new digital camera better than 35mm film?!

[super8man]

Have they figured out why digital cameras can't handle the color RED yet? Seriously.

I'll stick with film. Leave digital for real estate salesman and youtube.

[sarmoti]

Let's use this amazing new technology to transfer our films.
Then we have the best from both worlds 8)

Fred.

I doubt that it'll be a better option than using a telecine like the Spirit for that application. The machine vision cameras you pointed out can probably do that job cheaper than RED.

They don't have:

A Camera

It is my understanding that at this point Red only has proof of concept. They don't have an actual camera. Therefore the whole discussion revolves around a big "What If??"

I remember going through the same steps with the Viper, Dalsa, D20, SI, Genesis and countless other products. They were all announced and were "in development" for a couple years before we could shoot with them. I don't doubt that a determined billionare with a goal and the well known team of people behind him are capable of putting such a product on the market. Wether they can deliver it for that price and on time is another question.

Have they figured out why digital cameras can't handle the color RED yet? Seriously.

I thought this was solved when component video recording was released back in the 80's. Seriously, unless you use a composite video format or DV I can't say that it's been a real problem for a long time.

[VideoFred]

I doubt that it'll be a better option than using a telecine like the Spirit for that application. The machine vision cameras you pointed out can probably do that job cheaper than RED.

Yes, this was my point.
No doubt about the Spirit of cource.
But not for transfering huge amounts of film... to expensive.

Those new Cmos sensors could make 2K or even 4K scans possible for us, amateurs.

As soon as I have some time, I will try to get one for testing.
Not the RED of cource, but a Cmos machine cam.
Specs from these cams are looking very promising.
But this is all theory... The only way to know is testing.

I might do a test with 16mm, too.

However, capturing in high resolution is one thing,
working with these huge files is another thing!
Computers are not ready yet for this.


Fred.

[aj]

Have they figured out why digital cameras can't handle the color RED yet? Seriously.

I'll stick with film. Leave digital for real estate salesman and youtube.

You mean why digital sensors are overly red sensitive and why this is managed badly on cheap cameras? I.e. image record through human skin and such. You get to see people's vein structures and so on.

It is much the opposite of the early days of silver based film. I.e. pre pan-chromatic film.

It is usually handled with an Infra/red and red blocking or partial blocking filter.

But of course I cannot explain why. It is the physics of the sensor capacitor plates. These seem to be very sensitive to long wavelength light (red).

It good for astrophotography though Noise is a troublemaker there. Somebody should do something about that too.

[sarmoti]

sorry, double post.

[sarmoti]

However, capturing in high resolution is one thing,
working with these huge files is another thing!
Computers are not ready yet for this.


Fred.

You should check out the Cineform wavelet based codecs. I've been using it for a few years to deal with large amounts of high def content while travelling, it allows realtime workflows on even modest computer systems. It does a very good job and for most applications, it's virtually identical to uncompressed footage (yet better than DVCPROHD and HDCAM). Currently it's limited at 10bit 2K images but they've recently demoed realtime 4K playback and editing (with transitions!) on a current computer system.

[VideoFred]

Thank you for this hint!
I will certainly remember this for possible future use.

But can this be used as a stand alone codec?


Fred.

[sarmoti]

Thank you for this hint!
I will certainly remember this for possible future use.

But can this be used as a stand alone codec?


Fred.

To a certain extent, and you obviously need to have the codec installed in any system that's going to use it. There's a 15 day free trial on their website, it's worth a shot.

[idrawthings]

I'm surprised none of the other LA forum members went to the screening they had this week. Red screened their test footage, projected on a Sony 4k projector. I've gotta hand it to Red, the footage was damn incredible, looked like very clean 35mm, great DOF. Tack sharp too. Only drawback on the footage was they didn't shoot anything with a tremendous latitude. Then, the tests of the Red cam being posted on cinematography.net are pretty promising.

There is a camera called Frankie. Last I read, they begin prototyping the final design in 3 months, then production run in 6 or 9 months.

As for the storage, it lets you shoot out either RAW (something like 300Mb per frame) or encoded with their codec (about 72 Mb), which you can then down convert to whatever 2k HD or SD you want. They screened RAW footage against their Red codec, and I couldn't see a significant loss.

