5K is alive...

[Cedric Akins]

Why can't car companies just build a sedan that has 600 horsepower and gets 40MPG? Answer: They can. Just not for a few years.

Jim

That is assuming they are still around by then.

 

[David Dennis]

That is assuming they are still around by then.

Mercedes-Benz, which I believe will still be around, just announced a prototype that combines a 6-cylinder drive with a hybrid motor and gets the equivalent of 268 horsepower + 43mpg in a body similar to the current S-Class (Mercedes' largest car). This is to be compared to the previous generation S-Class, which gets about 20mpg with 306hp. So it should genuinely be possible to get somewhere around there.

The reason the Big Three are in trouble is that they lost interest in developing genuinely interesting cars. It's tempting to say that if Jim ran a car company that could never happen, but I'm going to take a wild guess and say Jim would rather stick to cameras.

D

 

[GPSchnyder]

As soon as Scarlet comes out, we're gonna go ahead and have our next baby: we'll have two Scarlets in the delivery room on jibs shooting at 5K 24p, possibly ramping to 60p for the cutting of the umbilical cord. We'll have two more rigged up in Postpartum and even have one of the nurses rigged up via steadicam. :w00t:

I'd go 3D on this ...

:bleh:

 

[András Puiz]

I don't think it is prudent to open that thread again. Not many understand exactly how our sensor, DSP and compression engine works so there is a lot of random speculation about what we can, cannot or "should" be able to do. Since we are not in the position to explain it...

Jim

OK, but I was just going to ask, would it be possible to introduce an uncompressed (or mathematically losslessly compressed) format for use exclusively in still mode?

I guess still photographers are more hung up on compression artifacts, and the speed or bandwidth issues inherent to motion picture wouldn't apply that much either.

 

[David Mullen ASC]

In my case, the most likely scenario would be to shoot REDCODE for a feature, but if the occasional bluescreen shot came up and the efx person wanted uncompressed files, I could give that to him. It's not a question of whether the efx person was right or wrong for wanting uncompressed files, it's just that generally it's their call, not mine. They would probably bring the recorder with them.

 

[Mark L. Pederson]

Speculation and vaporware are both usually much different than misconception.

Jim

I LOVE this comment.

 

[Jannard]

In my case, the most likely scenario would be to shoot REDCODE for a feature, but if the occasional bluescreen shot came up and the efx person wanted uncompressed files, I could give that to him. It's not a question of whether the efx person was right or wrong for wanting uncompressed files, it's just that generally it's their call, not mine. They would probably bring the recorder with them.

I guess the question here is whether we should educate the efx guys that (say) 3:1 or 4:1 compression REDCODE is absolutely as good for their purpose as uncompressed... or just let them waste time and money because nobody told them.

Jim

 

[Michael Lindsay]

I guess the question here is whether we should educate the efx guys that (say) 3:1 or 4:1 compression REDCODE is absolutely as good for their purpose as uncompressed... or just let them waste time and money because nobody told them.

Jim

I would guess, not guaranteeing your success of course, the right thing is to put up a test. This should include a small white paper that explains the test and outcome. From a personal point of view I'd rather you do this after epic is released.. ;-)

Everything I've seen and read would suggest to me that 4:1 wavelet on Large images (3k,4k,5k+) is enough for anybody (obviously I too would be happy to see some tests). Dalsa's cinema business failure is a lesson in what is enough and what is needed. Red's marketing challenge (and educational challenge) is that you are targeting so many different types of owners, operators, post people etc... the price of success? Good luck!

Michael L

 

[Eric Lange]

In no particular order...

Still cameras do not have a global shutter.

If you shoot a non moving object, even a slow read-reset sensor will look great.

There are widely varying degrees of read-reset times for a rolling shutter. Some as fast as a film camera.

A global shutter will "feel" different than a rapid-reset sensor or a film camera when used for motion.

Assuming that we all get used to a global shutter look (less "filmic" and more sterile), technology will take us there is a few years. It has only been very recently that we can have a high speed, high dynamic range, good light sensitivity, and a rapid-reset sensor.

As we move forward, you are likely to see all aspects of the sensor improve. We are spending a ton of time and research money on this. It is one of the main reasons why we believe in a sensor upgrade, starting with the RED ONE, in all our cameras without having to buy a completely new camera every year.

