Breakthroughs...

[Peter Majtan]

Any idea on pricing?

Mine, or retail? :devil:

 

[David Rasberry]

"estimated street price":matrix:

 

[Jason Badham]

Here's your black silicone right here.

I have been looking all over for that... thanks Jeff!

That was pretty funny...

 

[Austin Glass]

I think they invented a vector-based camera. Yes, that's right - A camera that takes images based on infinitely small points and creates a unique mathematical equation based around those points, thus doing away with this whole rasterized pixel nonsense we've been toying with all this time. Infinite zooming in post.

You laugh now, but I shall be the one laughing next week! muwahahahah!

 

[Tim Hole]

Oh thank you for being so insulting. Please excuse my relatively nonexistent knowledge of sensor technology.

Pray tell: when is your new sensor coming out Timmy?

Edit: I think I got the wrong end of the stick there. My bad

He He!! Yeah you got the wrong end of the stick. Thanks for the insult back though it was a good one!!!

 

[Kwan Khan]

yeooop'
any update?

 

[amateur6]

Here's your black silicone right here.

As for black silicon, I think we can only speculate at this point.

Thanks to UCLA, black silicon may be obsolete even before it hits the market... but someone who actually knows something about sensor technology will have to confirm that for me:

http://www.wired.com/science/discoveries/multimedia/2008/12/gallery_photodetector

(and I know, I know -- if WIRED knows about it, it can't exactly be news, but...)

 

[Eric Lange]

GLOBAL SHUTTER (I hope).

The only thing at the moment that holds the RED/EPIC/SCARLET systems back is the absence of a true electronic/digital GOLBAL SHUTTER. The twin rolling shutter of RED One’s is impressive, however what is still desperately needed for advanced 3d capture and VFX work, is a true global shutter. [This would also significantly help stills photography as well]. Some industrial cameras have global shutters and produce sequences that are true frozen motion per frame, but unfortunately most industrial cameras produce cheesy looking images (noise, low dynamic range, bad color and tables, and generally not portable). The holy grail for the technical and FX community would be cinematic quality high res images that are true global shutter; Skew completely eliminated.

That would be doable, but never the less would be a major breakthrough

 

[Radoslav Karapetkov]

Thanks to UCLA, black silicon may be obsolete even before it hits the market... but someone who actually knows something about sensor technology will have to confirm that for me:

http://www.wired.com/science/discoveries/multimedia/2008/12/gallery_photodetector

(and I know, I know -- if WIRED knows about it, it can't exactly be news, but...)

"polymer film"


Ha, seems that film is going to survive after all. :bleh:

 

[Eddy Robinson]

I think they invented a vector-based camera. Yes, that's right - A camera that takes images based on infinitely small points and creates a unique mathematical equation based around those points, thus doing away with this whole rasterized pixel nonsense we've been toying with all this time. Infinite zooming in post.

You laugh now, but I shall be the one laughing next week! muwahahahah!

http://en.wikipedia.org/wiki/Fractal_compression

I got to play with some of this technology back in the early 90s, though it never really took off properly. Which is a pity, because ~15 years later, I still remember how incredibly impressive the results were. Unfortunately a lot of patents are still tied up.

 

[Zach Swena]

notice on the wiki page that it calls fractal compression a lossy compression. Red Code is a not lossy compression. Don't think they match.

Might be cool to have if the image quality was sufficient and small space was critical. Assuming that it could be proved to be worth while to R&D

 

[Michael Hastings]

notice on the wiki page that it calls fractal compression a lossy compression. Red Code is a not lossy compression. Don't think they match.

Might be cool to have if the image quality was sufficient and small space was critical. Assuming that it could be proved to be worth while to R&D

I believe REDCODE is also a "lossy compression". It is claimed to be "visually lossless" but does in fact lose actual data.

Lossless compression is that which can completely reproduce all of the original bits of data. It basically uses the redundancy to create shorter descriptors of blocks of data that have the exact same value, but only to the level that it can completely reproduce the original. Obviously that is an oversimplification of some very sophisticated algorithms in use today - but it is the basic principles. Presumably "visually lossless" would go as far as it could using lossless techniques, then throw out actual data but by using knowledge of the structure of normal motion pictures, it would first throw out the data that would be noticed the least.

