The "Redcine in camera" Brain

[PaulClements]

Ok so this is probably technically impossible, but if there is even the faintest chance it is doable then it would dismantle every conceivable issue demanding individuals have with post and Red cameras and make the saleability of the cameras as a tool far greater.

What would be great is if there was an optional add-on brain for the DSMC cameras that had the ability to do on-the-fly processing to a host of different filetypes at usable resolutions. Effectively cutting out the Redcine/Redrushes/RedAlert! step.

Before a shoot you'd check to see what NLE will be used and then dial in the file type on camera. So for example you select Avid DNxHD at 720p. The camera then produces the RAW files as per normal and simultaneously creates the DNx files on the convertor brain. These can then be loaded up ready to go in the edit house's Avid. Options for the typical post resolutions such as 2k, 1080p, 720p PAL etc would be welcomed. Having all the various codecs built in such as ProRes, H.264, Tiff or DPX sequences etc etc etc, just as you do with the likes of RedCine would also be needed. Options for scaling and cropping etc (I'm thinking anamorphic lenses here) would also be great to have.

All of this would be contained in a subsidery and optional pro convertor brain that is added on to the back of the Scarlet or Epic brain. It's something I'm sure that many rental houses or larger budget shoots would come to see as an invaluable tool. Whereas users shooting their own short films or lower budget stuff need not go to the expense of purchasing the extra brain and simply render the old fashion way. I don't expect it to necessarily be something cheap but certainly something that in the right places would be invaluable.

Perhaps the adobe support for R3D files might make this seem redundant but at the end of the day there are a tonne of people who will be editing with other NLE's for years to come meaning without something like this or R3D support in every system there will always be a hefty conversion time. Furthermore any modifications in the R3D filetype need only be accompanied by a firmware upgrade to the conversion brain rather than every post application requiring modification to their code to make the new type of R3D work. There'd be no suprises for any post house when shooting Red, infact it'd probably become the preferred tapeless camera workflow to be used if they can opt to have their filetype of choice at their desired resolution delivered at the close of set.

Yeah I know it's a daft idea... I haven't been drinking, honest! :usd:

Paul

 

[Joseph Ward]

That would be cool! Here's another one.
What if you can have multiple Brains that sync together even if they are differ brains. They would sync with the lowest brain of course. That way if someone has a different size Brain when shooting, they can have different angle setups with whatever brains are there. :w00t:

 

[PaulClements]

Ok so the "Redcine in camera" brain is a shite name... something like a "Render Engine" is a bit more Red

Paul

 

[PaulClements]

That would be cool! Here's another one.
What if you can have multiple Brains that sync together even if they are differ brains. They would sync with the lowest brain of course. That way if someone has a different size Brain when shooting, they can have different angle setups with whatever brains are there. :w00t:

I've read it and now I'm wondering if I have been drinking... What? lol

Paul

 

[Joseph Ward]

I've read it and now I'm wondering if I have been drinking... What? lol

Paul

Haha! When I read mine, I thought the same.

I was talking about lets say. You have one 2/3 Scarlet Brain and your friend has a Epic FF35. Lets say you want one scene to have two angles views for editing different shots later. The 2/3 Scarlet Brain can be the ''Master'' and the Epic FF35 the ''Slave''. The Epic FF35 can accommodate the B-cam 2/3 Scarlet because it can go lower size/everything but the 2/3 could never go up.

I hope this starts to make more sense.

 

[Mark Pugh]

It's a shocking idea to have transcoding going on in camera while you shoot - you can't tie up the media like that for a start...

However,
I'd spend some money on an off-camera dedicated piece of hardware that reliably transcoded on set, with speed, ease of use, and portability that beat a laptop.

Imagine having great pro-res files available for viewing minutes after shooting, and ready for editing at the end of the day...

Post houses are charging hilarious amounts of money to suckers - I hear $2000 USD overnight is normal - $200/hour for machine time - for a Mac Pro and some cheap software.

 

[Jason Murphy]

A debayer/transcode box would be a very cool idea, though it might lose its usefulness as more post solutions move to native R3D support.

 

[PaulClements]

It's a shocking idea to have transcoding going on in camera while you shoot - you can't tie up the media like that for a start...

