I think I get The Goal of Red

Double read-out as applied to video images is not a double (or multiple) exposure and bears no resemblance to the primitive process you describe for still photography except in the ultimate result of extending dynamic range.
 
Greg, you'll have to explain double readout to me better then. Do you mean doing multiple readouts during the exposure cycle and post-integrating them? You do the readout hopefully before the bucket gets filled and then add together all those buckets? That certainly can be done, but it's not without it's problems either.

Graeme
 
I read ARRI do a double readout on their D20, halving their max frame rate in the process. Seems that you let some pixels integrate for the full exposure cycle and some for less time, then you combine the info together afterwards. Anyone actually seen this work in practise? Again, I can't see how this is meant to work on moving images due to reduced integration time being equivalent to increased shutter speed, and that produces a strong visual effect on motion.
 
Control...

Control...

As just one example, if you read the article on how 24 is testing (and already utilizing to some extent) 1/3" HD cameras within their regular production technique, you would know that the crossover has been underway for some time.

Independent filmmakers have always pushed whatever technology is available to them to find a place at the table, whether it be George Romero with 16mm news cameras or someone here with their DVX.

RED, depending on its price, will simply give content creators another acquisition option.

In the end, it isn't the camera that makes a filmmaker a filmmaker. Story is, and always has been, king. Filmmaking is simply the craft that services the creative. Certainly, the current crop of tools is an equalizer of sorts.

I would also state that RED, Genesis, D20, HVX, DVX and so on have very, very little to do with wresting filmmaking from the majors.

Radically changing existing distribution channels, where content is controlled exclusively by the creator, is the ultimate answer to that question. The music industry has been turned entirely upside-down in less than a decade by free-trade and third-party distribution scenarios - filmmaking isn't far behind.

e
 
There are some CCDs that have large sensors for the main image and small ones to capture highlight detail. The extended dynamic range is VERY impressive but only in RAW mode. (You get a stop or two at best in JPEG, far more in RAW.)

But it takes like 10 seconds to processes a single frame. IMO, having a higher bit depth A/D and a DSP which crushes the highlights into a smooth knee would be preferable and so much simpler for video.
 
Fuji do that small and large pixel thing, don't they? Didn't realise it took ages to process in camera though. They also sacrifice resolution to get it.http://www.dpreview.com/reviews/fujifilms3pro/page18.asp shows the extra dynamic range they can get, along with the 13 second RAW write time. http://www.dpreview.com/reviews/fujifilms3pro/page21.asp shows the trade off, that the Fugi's 12mp produce an image with lower resolution than Canon's 8mp. Again, my comment is that you don't get more than a pint from a pint pot. You're going to sacrifice resolution, or noise, or temporal resolution to get more dynamic range.

At least super-sampling will lower noise, and that's extending dynamic range at the shadow end of things.

Graeme
 
Graeme, the conceptual difficulty you're having (or had) is by confusing the capture/exposure function with the image processing function. The double read-out (clearly not an engineering term) is an image processing function, hence the trade-off in max frame rate. In the embodiment I know about it's still being developed but it's attainable without the introduction of temporal artifacts, and even the max speed compromise is ultimately soluable. I made no claims about immediate availablity but used this as an example of one the things that the flexible architecture of CMOS sensors makes possible. I do suspect that other innovations will negate the need for a double read-out technique.
 
Greg, perhaps I am getting confused. I've heard a few sensor experts talk about methods of dynamic range enhancement, and I thought the double readout you mentioned sounded like one of those I'd heard mentioned. AFAIK, if the sensor doesn't capture it, there's nothing you can do in post as an image processing function, because when those photosites saturate, any more photons that try to get in don't register and the highlights clip. The only ways to stop that are to do partial readouts to stop the photosites filling up, or to make the microlens smaller so less light is captured or to make the filter darker. All of these methods have been tried with some degrees of success, but I see so many more of them being used in stills cameras, not video cameras. For instance, the Fuji solution I pointed out above certainly works, but looses a lot of resolution, and doesn't help with noise at all, and indeed, has interpolation artifacts. Often what's acceptable in stills is unnaceptable in video due to the temporal dimension.

