max render quality always seems to take longer; as well, any filters, color correction, luma etc will amp up the render time. Oddly enough, I was using LT for some SD TV broadcast episodes, and ended up reverting back to standard 4:2:2, seemed to be a better fit.
Thanks for the quick answer.
I know that max render quality takes longer, but seems like it still shouldn't take longer then exporting 1080p file (also exported with max render quality).. Either way I tested it now, and unchecking the max render quality option does seem to make the export a bit quicker, but still a lot slower then it should be.
Since we are exporting final exports after merging AE rendered video + sound mix, the timeline in question has no filters or effects - it's just a basic timeline with 1 video channel and 2 audio channels.
Another thing that might have been unclear from my original post - the problem only exists with PAL files in 4x3 aspect ration. So this rules out problems with the codec itself (since ProRes LT w/ PAL 16x9 works just fine).. This might just be a bug, but since this is the format our client requires and we are trying to avoid having to convert the video after the export, I'm still looking for a workaround that doesn't require a change in codec.
I am not sure of this but I believe if you have a CUDA GPU that you do not need to check Maximum Render Quality as the GPU does that automatically,
just ran a few tests on my machine. I'm using an iMac 2011 with 12 G Ram. I started with a 15-second AVCHD clip in a qucktime MOV wrapper, just 1 vid + 1 aud track. Here's the render times for that 15 sec piece:
TO 1080p ProRes 4:2:2 // 14 sec Max render switch on, same with off
TO 1080p ProRes 4:2:2LT // 10 sec (same time with Max render on or off)
TO NTSC DV (720 x 480) 16 sec (Max off) // 1min 13 sec (max render on)
TO PAL DV (720 x 576) 16 sec (Max off) // 1 min 30 sec (max on)
So, you're in the same ball-park. The max render switch seems to take effect in re-sampling video for a different frame-size, and *note* PAL requests a different frame-rate as well. It's that switch that really amps up the render time. I would definitely use that Max parameter if you're exporting to broadcast; for web it's need would be questionable. Just bite the bullet and grab a coffee in the meantime!
Thanks again for your help, for some reason your export seems to be a lot faster than ours, even though we are using Mac Pro towers with stronger specs then an 2011 iMac.
The only thing you seem to be doing differently is having AVCHD in MOV wrapper as your original footage, our's is natively ProRes 422, and the export from ProRes (4x3 PAL) to ProRes seems to be extremely slow on Premiere. This should just be your run-in-the-mail "same as sequence" export, so I really don't see a reason for it to take so long.. I don't want to be "that guy" that comes into the Adobe forums with complains, but with FCP7 we obviously never had these issues.
We are indeed exporting to broadcast - we are making TV Promos for a couple of major channels and a big satellite company here in Israel. our kids channels unfortunately still air in 4x3, and since we are exporting 5 versions for each promo, and about 3 promos a day for that specific client - well.. that would be a lot of coffee!
I am also aware that PAL requires 25fps, that is the standard for broadcast here and the original footage is shot and delivered to us that way, so that shouldn't affect the export either.
The thing I'm still finding extremely weird about this whole thing is that this problem doesn't seem to exist with widescreen (16x9) PAL footage, but only with 4x3. I would understand it taking longer if this was a WS sequence which premiere had to re-conform during the export, but that is just not the case.. I'm basically taking a video that already has the exact specs I'm using for export (that was rendered out of AE), PAL 4x3 ProRes 422 LT, 30 seconds video, adding an audio track to it and exporting out with the same settings as both the sequence and original footage. There is no reason that should take 5-10 minutes on machines that do the same for 1080p in one minute.
Thanks again, and as always any further info would be greatly appreciated,
My understanding of the "Max Render" checkbox is that it only affects SCALING, for instance HD to SD. If the source and export formats match, then there is no reason to check that box.
Also, most users avoid using the "Match Sequence" option, just select an appropriate preset and take control of it yourself.
I remember someone telling me the max render checkbox also affects things like general sharpness of the exported image, color spaces etc.
If that's not the case, I'll try to test again without the max render box checked and see if I get different results.
Is there a known issue with the "Match Sequence" option, or is that just a general precaution?
The person usually doing the final export in our workflow is the sound editor - the AE artist creates a sequence in premiere with the correct settings and inserts the final versions (4-5 clips) of the promo. Then the sound editor just needs to open the project, insert the sound to the same sequence and render out using "Match sequence". The sound editors don't know much about video and I would prefer for them not to have to think about it too much.. In FCP7 they just used the 1-click export "same as source".
I'm pretty sure I already tested this issue with manual settings but I will test it again either way.
I concur with Jeff that Max Render is only needed if you're changing frame size (I would add "or changing aspect ratio"). I still use it for all renders destined for broadcast, just to be sure, even if it's SD material composited for SD output.
I did read a great article on the max render switches (and max depth as well), somewhere outside of the Adobe Forums. Google it, and I'm sure you'll find it.
The "match sequence" thing is indeed horrible. It always chooses some arcane MPEG2 format or something; I never go on "autopilot" that way ..!