Guess Who Holds the Keys to 4K Television and Broadcast?

[Peter Moretti]

And Panasonic has a competing 4K resolution I-frame codec based on H.264 ...

I believe this has potential to be "better" than 4K SR b/c 4K SR will continue to be H.263.

 

[Blair S. Paulsen]

What seems likely to me is that SR tape will dominate HD delivery for years to come. Ironically, it is my guess that Sony's file based SR specification will become the de facto "standard" for file based deliverables:

"MPEG-4 Simple Studio Profile is an Intra picture compression algorithm designed by the International Standards committees, ISO / IEC, and is defined in the MPEG-4 14496-2 specification. Sony adopted MPEG-4 for HDCAM SR technology and is in use in the current hardware products. It was initially made to comply with up to four Levels. These are Level 1 through Level 4. Successive development efforts expanded the standard up to Level 6 for 4K 4:4:4 50p/60p capability with a data rate of up to 3.6Gbps."

As Mike noted, most industry groups and standards bodies are looking exclusively at open source solutions. It would seem that Sony has adopted an open spec and since they have such a long track record in the broadcast space...

Obviously the wild card, on so many levels, is RedRay with it's ~20mb/s bit rate. Yes, the file format is likely to remain proprietary, but if licensing is cheap and/or the hardware affordable that might not kill it. I believe that if RED can propagate an easy to use system that works a lot like a tape deck (simply using SSDs instead of tape carts) that can integrate fairly simply into broadcast infrastructure using established protocols they can own 4K distro.

As many have pointed out, the early 4K ecosystem is likely to develop outside of the risk averse, margin challenged broadcast "channel" (pun intended). If RedRay (or whatever it ends up being called) can get a foothold in the "pre-broadcast" phase of 4K development (and its cost/ease of use equation is right) then it could become a train rolling downhill - especially as the distinction between IP delivery and traditional broadcast erodes.

One big unanswered question for me is whether RED will go after the cinema distribution space. Can they make RedRay the next logical step after DCI?

Cheers - #19

 

[Peter Moretti]

Blair,

I agree that Sony is well positioned. There are a few things about using Sony's SR MPEG-4 for file based workflows. One is that SR was designed for tape, which means it uses the older codec H.263, is constant bit rate and can't easily uncompress to partial resolutions.

Now how much of an advantage of H.264 is over H.263 is not clear. I've read 30% and I've also read "not much" in real life and with high bit rates. Cineform (representing a modern codec) and Sony have both tested SR vs Cineform and have come to the conclusion that they are essentially the same, with one out performing the other depending on source material.

A constant bit rate is essential for tape b/c the tape has to run at a certain speed for recording and playback. But file based recording can record at a variable bit rate which will give better results for the same file size.

I think the biggest Achilles heel w/ Sony's newest additions to SR is that the files can't be easily down-rezed. With a wavelet compressed file, you can access a lower rez that is a multiple of the highest rez. For example, 4K can be accessed as 4K, 2K or 1K. This has to do with how wavelets compress with multiple layers, each of which builds upon a lower level that is visually intact. So wavelet compression offers a real advantage in terms of storage and simplicity, b/c you don't have to create and save multiple file sizes. For example, Youtube presently creates additional lower rez files for HD material that is uploaded. This means more storage and more database entries and files to manage.


P.S. The image you attached must have been compressed using MPEG-1 at 24i, b/c I can't read it at all, LOL!

 

[Blair S. Paulsen]

Agree with Peter that the ability to extract lower res versions from the master wavelet file is a big plus. If JP2K were to be implemented on the scale of MPEG-2 with commoditized encode/decode chips under $10 each that would be great - but not likely IMHO.

In any case, the real issues are reliability, consistency, ease of use, smart compression levels/type, ease of integration with existing infrastructure and, bottom line, widespread uptake.

Cheers - #19