hi musicvid,

thanks for the reply....

i don't actually transcode the hdv stuff - it plays fine off the tl, it's the avchd stuff i usually transcode since it's usually off one or other of those small consumer camcorders (sony sr9, panny sr9, or whatever number either is that isn't 9 ;-)) in full hd*

my usual end product is a m2t 'master' for archive* and mp4 for web / intranet.

i've used the black magic codec a couple of times while playing around, also the dnxhd with alpha out of ae - which i'm very happy with.

i was just wondering if i was doing right having read so much recently hereabouts about various other options.

*i also usually 'pack' the whole final edit .veg onto a external usb hd just for safety in case the client realizes a couple of months later they've misspelt the ceo's name ;-)

Most neophytes go through a phase where they demand pixel-perfect lossless archives, hang the size or efficiency. I'm sure hard drive salesmen have done a lot to perpetuate that fantasy. I tell them the "best" format is always a copy of the original. There's no need to make it ten times larger to store the same information, and we really don't know what "future-proof" is (anyone remember MCI?). Those of us who have been around for a while (like you and I), have shed that cavalier mindset long ago.

AVCHD (and future iteratations) are a bit of a red herring, though. Great for acquisition, and a relatively small footprint but it doesn't play nice with editors (or players). So, do we store and edit it as lossless RGB or YUV, compressed I-frame or Interframe, or something else? Are we willing to live with huge files, some compression losses, or long transcoding times? These are the questions I was researching until I got bogged down with some health issues last year.

Actually, AVC in itself is not a bad editing format, once we've eliminated b-frames, weght-p, 8x8 dct, and all those other compression tricks that gobble up all our system resources. If there was a smart muxer that would take our AVCHD back to near-baseline (with the exception of CABAC) without actually recompressing, I predict many of our timeline decoding issues would simply go away. Unpacking a transport stream for editing and sloppy decoding for playback are two entirely different things. Something I need to look into more.

In the meantime, I'm still not convinced there is a one-size-fits-all solution for long-GOP source. There are certain codec names that keep popping up more often than others, and people come up with some truly surprising solutions (who would've thought of EX MPEG-2 as an archival solution?).
The sage advice, "Speed, Quality, Size. Pick two." is probably still the best guide for making individual choices.

There's my Sunday sermon. It's an interesting area of inquiry, one that has occupied more of my time than I should permit.

> "I tell them the "best" format is always a copy of the original. There's no need to make it ten times larger to store the same information, and we really don't know what "future-proof" is (anyone remember MCI?). Those of us who have been around for a while (like you and I), have shed that cavalier mindset long ago."

Yea, I archive my HDV to HDV. Call me silly but it works for me. ;-)

~jr

at last, a sermon i can sit through ;-)

seems i'm in good company with jr too ;-)

i have to confess to not particularly worrying about 'lossy' codecs (within reason of course!). coming from the lowband era through all varieties of tape up to betasp it was in ones nature to have as few generations as possible. with bvw and sp i was comfortable with 3gen, though often was forced to 5 on occasions. of course there was loss, but rarely did a client ever see it and by the time it was transmitted it still looked bloody good.

now, with most editing codecs 3 gen is nothing, and any loss with 5 only to be seen by serious pixel peekers, none of whom are clients.

and, to cap it off, when your material is viewed on a mobile or 50" plasma after the usual bout of youtube or cheap cable it's always going to look spectacular in comparison ;-)

I use XDcam mp4 or XDcam mxf over HDV because of the uncompressed (data compression that is) audio. Every time you go from HDV to HDV you add another generation of data compression to the audio. The video may smart-render, but the audio doesn't. With the XDcam formats not using data compression, you aren't doing the same damage to the audio.

Also, I find that in smart-renders, there are bits at the beginning and end that aren't smart-rendered. I find that there is less damage to these bits with the XDcam formats. I don't know why that is and there shouldn't be, but I can see it quite easily.