Sony NEX-FS700R announced!

[David Heath]

Originally Posted by Cliff Totten

Alister, that is pretty interesting.

So, the decoding process seems to be very complex.

Not necessarily - it may not be a very precise comparison but think of 1 pass versus 2 pass encoding. In 2 pass, the first analyses the data to determine the best way to allocate bits - the second does the encoding according to that pattern. The resultant is no more difficult to decode than encoding via one pass.

Yes, that requires all the file to be available before encoding can start (so not much good for real-time.... :-) ), but the principle is that just because encoding is intensive, it doesn't necessarily follow that decoding also will be. In this case, if it stores information about how a frame is encoded, it's conceivable that may let the decoding be easier, as well as giving coding efficiencies.

It doesn't surprise me that XAVC coding is done by a dedicated chipset - so no, don't anybody expect it to be available to any other camera as an upgrade.

Question for Alister (which maybe needs a separate thread?) - I'm getting a bit confused by all the different flavours of XAVC (differing colour spaces, bitdepths, framerates etc) and the datarates each requires. Do you have a reference to any table that lists such? So is it possible to define a given bitrate for (say) 3840x2160 10bit, 4:2:2 50/60 fps XAVC?