subject. He's dead right about 4K.

Overall, digital is going to blow a hole through film, just like it did with stills photography.

Yesterday I was testing out the look of various Bluray rips on my 50" screen. I wanted to compare film scans vs digital acquisition. For film, I chose some serious Bluray contenders -- Baraka (65mm scanned at 8K), IMAX Space Station (65mm 15-perf), and The New World (recent 35mm Panavision anamorphic). To be honest, none of those films could hold a candle to Revenge of the Sith (F950R 4:4:4) in terms of detail, clarity, color, and cleanness of the image in 1080p digital display. And my Sith 1080p copy was an HDTV rip, probably much lower bitrate than true Bluray. All of the scanned films, including 65mm and IMAX films, also got blown out of the water by the insanely clean and beautiful Bluray image of WALL-E.

I don't care what anyone says. Scanned film looks worse to me than digitally acquired material. I remember shooting on 35mm Kodak film on my old film Canon film camera, and then scanning it. The scans always looked like crap. Then I bought a $250 digital Canon camera and all of a sudden, my photos looked amazing in digital format!

Digital 4K display and 3D are the way of the future.

One thing that is rarely discussed about film scans is the introduction of more glass, glass that is never perfect. The scanner has many glass elements, including some form of lens or at least an OLPF. It's a bit like the problem with 35mm adapters. Light bent too far too many times can get a bit broken.

I've also been learning more about resolution. Maybe average eyes cannot make out more than 1080p at "normal" viewing distances (whatever normal means - Imax films are shot, and the theaters set up - for closer than THX recommended viewing for example, and some people sit very close to their flat-panel TVs). But your eyes use a kind of super sampling to determine gradients. The more points you have to perform this psycho-visual feet, the smoother the gradient will appear, and the more 3D-like the image - this well beyond the abilities shown by the resolution-chart like tests used to determine what resolution we can see.

This ability makes sense from an evolutionary perspective: in the forest subtle gradient differences on leaves would tell you what time of day it was (better get back to camp before dark!) your direction of travel and many other bits of info most useful to your survival.

 

[Chris Clogg]

Chris, I am in pretty much the same boat as you. I've been using a Canon XH-A1 for the past 2.5 years for commercial type work, Sony VX2000 and various others before that. Although HDV has been good, I've always been waiting expectantly for something like RED. I've been scraping the money together for a feature film which has been in the slow cooker for about 3 years.

I didn't want to go down the 35mm-adaptor-on-HDV route, as although you get the depth of field, you are still at the end of the day dealing with a highly compressed 1440X1080 resolution. I am so used to fantastic image quality from my Nikon DSLR's, I wanted the same kind of resolution in moving pictures. I had at one stage thought that 16mm would be my only option for decent, reasonably economical acquisition, but of course, RED has changed all that. Yes $17,500 for an R1 is a major outlay (+ various essential accessories and storage), but it's an incredible jump to pro-levels of image quality. All this for not much more than the cost of stock and processing of 16mm... and of course with a far superior image quality.

I think many people on reduser are in the same boat... not particularly blessed with major finances, but knowing that RED can allow us to realise the dreams that before were pretty much out of reach.

That's good to hear .