You should not be using QuickTime pro to export h.264 Quicktime. H.264 never worked very well in a QT container and is not being developed or supported by Apple. You should be using the Adobe Media Encoder and one of the H.264 presets that match your frame size and frame rate. The only thing I would change would be to enable multi pass rendering. It will greatly improve the color artifacts in detail areas.
Render times depend on your project. In the latest version of AE you are not going to see a significant speed increase in basic rendering. AE just isn't using system resources efficiently enough to take full advantage of the nearly double multi core speed in the new machine. That may change a bit in the next release but I would not expect efficient use of multi core/multi thread processing any time in the next year or two.
Depending on the project you may be better off rendering a visually lossless DI (digital intermediate) using a suitable and fast rendering production format and then using the Adobe Media Encoder to render the compressed deliverable product. It all depends on what is going on in your comp. Rendering compressed formats is always going to be slower than rendering to an image sequence or most visually lossless production codecs.
If your projects involve more than one shot I would use AE to create the shots you cannot create in Premiere Pro, render a DI, finish the edit in Premiere Pro and if there are sound edits that are important polish the final in Audition, then render your final deliverable using Premiere Pro and one of the AME's presets. In most cases, if you have more than a shot or two in your final product, this is the fastest and most efficient workflow.
More than 90% of all my comps are one shot under 7 seconds. Most of my edits are a half hour or longer. About 10% of them are five or ten minute shorts or explainer videos. I never create an AE comp that is longer than a single phrase of music, copy or narration. It's just too inefficient. Typically a 2 minute project for me would have 20 or 30 shots. If it was an explainer video or a demo video two minutes may be 7 or 8 sentences of copy. Each sentence wold be a separate comp. I would design at the maximum client required size or boost it up from HD to 4K if I thought they would come back to the project in a few months, then render a DI at full comp resolution, then use the AME to render the different deliverables. I have found this to be the most efficient workflow, especially if you have a client that requests changes.
I hope this helps.
thankyou so much for your time and detailed explanation! really interesting.
my project sounds very similar to what you mentioned.. i do however create the whole thing in AE. mainly because i dont really like jumping back and forth when client changes, timing changes come in, i think i might give it a go next time though. eg. if theres 20 text animations, i will precomp each one and have a main edit comp in after effects where i compile everything. i guess having to render each one out and put it into premiere will slow me down a little bit, but at the end of the day, the final render will be faster.
by DI, do you mean rendering something like a prores file? would you be rendering the graphics with the footage and put that into premiere? or render with an alpha? not sure if rendering with an alpha slows things down. alpoha makes sense as if its just a timing change, or slipping the graphics back for example, i wouldnt need to render the shot again.
Jpeg 2000 supports alpha and is fast to render. Cineware (from GoPro and free) can be 10 bit and supports alpha. Sequential tiff files are extremely fast to render compared to a lot of other formats and can be up to 32 bit. The format I use for the DI depends entirely on the project. Rendering an alpha does not take any longer than rendering without alpha if you pick a fast to render format.
I've been making client changes for all of the 20+ years that AE has been around and trust me, jumping back and forth between a NLE and an AE comp that is no longer than a sentence of dialogue is way way way faster than making a small change in the middle of a comp with 50 or 100 layers. That's the way they cut feature films - one scene per reel, then assemble the reels. Reels became sequences in NLE's like Premiere Pro. All of the sequences are assembled in a final full movie. It's the best way to get the best edit and make changes quickly. You just have to get used to adjusting your thinking.
Someday I'll do render tests on the most common mezzanine codecs (look it up) and publish a list for Mac. I'll publish the project files so Windows users can also run tests on their favorite mezzanine formats. There is a lot of information on production formats out there that 90% of all AE users should look up but most don't bother.
What about dynamically linking the ae comp instead of rendering it? Is that slow to work with in premiere?
Also, regarding archiving, i guess the downside is that you are left with a much bigger project file size because of all the extra linked renders.
Dynamic link works with simple quick to render compositions. Anything complicated will be faster if you render a DI and put that in PPro. This is because Dynamic link opens up AE in the background and uses even less of your computers processing power to render than AE does when it is in the foreground.
You can render DI's using the Adobe Media Encoder and continue to work in AE on other parts of your project while the AME is rendering away. I do this all the time. I usually do not spend any time at all waiting for AE to render. I just send the comp to the AME, push go, then open up a new comp or even a completely new project in AE and start work again.
I also seldom of ever do motion previews at 100% magnification and full resolution. Normal workflow is to do any animation with all effects off and the comp at 1/4 or maybe 1/2 resolution and check the timing and flow. I call this a pencil test after the pencil test cell animators use to check the motion before things go to ink and paint. I then start adding effects and the rest of the comp. Critical frames are checked at 100% magnification with resolution set to auto and a final check of things like motion blur and color correction are checked. Then the comp goes to render. After a while you'll know exactly what your renders are going to look like without ever waiting five or ten minutes for a preview to render. That's is when you start building efficiency and meeting deadlines.
As for project size, the size of a comp or a Premiere Pro project doesn't change when you add footage. Rendered intermediates do eat up a bit of storage but drives are incredibly cheap. I archive all footage that I have shot, all images, and all project files based on Client - job and date. I use a Raid drive system with two mirrored drives. At least twice a week I change drive #2 with a fresh drive so I always have at least 2 copies of everything, including the footage and most of the time I have 3. Every Friday I put one of the copies in a safe and send one to a safety deposit box so I can go back a year later, find the footage I need or the renders or the projects I have done for a client and make changes. In the early days I spent close to $20k for a few gigabytes of Ultra SCSI storage but today I picked up 3 new 2TB drives from Fry's for under $200. Storage is such a small part of my cost of doing business that it's not even an issue. A really good project management system and backed up storage plan is critical to being a successful film maker. Try to get one in place as soon as you can.