Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Jim and CO. take note...

O

Obin Olson

Well-known member
Joined
Dec 28, 2006
Messages
2,004
Reaction score
0
Points
0
Website
www.dv3productions.com
This company, if they have the goods hardware wise needs FULL red support ASAP:

http://www.pixellexis.com/

just got off the phone with the owner and what she has built could turn the post industry upside down on the magnitude of RED Digital Cinema.

Lets give her the support she needs to make 4k online from capture to finish a reality. I think her hardware can do it, she just needs programming support for post software plugins etc.

Think about this: realtime transcode of 4k from red to jpg2000, realtime editing of 4k jpg2000, realtime color of jpg2000, realtime rendering of jpg2000, realtime 3d animation rendering etc......as if you had a render farm of 500 machines!

All it's going to take is software to make this work. She wants to know what WE want, lets start the support!
 
They will, and will use it. What I am saying is this is HUGE for POST, as most of you know post is lots of render time...every day, all day long, if this system can get rid of some or most of that..........:)
 
Red takes note because they need to make sure whatever they are doing they work with them. I'ts going to be lots of outside sources that are going to write code for this new device.
 
Great News
 
CineForm also gets rid of almost all the current render time. With build 15 CineForm can transcode RedCode Raw to CineForm Raw at about 12 frames per second. That means one hour of footage is ready to edit in two hours. When CineForm gets the SDK, we can all do this with an eight core PC or Mac. Color correction can be done with 3dLUTS, all non destructive. But who knows, this new processor situation you're talking about might even make the CineForm thing all real time. Lots of great things coming.
 
I told them I would order one last week if it worked! ;)
 
Ummm just a quick question?

Is proccessing power really the bottleneck? I wouldve thought that hard drive speed would be. Once you have an octocore (and once all the cores are used) normally i would think that the CPU could shovel through much more data than the memory can give to it?

Somebody enlighten me if this isnt true...
 
debayering is CPU job. But instead of thinking transcoding, i would be more interested to see it playback R3D 4K full quality realtime. Most people/some people dont do transcoding to jpeg2000 to retain flexibility on metadata.

I see the usage in DI system more. I would love to see it implemented on Assimilate Scratch mainly and RED.
 
I understand that the Debayering process is the CPU's job....

But the information to do the debayer has to come from somewhere, and surely that would still be the bottle neck for this kinda stuff, not how much CPU power you have. Unless you could load it all into ram, even then i think you would be struggling to saturate a CPU.
 
obin said:
All it's going to take is software to make this work.
Click to expand...

My favorite quote of the week. :)

Walk yourself down the path. What DO you want? My guess is that you want the Redbrix to support R3D rendering and accelereate it. Ok... but through what application?

And there lies the rub.

Pixellexis is not going to do the work to support every application under the sun. They can't, and plus - they are obviously a hardware company. It's not their business or their core competency to make that happen. That means they need applications to make their solutions useful. I'm sure they have a well-developed API so that any current software manufacturer can take advantage of their acceleration hardware.

Now you have to go to the individual software companies and convince them that spending significant amounts of time, energy, and man-hours to code to their API makes sense. Programmers are extremely expensive resources. A software company would have to feel like the Pixellexis technology was fairly core to their future success and would be something that a LOT of their customers would want (and buy) in order to dedicate those resources.

Redbrix is $10,995. Redcard is $2,995. Redcard looks pretty cool. How much Redcard would actually accelerate... what? Rendering? Is that the primary concern here? My guess is that Redcard would do a pretty good job, but you would really need Redbrix to make high-rez rendering realtime or pretty close to realtime.

It's $10,995. For that price, how many blade servers could you buy to set up a Redline render farm? Maybe 2 or 3? And that would increase your rendering time so that it was maybe 2-to-1 or maybe 3-to-1. Redline is available now. And it works. Forget about Quicktime here, by the way. Quicktime rendering is largely limited by the Quicktime API, which is not very optimized for speed.

So... is getting that last bit of rendering performance to some file format other than Quicktime worth the $10,995? Would enough people buy it for an extra 11K to make it worth the time for a software manufacturer to code to their API and restructure their rendering pipeline in such a way?

Not trying to pee on your Cheerios here - just walking you down the road of how this actually happens and the real-world mechanics that would have to be put in place for Pixellexis technology to become part of a RED pipeline.

Best,

Lucas
-----
ASSIMILATE, inc.
LA, CA, USA
 
Yeah I appreciate the enthusiasm and all but other than the fact that these machines happen to also be called Red I really don't see why Jim & Co should take note. Really the other way around considering this is the first time I've even heard of Pixellexis. Just don't see how adding in a whole new system architecture when we have a developing workflow for Mac and PC is lagging behind. There's Linux boxes out there too with massive parallel processing but it's the same issue- too niche to make it profitable or well supported for the consumer. Stay the course....

Noah
 
You must log in or register to reply here.
Back
Top