For as much as I love film, I welcome the Red and hope it stirs up the industry a bit. A 4k camera that's cheaper than say, an IMX, is likely to do so.

[reflex]

I've gotta hand it to Red, the footage was damn incredible, looked like very clean 35mm, great DOF

Interesting to hear, although don't mistake shallow DOF as a sign of image quality. It's a function of lens aperture, focal length and frame size (or sensor size in this case) and has nothing to do with the quality of the image sensor. Effective use of DOF merely demonstrates that the cinematographer has technical skill.

[christoph]

As for the storage, it lets you shoot out either RAW (something like 300Mb per frame) or encoded with their codec (about 72 Mb)

make those numbers "per second" ;)
++ christoph ++

[idrawthings]

reflex - yeah, they kept touting the size of the CMOS sensor. It's a hair bigger than a super 35 frame. but then, they never disclosed any details of the shoot, particularly the fstops.

christoph - dyslexia kicked in, their codec scales it down to 27Mb, not 72, virtually loseless. the onboard flash drive (32 - 128 GB) only allows their red codec. To dump raw you'd need one hell of a drive array (it does have eSATA out) with a whole lotta empty platters.

oh yeah, it can shoot variable frame rates - at 4k 1-60 fps and 2k 1-120 fps. wicked.

check out red.com for the specs. a 'hazzah' to them if they pull it all off.

[Nigel]

Lets wait and see...

I wonder who is making their "Mysterian" or whatever it is called CMOS. Time will tell how everything shakes out.

Good Luck

[sarmoti]

It's funny reading these old posts knowing all that has happened in the past few years.

[carllooper]

Well, the Red introduced the shakeup, but the concept of a camera like the red was in the works for at least a decade prior to that. The dream was a digital camera that could provide at least the same kind of image as 35mm. I was working on various designs for such a camera in the late 90s until I realised there was a particularly false assumption in video and digital technology.

The problem with the Red, and the cameras on which they were based, and those that have followed such, is that they are based on a Cartesian model of the image, rather than a statistical one. If the desire was to create an image the same as film they were going about it in completely the wrong way. What was required was a more novel approach to sensor design rather than just refinement and scaling of ideas from the television/video era. Basically they didn't really appreciate the nature of the film image which inspired them. Add to that a kind of arrogance about the superiority of digital and it didn't sit very well at all. But it made a lot of accountants happy.

The big change wasn't the cameras. It was the digital projector. It was the colonisation of the cinema, by digital projectors, that meant it was no longer necessary to deal with film at all.

But for the digital pioneers it's still very much early days - there's still plenty of ideas to be explored in the digital domain, for producing new types of digital photography, that might satisfy some of the original inspiration as far as look and feel goes.

For example, instead of a rectangular grid of sensor cells, the sensor could (or should) adopt this sort of architecture:



In addition the sensor could (or should) move around in a random fashion (at a frequency higher than the frame rate).

The combination of these ideas could (or would) provide something more comparable to the way in which film finds an image, while at the same time retaining the virtues of digital sensors, such as low light sensitivity and high frequency response.

Until then digital will exhibit a sampling frequency cutoff (no matter how many pixels there will be smaller ones possible), and the regular grid like spacing produces aliasing, both of which gives it that video look, and that unavoidable sense of a world in which something fundamental is missing.

An alternative is to treat any digital technology (and video before it), not as some way of obtaining something that looks like film, and definitely not as a replacement for film (because whichever way you look at it, it simply isn't film anyway), but as a technology in it's own right, with it's own particular look (or a new look) that can be exploited in it's own way, for what it is in itself, rather than the extent to which it does or does not coincide with the look of film.

And of course, for those so inclined (such as myself) to exploit film technology for what it is, in itself. There is so much you can do in film that you simply can't do in digital (and vice versa of course). They are just so completely different in terms of how they are made and how they work. Lenses, and a light proof box remain about the only thing in common.

And analog/digital hybrids. I'm into that sort of thing as well.

Of course, most people don't really care, but the sheer number of such people provides no reason whatsoever to join them.

C