It is impossible to find what you are looking for today. Sometime in the future, when we have delivered all you are asking for today (assuming you really want it once you have it), you will have a whole new set of wishes. That's how this whole deal works.

Why can't car companies just build a sedan that has 600 horsepower and gets 40MPG? Answer: They can. Just not for a few years.

Jim

Jim thanks for the information.

Yes global shutter does look sterile, but we really look for the spatially accurate series of frozen moments in time for our 3d work. Nice to get an answer straight from the “horse’s” mouth about the likely hood of higher res and frame rate globally shuttered sensors/processing. We will consequently work on special algorithms to glue 3d fragments extracted from multiple stereo camera views that fudge the 3d spatial errors caused by rolling shutter artifacts for dynamic subjects. In this sense (from your comments) we do not feel obliged to provide a hardware solution to the problem but rather find a near fix in software (I.e. maybe we don’t throw several/many millions of dollars at the problem either!).

As far as industrial cameras go, true global shutters are of the order of (2k x 2k 15fps 12bit @ $8k), (1K x 2k 30 fps @ $6k), (11mp 3fps (Kodak 35x24mm sensor) @ $9k) (fairly dismal looking images and frame rates but they are in some cases spatially and temporally very precise). So it seems at least for a few years we (in our work) will be stuck between a rock and hard place, as we can provide very precise 3d texture mapped models but the imagery looks lack luster (from globally shuttered industrial cameras), or (my preferred option) we can have gorgeous looking effects and images from RED One systems where the 3d fit has to be munged/fudged quite substantially for dynamic subjects.

Like RED it is very important to communicate what is currently unfeasible but yet may be possible in the next few years. That way if we communicate the same to our customers/partners then there is not the expectation that the impossible can be achieved right now. Instead we can steer people/customers to the next best thing , and if that thing happens to be RED ONE then that’s still pretty miraculous. Its good to understand the limitations.

Cheers very much for the input,

Eric

 

[Friedrich Moser]

Amazing! Just amazing! You guys are artists in technology!

Fritz

www.blueandgreen.info

 

[Rob Castiglione]

"artists in technology" ... yes, a beautiful description that captures the essence of Red.

The only other company I can think of that openly showed its development process and which ended up with a technological marvel is Sound Devices with their amazing digital recorders.

Rob

 

[Sanjin Jukic]

In no particular order...

Still cameras do not have a global shutter.

If you shoot a non moving object, even a slow read-reset sensor will look great.

There are widely varying degrees of read-reset times for a rolling shutter. Some as fast as a film camera.

A global shutter will "feel" different than a rapid-reset sensor or a film camera when used for motion.

Assuming that we all get used to a global shutter look (less "filmic" and more sterile), technology will take us there is a few years. It has only been very recently that we can have a high speed, high dynamic range, good light sensitivity, and a rapid-reset sensor.

As we move forward, you are likely to see all aspects of the sensor improve. We are spending a ton of time and research money on this. It is one of the main reasons why we believe in a sensor upgrade, starting with the RED ONE, in all our cameras without having to buy a completely new camera every year.

It is impossible to find what you are looking for today. Sometime in the future, when we have delivered all you are asking for today (assuming you really want it once you have it), you will have a whole new set of wishes. That's how this whole deal works.

Why can't car companies just build a sedan that has 600 horsepower and gets 40MPG? Answer: They can. Just not for a few years.

Jim

Thanks Jim.

Amazing!!!

It sounds very promising.

Also with a great hope that the future sensor upgrades with Mysterium-X and Monstro

can solve the problem with shutter speed and flash shown in the picture below:

Shot on RED1 with Canon 50mm FD L f/1.2 @ f/1.2, available light, 4k 2:1, 24 fps, ISO 320, Shutter 1/48, scaled to 1K, no any CC.

 

[Jannard]

Thanks Jim.

Amazing!!!

It sounds very promising.

Also with a great hope that the future sensor upgrades with Mysterium-X and Monstro

can solve the problem with shutter speed and flash shown in the picture below:

Shot on RED1 with Canon 50mm FD L f/1.2 @ f/1.2, available light, 4k 2:1, 24 fps, ISO 320, Shutter 1/48, scaled to 1K, no any CC.

