In development...

[Copernicus]

I'd love to see tests on the original sensor data and the Rfiles comparison. Then the value of what you are saying can be taken into consideration.

I fully agree and such a comparison is on my wish list.

 

[number6]

Number6. What RED has achieved is really impressive, I'm a huge fan. I only give out my thoughts here, hopefully in a constructive manner.

Regarding your statement, I´m not so sure of 95%. An example, for 4K RAW with e.g 12 bit depth you have an image size of approx. 12 MB uncompressed, i.e. 3 GB for each 10 second sequence. Memory isn´t so expensive and remember you only store that once! Then you can pick it up in your favorite postprocessing tool, set workspace to 16 bit color depth if you want, do your final touch and store them in any compressed size, even in uncompressed 48 MB images if that´s what you need (of course then 10 sec. equals 12 GB). If you start with an already compressed original, my belief is that you loose some crucial freedom to operate in you workflow. Thats my belief in theory, I have not done a comparison. Yet...

Copernicus, not saying what you suggest isn't somehow better, but it's just that having REDCODE making a more manageable file means I can get by with a less expensive and less complicated hardware and software workflow.

For the pro-iest of pros... you know, the REDi Masters... I'm sure the more control over the entire process the merrier. For some of us, well, let REDCODE do the heavy lifting before times. Again, just a different take on things. There is no right or wrong position AFAIK.

 

[Blair S. Paulsen]

The sweet spot

The sweet spot

No format can do everything. RedCode RAW comes as close as anything on the market. RedCode is well matched to the existing and near term performance capabilities of computer based post production hardware and software. Completely uncompressed formats are great but expensive to handle. HDCAM-SR is a robust and lightly compressed format that isn't popular just because of the Sony hegemony, but because it looks damn good - unfortunately the decks are very pricey.

My point is that 4K RedCode RAW treated with care can hold its own in big theatrical presentations. With enough time and patience even a desktop based DIY workflow can achieve stunning results. With a Scratch system and high end talent you can get there substantially faster.

A few years down the line the parameters of data based workflows will likely be substantially different, but for now, IMHO, Mark is right, RedCode is an under appreciated leveraging technology in the Rebellion.

Note to Bruce and others pimping for real time output at 1080P - it is my guess that once the internal operational particulars of the algorithms and such of RedCode are more locked down we will see a one-trick pony board built that will do full quality de-bayer on the fly. Just a guess.

 

[Bruce Allen]

Note to Bruce and others pimping for real time output at 1080P - it is my guess that once the internal operational particulars of the algorithms and such of RedCode are more locked down we will see a one-trick pony board built that will do full quality de-bayer on the fly. Just a guess.

Of course. But by that time, Epic will be out.

All I am saying is that perhaps just a nice 4:4:4 HDCAM-like codec that gave proper 1080P output from the get-go would have been a perfectly acceptable choice for Red One.

All of those Red owners could have made more money on their rentals and hence there would have been more people able to buy an Epic.

With the extra cash from extra Epic orders, Red could perhaps have licensed Cineform RAW and still come out ahead. No REDCODE needed.

BTW, just to justify my earlier statements that REDCODE likely didn't come first, if you read the blog for the film Spoon (which I read at the time and then promptly forgot about , you'll notice that they were already shooting a feature using the Silicon Imaging camera and compressed Cineform RAW back in May 2006.

http://indiefilmlive.blogspot.com/2006_05_01_archive.html

Bruce Allen
www.boacinema.com

 

[Lawrence Bansbach]

The faster Scarlet's lens gets the better... because we get closer to 35mm DOF.

Not to mention needing less light at a given ISO rating.

 

[Lauri Kettunen]

compressed Cineform RAW back in May 2006

In fact, the historical explanation of Cineform Raw can be found in David Newman's blog, see

http://cineform.blogspot.com/search...-max=2007-01-01T00:00:00-08:00&max-results=22

and the stuff "CineForm RAW™ -- this skunkworks project becomes a reality" he posted April 16, 2006.

What often happens in science and engineering is that people invent the same idea about the same time independently of each other. Then, everybody feels rightly it's their own invention and it's just another matter of fact who can be shown to be there the first one.

 

[Esteban Sosnitsky]

I think the 1080p is not a matter of time nor a priority.
My Guess is that its just not possible because the camera doesnt have such a processor inside and was not conceived in that way either. It would require a brand new board that could process the raw data and convert it into a processed signal; which I think its pointless because you would be trying to convert the red into a regular HD cam. I think the cam is great as it is, it makes you look for your own resources and re think creativity. I like this challenge very much with all and its limitations, specially due the fact that it makes you re think and be creative=talent.
The video outputs are a video assist, thats it. Read the leonetti's interview, very interesting.

