Is Apple ProRes 4:4:4 ?

[Jeff Brue]

Doing a ton of studio finishing wor.k...it's all 4:4:4 SR... Mezzanine files are all Mpeg-2 4:2:2 i frame.


As far as Apple's new 4:4:4 codec anyone else notice that it's saying its 12 bit ? Doesn't that put us out of sampling terminology like 4:4:4 ? None the less the KPro just became a much more interesting proposition. Mike M thoughts ?

 

[Blair S. Paulsen]

KiPro plus ProRes4444 creates new bang for buck benchmark

KiPro plus ProRes4444 creates new bang for buck benchmark

Jeff makes a great point. If the AJA KiPro can write 4:4:4 at 12 bit to full raster 1080 by 1920 at up to 30/P it would rock the house.

Will the KiPro start replacing SR decks all over town if the top ProRes codec is at least as good as, if not better than SR quality?

Anyone know how many tracks of audio you can have? Gary A.? John Thorn? Maybe Mike M. knows...

 

[M Most]

As far as Apple's new 4:4:4 codec anyone else notice that it's saying its 12 bit ? Doesn't that put us out of sampling terminology like 4:4:4 ? None the less the KPro just became a much more interesting proposition. Mike M thoughts ?

The numbers 444 now represent a ratio rather than actual sampling rates. The numbers were originally based on standard def video sampling rates, and the multiple of the subcarrier frequency that was used as the actual sampling rate. As we've gone to higher definition video systems, the numbers were retained (rather than use, say, 33:33:33) because it was simpler to use the same terminology, but have it refer to the ratio of sampling luminance vs. the chroma. So 444 in this context simply means that the luminance and chroma are sampled at the same rate, where 422 would mean that the chroma is sampled at half the rate of the luminance. 12 bits is a measurement of precision, the word length, which relates to each item that is sampled. It has nothing to do with the sampling rate, which is basically a frequency measurement that indicates how often the sampling occurs.

The Ki Pro has not shipped yet, at least not to my knowledge - which could, given their rather tight relationship with Apple, indicate that it was held back until the new codec was completed. Of course, 444 sampling doesn't mean anything unless they change the hardware to accommodate dual link input or allow 3Gb operation on the HD-SDI input, which they might do. I would also be concerned about the ability of a single drive to reliably do sustained writes at the rate demanded by the 444 codec. I don't know about audio channels, but my guess is that it can likely handle at least 8 embedded channels. Of course, that means whatever feeds it must have the ability to embed audio in the HD-SDI, which most newer cameras are capable of.

 

[Dominic Jones]

330Mb/s is about 42MB/s, which should be well within sustained write speed for modern high-density disks, at least for most of their capacity. We've got some of the new Seagate 1.5TB drives in our newest Mac Pro, and they're sustaining nearly 100MB/s, single disk, 50% full! Obviously SSD disks will eat this alive as well...

It's a very exciting advancement, although, if it's 12-bit and 444, then the compression ratio is nearly twice as high as for PR422HQ, so I guess we'll have to see how well it holds up compared to SR - very well for a couple of compression cycles, I would have thought, but perhaps not so well after that...

Great codec for RT playout of a non-RT finished RGB/444 project for lower budget productions, though, if nothing else.

Dom.

 

[Evangelos Achillopoulos]

I already asked AJA about KiPro and 444... its a no go... a next version will probably have it... at least a year from now...