We'll probably solve this as quickly as a film camera does. This happens with any camera that has a mechanical shutter or rolling shutter with un-sync'd flash.

No matter what you accomplish with technology, some other problem crops up. We all wish for cameras to act like our eyes... oh, no we don't. It takes our eyes several minutes to adapt from bright sun to a dim lit room. A camera will do it as fast as you can spin the aperture or ISO dial. That means... our eyes suck?

People tend to get fixated on a specific issue (like split frame with un-sync'd flash). The obvious solution, in the minds of many, is a global shutter. OK. Here is your global shutter. But you can't shoot high-res at high frame rates. And "can't we get that motion to look better... you know, more like a film camera or a rapid-reset rolling shutter?". You get the idea...

Jim

 

[Eric Lange]

We'll probably solve this as quickly as a film camera does. This happens with any camera that has a mechanical shutter or rolling shutter with un-sync'd flash.

No matter what you accomplish with technology, some other problem crops up. We all wish for cameras to act like our eyes... oh, no we don't. It takes our eyes several minutes to adapt from bright sun to a dim lit room. A camera will do it as fast as you can spin the aperture or ISO dial. That means... our eyes suck?

People tend to get fixated on a specific issue (like split frame with un-sync'd flash). The obvious solution, in the minds of many, is a global shutter. OK. Here is your global shutter. But you can't shoot high-res at high frame rates. And "can't we get that motion to look better... you know, more like a film camera or a rapid-reset rolling shutter?". You get the idea...

Jim

Even if a special technical/FX imaging camera is not in the works or on the cards (with global shutter and RAW [ha. ha]) I think most techies and tech firms would rather knock on the door of RED to seek a technical solution than knock on the door of DALSA (not that there is anything wrong with DALSA…); and one usually does not knock on the door of SONY unless you have a very astute legal team for IP and licensing.

Besides I think you have gone a very long way to almost completely/completely solving the “high-res at high frame rates“ and the "can't we get that motion to look better... you know, more like a film camera or a rapid-reset rolling shutter?" thing.

I hope you take it as compliment and terrific achievement that the best in their respective technical fields look to RED for REAL solutions. Three years ago when you started RED did you really have a sense that you would arrive at this point as well as you have, or do you have to pinch yourself now and again to suspend disbelief on such an amazing accomplishment?

Happy third anniversary.

Very Best,

Eric

 

[Roberto Lequeux]

I'll take the flash issue without complaining, especially if film cameras have it as well... I wouldn't presume film cameras could be improved upon, that would be a sacrilege. Except of course, when that is exactly what happens time and time again, to the benefit of all of us in need of a good camera at a more than fair price. Of course, Eric, you didn't complain that is certainly not what I am saying here.
What I did thought I'd mention was that I'd gladly be OK with Jim and the guys forgetting about the flash issue for now and used their effort to somehow squeeze more dynamic range out of their new sensors... not to sound greedy either

But perhaps a simple solution could happen with firmware, if Red allowed sync with every single camera and flash out there that are sync capable, then instead of firing when you press the button it merely needs to wait to the next hole 24th of a second when your favorite Red camera is going to be recording the next frame. Not a big deal, and as good of a fix as you might be able to get for a while... no?

 

[Eric Lange]

I'll take the flash issue without complaining, especially if film cameras have it as well... I wouldn't presume film cameras could be improved upon, that would be a sacrilege. Except of course, when that is exactly what happens time and time again, to the benefit of all of us in need of a good camera at a more than fair price. Of course, Eric, you didn't complain that is certainly not what I am saying here.
What I did thought I'd mention was that I'd gladly be OK with Jim and the guys forgetting about the flash issue for now and used their effort to somehow squeeze more dynamic range out of their new sensors... not to sound greedy either

But perhaps a simple solution could happen with firmware, if Red allowed sync with every single camera and flash out there that are sync capable, then instead of firing when you press the button it merely needs to wait to the next hole 24th of a second when your favorite Red camera is going to be recording the next frame. Not a big deal, and as good of a fix as you might be able to get for a while... no?