 

[Even Solberg]

RED would be better producing something to compete with the medium format digital market (eg 40-60MP Hassys) at DSLR prices. That is where prices can be cut by large amounts. All RED need is a full frame 120 sensor...

This would have me very, VERY interested. To the point where I'd take a long, hard look at my $30k worth of Canon gear. Especially if I could use it in harsh environments such as the arctic.

 

[Eric Lange]

A Very high res still shooter with high res video cannot come soon enough!

We are currently working on what can best described as an Inverse 3d animation system using very precise stereo cameras. The software we develop, very efficiently and cleanly extracts very accurate correctly texture mapped 3d models as well as a new type of VR model that is 400x more efficient than conventional techniques.

We are currently experimenting with high end medium format backs (such as Leaf) and are starting to experiment with RED ONE as well as some rather lack luster industrial camera systems.

What we really need is system that can capture high res stills at a rate of 2fps (to capture high res geometry) and then to have interleaved runs of low res 30-60fps to provide motion capture data. That way the high res stills provide high fidelity 3d key-frames and geometry and the low res video directs the character movement and rigging.

It is very frustrating for us as we feel we are being held back by rather slow thinking from the DSLR industry and to an extent by film/digital industries (apart from RED ONE of course). We know that all of the hardware that we are currently using (except the RED ONE) will be obsolete, expensive and inefficient if RED reaches the goal of medium format res with lower res video capability and would greatly reduce the amount of instrumentation and equipment to achieve effective 3d live capture.

RED please do not (I am begging now) use a mirror or reflex in such a camera. The reason is that a mirror or reflex prevents the use of non-retrofocal lenses. We very much would like to stick with BIOGON type lenses as these have much lower distortion at wide angles than retrofocal lenses. This is good for 3d stereo cinema and good for 3d inverse animation. Lenses such as ZM (range finder type lenses) and Schneider (symetric design) lenses are far superior for our applications. [Think RED RANEGFINDER instead of DSLR].

In terms of functionality and form factor you could have a very large eye cup, an OLED display with diopter sandwiched onto it (much like a head mounted display). The camera could be cylindrical in shape, fit steady to the eye leaving the other eye free to view the scene like a leica rangefinder. Cable release and control could be long and be in the left hand. Biogon type optics would make this very high quality and compact. I personally think the design and formfactor of conventional DSLR's to be reminiscent of victorian jelly moulds. They are bulky, inefficient and greatly limit the optics that can be used. RED, please deliver us from mediocre and stayed camera design.

Yours hopefully,

Eric.

 

[Peter Majtan]

Not sure if this has been mentioned, but in my opinion the key to succes for such a "DSLR-like" camera would be unique and the first thing that comes to my mind is to provide interchangeable lens mount. This might sound silly for DSLR, but You have to consider that most pro still photographers (like myself) have a ton of investment in a great lenses. The thing we keep buying almost every year or two is a new body. At the very least it should offer both Nikon and Canon mounts. Combine that with something like the Birger brain and You will have one hot camera to look for...

Just my one peso...

 

[Peter Majtan]

One thing to keep in mind is RAW. RED is not likely to steer away from RAW acquisition. But RAW has one important limitation - it can't be scaled down to provide lower resolution without either converting to RGB (or another non-RAW image format), or cropping the actual sensor - reading out only smaller centralized portion of the sensor. This inadvertently result in loss of wide angles through the associated crop-factor. There are techniques that allow for reading every other n-pixel to get 1/n resolution, but this brings specific problems and image degradation via strong aliasing artifacts...

To summarize this - I just don't see how could You create a RAW camera capable of recording super-high resolution still and lower resolution motion pictures unless You put in two independent sensors, each dedicated to specific function. This will however result in significant light loss...

Just my one peso...

 

[Eric Lange]

Peter,

I agree, "but in my opinion the key to succes for such a "DSLR-like" camera would be unique and the first thing that comes to my mind is to provide interchangeable lens mount."

We spend a huge amount of time redesigning our systems around different lens mounts, this becomes expensive and very redundant. I wish there was a more open and basic mount that allows one to get closer to the sensor, and therefore would be much more flexible for engineering purposes.

 

[Eric Lange]

If you look at the transistor arrangement on CMOS sensors almost anything is possible. Not so for CCDs. As a programmer I know that it would be possible, but rather inelegant perhaps compared to a straight ahead flow of RED RAW. Its definitely something worth going for and we would love to get rid of the redundancy of two different camera systems. This would reduce our rig from 16 cameras to eight for full 360 degree 3d capture.

We can hope...

 

[Eric Lange]

Nikon D-90 aside, in terms of higher res video from a DSLR, the SONY Alpha A-900 (24 MP sensor) seems set for launch for photokina 2008, sept 23-28. Sony are keen to push the idea that HD is in their "HDNA" (puke..) but I wonder if the A-900 will produce any decent or usable HD output?