The only mentions I could find about double readout mentioned differing integration times on pixels, which leads me to believe it is an engineering thing in the sensor and happens during capture / exposure, and also needs some significant horse-power to do the interpolation to get some more dynamic range from this extra information.

Graeme
 
Graeme, I misspoke somewhat in saying that it's not a capture function. It's not a matter of double exposure per se, but a combination of capture and signal processing. i'll see if I can dig up some reference to clarify.
 
HDkilledFILM. said:
Sick! The red camera is going to be sick. As for the complaints about the HVX, I just saw some footage that 5 years ago I would have thought was film. Yeah the no interchangable lens stinks, but how long before a G35 type addapter gets made? two months? Three? the bottom line is: the DVX100 changed the game for SD, the HVX is going to make HD a reality for the independant film maker on a budget, no more HDV dropped frames. I love it.
Click to expand...

for the record if you get a dropped frame using HDV, you are doing something seriously wrong.
 
MarcusX said:
What do you need to shoot your movie and take it to the big screen?
Click to expand...

You forgot to mention somebody to pick the movie up and distribute it.

That being said, Red sounds like it is going to be a little out of the price range of the person who would use an HVX. Or in 1990, someone who would use Super 8mm or Super VHS.

Of course, I hope not...
 
Jannard said:
what if you could use your B4 or 16mm PL mount lenses with a cropped image size?

Jim
Click to expand...


I would be very happy. It would be nice to put my Optar Illuminas to use on a video camera.
 
God only knows what an "indie" filmmaker" is since if you control the distribution it doesn't really matter what it's shot on, and the studios will control distribution of vehicles with STARS for some time to come.

Be that as it may, an "indie" could get by with an HVX with 2/3" chips and interchangeable lenses. This would wipe out most of the criticism of the HVX and come in at about 6k san lens. The lens could then run you 1k to 40k, depending on what you bought....I'd take this in a NY heartbeat.

Your picture is then probably "good enough" for widescreen distribution, should the one in a billion chance hit that you get distribution.

I think 4k is overkill, 2k is even too much if you can't edit it on your home Apple system...If you can edit 4k, you can rent 35mm or anything you want....

For this reason there almost has to be two versions of Red. A low end version for the "indies" of the world (translated as "I make films with my crew of 10 and eff hollywood") and the 4k version which is designed to attack the Sony 950's of the world.

And FCP will be supporting you, Mr. Jannard? ;-)
 
Last edited:
It seems to me that 2k, then 4k will be on everyone's roadmap as CPU speed and storage increases. Until then, you can use the RED camera as a high speed 1080p or 720p camera that has 35mm DOF. Record to codec, 4:2:2 or 4:4:4

Jim
 
Jannard said:
It seems to me that 2k, then 4k will be on everyone's roadmap as CPU speed and storage increases. Until then, you can use the RED camera as a high speed 1080p or 720p camera that has 35mm DOF. Record to codec, 4:2:2 or 4:4:4

Jim
Click to expand...
This is just outstanding stuff. I'm wiping the drool off my chin right now.:thumbsup:
 
Jannard said:
It seems to me that 2k, then 4k will be on everyone's roadmap as CPU speed and storage increases. Until then, you can use the RED camera as a high speed 1080p or 720p camera that has 35mm DOF. Record to codec, 4:2:2 or 4:4:4

Jim
Click to expand...

I really like this idea, it's definitely forward thinking. But I believe what concerns some people about that idea is that if it's 4K they won't be able to afford it even if they're only doing 1080 at the start with it. Simply because it's more advanced technology that cost more money. But if it can only do 1080 chances are it will be more affordable as we already know 1080 cameras can be produced for under $10,000.
 
What do you mean with up-rez? Pixelshift?? By the way: Don´t you think sooner or later there will emerge a couple of little reddish 1k or 2k cameras out of the digital still camera segment?
 
You must log in or register to reply here.
Back
Top