 

[albert rudnicki]

I Remember the Matrox full digisuite calling their boards mathematically lossless.
I think in reality there is no such a thing as lossless compression.

 

[Jeff Carpenter]

here's a breakthrough...

here's a breakthrough...

here's a breakthrough...

but one we won't see in commercial cameras for several years.

sounds cool though

http://www.wired.com/science/discoveries/multimedia/2008/12/gallery_photodetector

from the page:

Thanks to this lucky mistake, a new breed of camera sensors that are cheaper, higher-resolution and have lower distortion could be on the horizon.

FYI here is the paper in Nature:

http://www.nature.com/nnano/journal/v3/n9/abs/nnano.2008.206.html

The correct, full content publication is as follows:

http://164.67.192.163:81/papers/nnano.2008.206.pdf

P3HTCBM doped with CdTe has QE (Quantum Efficiency) of 8000% at 350mm! Again, CMOS and CCD QE is usually around 50%, but there are ways to get it up to 90%.

 

[Teague Kennedy]

Love it Jim. Waiting with eager ears. I don't mind paying for the best, but please don't forget the independent..... I like to think if you sell it to the low budget independent, you will sell a heck of a lot more. Nobody wants you to undersell yourself, but we all want to be able to buy it -- not rent it..... Remember the little man, and get rich doing it.... That's my motto.

 

[Steven Caesare]

I Remember the Matrox full digisuite calling their boards mathematically lossless.
I think in reality there is no such a thing as lossless compression.

Accepted meaning of the terms-

Mathematically lossless: bit-for-bit reproduction of the originial after compression/decompression. This is the classic ".zip" file.

Visually lossless (aka perceptually lossless): Compression applied to image exploiting certain characteristics of the human visual system such that the image appears to be (practically) identical to the original, despite not being a bit-perfect reproduction. REDCODE is considered in this category, as can be AVC/VC1 at sufficient bitrates.

Lossy: Compression wherein visual fidelity is compromised in order to effect greater sompression ratio. Examples: low-bitrate MPEG, highly compressed .JPG files, etc...

-Steve

 

[oldphart]

http://en.wikipedia.org/wiki/Fractal_compression

I got to play with some of this technology back in the early 90s, though it never really took off properly. Which is a pity, because ~15 years later, I still remember how incredibly impressive the results were. Unfortunately a lot of patents are still tied up.

The problem with fractal compression is that you need a good mathematician tweaking it manually to compress even a single frame. They never succeeded to automate the process, and it may actually be impossible.

Fractal compression is not lossy in the way other compression is. It produces a highly detailed image which is more or less subtly different from the original, including detail that was not there in the first place. hmy:

 

[Mark Petereit]

We had several nice breakthroughs today that will translate into Scarlet and EPIC. Details shortly...

As we stated... drastic changes, expect it and you won't be disappointed.

Jim

Hmmm. Have you been hanging out at UCLA lately? :detective2:

 

[Eryc Tramonn]

The problem with fractal compression is that you need a good mathematician tweaking it manually to compress even a single frame. They never succeeded to automate the process, and it may actually be impossible.

Fractal compression is not lossy in the way other compression is. It produces a highly detailed image which is more or less subtly different from the original, including detail that was not there in the first place. hmy:

WTH? It's capable of producing details that were not there in the first place? :help:

 

[Gavin Greenwalt]

WTH? It's capable of producing details that were not there in the first place? :help:

Sure. You can do this easily in photoshop right now. Load an image. Save it out as a JPG. Then reload it and apply 10x noise. The details is now pixel for pixel. It's just not the right detail.

Or you could trace a person walking along a river. That black and white image may be far 'sharper' than the original. It's just not a funcitonal image.

Detail is frequency. Add noise to any signal and you can increase the frequency. Fractals try to extrapolate the 'signal' noise and all. So it adds in visual details that aren't necessarily 'motivated' by the original.