Zeiss seems to be poised to launch some new primes for that camera at very least a 35mm ZF type lens. Many folks seem to think even five minutes of HD video from an SLR is worth while, and certainly very useful for the kind of work we do.

check out http://www.photographybay.com/2008/04/15/photokina-2008/
for info on the A-900

also take a peek at http://www.roumazeilles.net/news/en/wordpress/2008/04/05/sony-alpha-900-first-analysis/

Those of you that like the idea of Zeiss glass on Canon equipment might be interested in the new ZE range of lenses. See http://www.zeiss.com/photo So far the most plausible explanation of ZE is Zeiss EOS, but you never know...

 

[Rocco Schult]

Peter,

I agree, "but in my opinion the key to succes for such a "DSLR-like" camera would be unique and the first thing that comes to my mind is to provide interchangeable lens mount."

We spend a huge amount of time redesigning our systems around different lens mounts, this becomes expensive and very redundant. I wish there was a more open and basic mount that allows one to get closer to the sensor, and therefore would be much more flexible for engineering purposes.

I hear P&S IMS again and again...
Veery open, it is meant to be adopted.
And the adapters are neither complex nor expensive.

 

[ThydNostrum]

RED DSLR: DSLR form factor, medium format full size sensor, Interchangeable lens, Live view, 5 min video capability, 5" LCD, body only for $1750.
Now, that will revolutionaize the Still photography market.

 

[Eric Lange]

I hear P&S IMS again and again...
Veery open, it is meant to be adopted.
And the adapters are neither complex nor expensive.

Thanks for the tip Therock, I'll check it out.

Generally we have to engineer special mounts that allow a lens to be adjusted to within 3 arc seconds of squareness to the CCD for 3d and photogrammetric work using an autocollimator. We like mechanical and thermal stability and symetry in the design. With proprietary lens mounts you only have a millimeter or two (at most) to effect mechanical adjustments. Something that would give us four or five mm depth of an envelope to play with would be fantastic and make our lives much easier.

 

[Eric Lange]

RED DSLR: DSLR form factor, medium format full size sensor, Interchangeable lens, Live view, 5 min video capability, 5" LCD, body only for $1750.
Now, that will revolutionaize the Still photography market.

Possibly...

However to design and manufacture the shape of a classic DSL is frankly a pain in the bum. Like the other RED designs, simplicity is the key for efficient relatively low scale production that still looks good. Even though they use 5 axis mills and ProE CAD/CAM software a simple design is much better for the bottom line.

I think the camera will look like the offspring of an old Kodak 16 MP back that has had sex with the RED EPIC. The camera unit itself will essentially be square or cube block, measuring 3.5" x 3.5" 2.0", but will be mountable on rails so that you can use long lenses, matt boxes and filters etc. So it will look like an exact cross of a medium format stills camera back and a small studio camera. Might be cute, as you could pull focus and play like a movie camera. Price $4k , + loads of accessories to really go to town on.

maybee...:usd:

 

[Cory Mitchell]

Eric/VERAXYZ it sounds like you guys are working on some pretty awesome stuff. Even if it could be used to extract keys without manual roto or greenscreen, that would be awesome.
Even with the newer movies, greenscreen scenes pop out like sore thumbs (and with some movies that means a lot of sore thumbs).

Good luck with your work!

As always, I'm very interested with the Red DSLR. Because of the extremely low price of Scarlet (3k and 120fps for $3k wth?!) I really wouldn't be surprised to see their price for the DSLR to be amazingly low (relatively speaking to other PRO DSLR's or maybe DSLR's in general). Keep up the great work guys (and girls)!

 

[Eric Lange]

Eric/VERAXYZ it sounds like you guys are working on some pretty awesome stuff. Even if it could be used to extract keys without manual roto or greenscreen, that would be awesome.
Even with the newer movies, greenscreen scenes pop out like sore thumbs (and with some movies that means a lot of sore thumbs).

When we go for launch of our software (Q1 2009) I hope others will think it awesome too. I guess vision mind is good handle for you, as lot of people wouldn't figure out that full 3d capture has the capability/potential to mitigate green screen errors because of the angles involved. Very good thought indeed, our software has a lot of applications, we hope we find the right markets before we loose the farm as a five year SW and hardware development program is expensive!

"green screen scenes pop out like sore thumbs". I think in SIN CITY and the 300 they did this funny thing where all of the characters hair was lit like an atomic explosion, so every one had their own gigantic radiant halo. An interesting stylistic adaptation to a very difficult technical problem. Hair is tricky in 3d and in stereo too, but we also have some funny tricks up our sleeves.

Cheers,