I’m not the guy that has a bee in his bonnet about flash, that’s Sanjin, I was trying to bring a little perspective to the discussion and stick up for Sanjin a little (not that he needs it). I’m the crazy global shutter/Quasi RAW guy, as a high quality FX camera with true electronic global shutter is desperately needed for our kind of 3d work. Such a camera does not have to come from RED of course, but it sure would solve a lot of problems for us. Jim thought that such a camera would be possible (in a few years) but is in essence counter to the whole architecture and concept of the RED/REDCODE system.

As what you are suggesting is it possible to synch a strobe to a RED One so that you have complete illumination of the frame (I thought this was impossible)? I was surprised on Sanjin’s photo he posted that you could get 80% coverage from a flash bulb? I guess I'm not thinking still versus dynamic properly (dynamic sequnces to "freeze capture" motion using flash strobes vs a standard "sloppy" flash unit)

 

[Joe G.]

A while back there was a thread that showed a slower timed electronic flash unit (adjustable) that would be pretty useful, going across several frames. Has anyone experimented with this, and can they show results?

 

[Jannard]

What is on our list is to continue to push for more dynamic range (hence the 16-bit A/D in Monstro) and to give faster read-reset times... as we are doing in both Mysterium-X and Monstro. Technology only goes so fast. Sensor development is painful. Very time consuming and expensive. You might not know how long it takes to make a test wafer. Ask one of your silicon buddies.

We have a pretty comprehensive roadmap of sensor development at a time when others are cutting budgets. Our 1st effort was pretty amazing compared to anything done before. And we really didn't even know what we were doing!

Give us a bit of slack. We don't know of any other sensor alternatives that come close. And we have the most aggressive program out there for resolution, sensitivity, dynamic range, and speed. We know what we need to do. Better than you think.

Jim

 

[Eric Lange]

What is on our list is to continue to push for more dynamic range (hence the 16-bit A/D in Monstro) and to give faster read-reset times... as we are doing in both Mysterium-X and Monstro. Technology only goes so fast. Sensor development is painful. Very time consuming and expensive. You might not know how long it takes to make a test wafer. Ask one of your silicon buddies.

We have a pretty comprehensive roadmap of sensor development at a time when others are cutting budgets. Our 1st effort was pretty amazing compared to anything done before. And we really didn't even know what we were doing!

Give us a bit of slack. We don't know of any other sensor alternatives that come close. And we have the most aggressive program out there for resolution, sensitivity, dynamic range, and speed. We know what we need to do. Better than you think.

Jim

As with any complex system, there are always trade offs. As you say one has to be practical and the notion of the rather Germanic “uncompromising design” can lead to real disasters or a system that is over designed can all to easily fail as a product. Sometimes you have to be ruthless about what goes into the mix and what does not. I believe good design is always a compromise. Generally we try engineer the most precise systems we can first and then see how we can cleverly down grade to achieve 89% of the same performance at a fraction of the price/effort. However I am constantly reminded to strive for excellence not perfection as this can be the road to failure, (I am reminded of John Harrison and the Longitude problem ). There is no doubt in my mind that RED is very practical, and I find it very amusing that other companies make rather spurious arguments as to why no one should ever need RED rather than figuring out how to keep up.

Jim do you have any sense as to whether various competitors are going to try and develop rival systems or do you feel that RED has the field to its self?

What advice would you give anyone who is contemplating building their own camera system?

P.S. I think of the RED one very much in the same way as the Bell-X1in terms of breaking the sound barrier, and the EPIC as the SR-71. I think some people understand RED better than you think, and still remain absolutely blown away; as Arthur C Clarke wrote (misquote) “any sufficiently advanced technology is indistinguishable from magic”, and I think people see JIM et al. as the wizards, that’s why people ask for magical things.

 

[Jannard]

Jim do you have any sense as to whether various competitors are going to try and develop rival systems or do you feel that RED has the field to its self?

What advice would you give anyone who is contemplating building their own camera system?

No idea what they (competition) are doing... I hear that many are cutting development budgets, but we act as if everyone is coming at us full steam. Plus, we are motivated to move as much and as fast as we can by someone other than out competitors... us.

What advice would you give anyone who is contemplating building their own camera system? Good luck with that...

Jim