- ProresRAW is rare. I’d love to edit in it, but do I really want to wait for an encode? If only there was something that could speed that process up....
The primary point of ProResRAW is to record the primary (master) copy in the format. It would get "encoded" when the camera captures the images. You're at some other end of the process. That isn't where Afterburner is trying to be ( any more than the RedRocket card was trying to be there). [ literally 'Afterburner' suggests the tail end of the a jet engine but that positioning isn't where the product actually fits. It is more so that it is "extra gas" thrown on the thrust the Mac Pro already can produce. It enables an narrow offload so the processor and/or GPU resources can be plowed into something else. ]
Are there a relatively (to whole video capture camera market) low number of cameras that record in ProResRAW ? Yes. However, in the same ballpark as RED and it has it own card for its own format.
( ProResRAW has some upside over time since it covers more camera makers which should also grow over time. Also Atomos recorders which will deepen the supply of ProResRAW footage over the near term time also. )
Afterburner is part of the bootstraping to get ProRes RAW going. If it is a harder to work with format then less folks will adopt it. $2-8K camera and perhaps $2-4 ProResRAW recorder and/or monitor) and the adoption process doesn't have a $15-20K camera cost barrier to entry.
Are there an order of magnitude more cameras that capture HDR 4K video in a non ProRes format? Yes. Is Afterburner going to help with those? No. Is that one of the primary markets Apple was thinking of for Afterburner? Very probably not.
- 8k delivery is extremely rare. Vast vast majority of video industry is not delivering in 8k. Most still in HD for broadcast and streaming. Some in 4k for streaming.
While 8K delivery is rare there is substantive number of folks who are capturing in > 4K formats. Part of it is "future proofing" the content ( perhaps reissue it again later when delivery systems improve). Part of it is "capture and crop" and "capture and adjust deeply".
So the issue is how to you rapidly 'scrub' through "too much" without creating proxies.
Even if not 8K, it is also a high multiple 4K stream multiplier of doing shots with more than a handful of cameras. If there is only 1-2, 4K (or less ), video streams in the mix, then Afterburner doesn't really have much traction.
Higher end "filming" productions is where Apple mainly targeted the Afterburner. It isn't for "most classic Mac Pro users" at all. ( anymore than the Vega II Duo is either. ).
Smaller scale productions it won't have an impact, but I highly doubt Apple was trying to make an impact there with this specific card. The cores and the other GPGPU core in the Mac Pro are targeted at that. Throwing $2K at those would make more of a "bang for the buck" difference.
We need software that is optimized for editing in 4k prores.
That seems more of editing what content than of editing. For relatively high storage footprint ProRes ( as high as can 'tolerate' ) Afterburner makes a difference. For shops that don't ingest high footprint ProRes then it won't make a difference.
"I get a wide variety of stuff and would like to covert all of that into ProRes" . That isn't what Afterburner is for. And frankly that probably not a good job for a FGPA based product. FPGA are good for configuring the gate array to do one job extremely well. If trying to ingest and covert a smorgasbord of formats that isn't "one job". A more general purpose compute solution is highly likely a much better fit ( since it will have to adapt to multiple jobs. )
. But we also need hardware optimized for encoding 4k prores so we can edit in it. And optimized for rendering! That’s the stuff that gets in our way.
Rendering isn't decoding. (at least to me. It seems folks have definitions. ). Again rendering ( adjusting colors , putting "effects" on top of the video ) is adding stuff to what was captured by the camera. Again isn't a particularly good job for a FPGA unless it is just one, narrow type of job. ( green screen effect . title effect. )
Implementing a wide variety of editing knobs and sliders is a general purpose compute task.
Afterburner is also not trying to be hardware solution. It is basically hidden behind the Apple APIs for handling video. If it is there and appropriate it gets used. If not then Apple's software implementations are used. It isn't about other folks software or (plug-ins). [ Again goes toward Apple's vision for handling video at the foundational level. ] For the folks that ignore Apple's library there are substantive adjustments to make. For folks who don't then not as much.