No I mean the sensor is sending the data to two boxes, one records as normal to the normal media, the second box transcodes on the fly to a separate media. Giving you two sets of data at the end of the day (linked with timecode naturally). You edit with your offline files that were generated in the render brain and then conform to the original 5k, 6k or 9k files. Alternatively you use 2K or 1080p files generated from the render engine and online.

Paul

 

[PaulClements]

A debayer/transcode box would be a very cool idea, though it might lose its usefulness as more post solutions move to native R3D support.

Perhaps, but I don't see Avid going the R3D route anytime soon so there is still a transcode there and it's still going to be a major player for some time to come. Too many companies use it and prefer it to FCP and Premier. Also updates to redcode take time to implement in each and every post solution. Look at the issues with the new anamorphic setting for example. Fixing the render brain to understand the new redcode would mean any updates would be instantly available to any post workflow solution. As I mentioned in the last post as well you could even use the converted files as an online such as 2K or 1080p and remove the need to conform or transcode anything further down the post pipeline.

Paul

 

[conrad gaunt]

What I'd like is a hardware R3D decoder/ debayer acceleration card, and raw R3D trimming for archiving.

 

[Jason Murphy]

Perhaps, but I don't see Avid going the R3D route anytime soon so there is still a transcode there and it's still going to be a major player for some time to come. Too many companies use it and prefer it to FCP and Premier. Also updates to redcode take time to implement in each and every post solution. Look at the issues with the new anamorphic setting for example. Fixing the render brain to understand the new redcode would mean any updates would be instantly available to any post workflow solution. As I mentioned in the last post as well you could even use the converted files as an online such as 2K or 1080p and remove the need to conform or transcode anything further down the post pipeline.

Good points both. The transcode box is still a great idea, if feasible.

 

[M Most]

It's a shocking idea to have transcoding going on in camera while you shoot - you can't tie up the media like that for a start...

What do you think just about every consumer HD camera on the market does? Granted, they're dealing with less data, but essentially all consumer cameras - particularly the tiny ones - are Bayer chip designs, with an internal debayer and compression engine to write whatever format they write, be it MPEG based, DVCPro based, or AVC based.

 

[Brent J. Craig]

If there is an output module that could handle the data rates (fiber perhaps?) I would much prefer the Render Engine to live somewhere other than attached to the camera.

Why load up the tripod with tons of extra stuff, and try to power it all off of batteries, and lug it around every time you move the camera when you could simply be connected to a nice flexible fiber optic line?

This would probably make the Render Engine much less expensive because it wouldn't have to be ultra-miniaturized.

 

[Bing Bailey]

thats a great idea, specially if you could set profiles and kick out your dailies. the transcoder module could be like a red ray player too. you hook it up. set the profile. it transcodes to the format and size you want then you disconnect and hook it up to a display with hdmi. licensing all those codecs might get expensive. but it is a cool idea. it wouldn't take as long if its doing it as you are shooting

 

[PaulClements]

If there is an output module that could handle the data rates (fiber perhaps?) I would much prefer the Render Engine to live somewhere other than attached to the camera.

Why load up the tripod with tons of extra stuff, and try to power it all off of batteries, and lug it around every time you move the camera when you could simply be connected to a nice flexible fiber optic line?

This would probably make the Render Engine much less expensive because it wouldn't have to be ultra-miniaturized.

My thinking came from the fact that the camera is built to be modular, this would be a module therefore it would stick onto the camera. I also figure it'd be a lot easier to make this doable if it's close to the sensor and doing as mmost pointed out the same thing most video cameras are doing anyway. Personally I prefer it being on the camera than being tethered. It'd make the current body spec weigh about the same as the RedOne if not less so it wouldn't be excessive (I'm guestimating here but since the R1 body weighs 9lbs and the Epic is supposedly 4lbs, I wouldn't imagine this body weighing much more than 4lbs again).

Paul

 

[tj williams]

HEy Paul I have one of these that sets on the desktop and renders whatevr U want onto the disks.... strangely it looks just like a couple of Mac Towers?

Maybe 4 miniturized quad core macs in a tiny box
..... well at least it would keep your hands warm!!!!