-
Prerender/render solutions plox!
Hello creative ones, I could use a bit of help as I’m nearing the end of what I know or what accessible tutorials can teach me. In advance, I apologise for my lack of terminology.
Here’s the outline; 5 minute comp, takes 6 1/2 hours to render out. I’ll try to include all the details.
First, here’s the final render -> https://www.youtube.com/watch?v=WkpoyoyZBow
Here is how it’s structured: there are 6 pre-comps that make up this final comp.
[Main Text/Logo] Simple hologram, has a static block looping from bottom to top. Also has static noise. Shatter applied to add several pixels of depth.
[Left/Right AudioSpectrum] Pretty straightforward, also has Shatter applied for depth.
[BottomLeft text] 3 total precomps make up that text. Top two lines are marquee with same effects applied as main text. The bottom line is a timescale, obv, same effects applied. No reflection on these comps.Ontop of this I had some translucent Particular particles swirling about. Had to strip this to reduce render time but it was negligible in the end. I also plan to go back in and add that effect again, as well as several others.
Now the reason I ask this question about pre-rendering is because, as you can tell, I will be uploading music video’s. My goal is to be doing one (or several) uploads per day. Unfortunately I’m being trolled by Earth’s rotational speed so I don’t have enough hours to spare nor other computers to use.
So guru’s, I need tips on how to pre-render, and the best way to go about it.
Keep in mind the only compositions that need to change depending on the audio track are; the left/right audiospectrums respectively, and the 3 marquee text lines at the bottom. So is it worth pre-rendering the main-text, and any other compositions I may add (that aren’t audio-dependent)? How much time would it save me? Remember, 6 1/2 hours.
Here’s my specs:
AfterEffects CS5
Windows7 64 pro
AMD 1100T hexacore @4.1ghz
8gb DDR3
No OpenGL card (ATI 5870 Eyefinity6 2gb)Here’s what I’ve done:
Configured my memory and multiprocessing to 1.5gb allocated across 3 cores. I played around with the settings, using 5 cores with 0.75gb allocation, after settling on 1.5gb over 3cores, the difference in rendering times was negligible 7hrs vs 6.5hrs.Here’s how I output:
H264. I tried lossless but not only do I not have enough HDD space to contain one of these files per day, but the rendering time difference only shaved off about an hour to 5 1/2hrs.
My comp was created at 1920*2×1080*2. Rendered out at 1920×1080, but will probably settle on 1280×720. Also, rendering at 1280×720 saved no noticeable time.Whew. You’re all champions if you’ve made it this far in my post. So, ideas on how I can start pushing out 5 minute renders in under an hour each? I know it can be done, I’ve been making video gameplay trailers for over a year and Sony Vegas can output a 10minute render in about 90mins on my rig.
Thanks all for reading. I could use a Flintstone tablet or two after that..
-
4 Replies
-
That’s quite surprising, it was a trick I picked up from a random YT tut about making text animations. The theory ofcourse is to start large then shrink, the tip that the tutor gave was in composition settings instead of typing in 3840×2160, to just type in 1920*2×1080*2. Well, it works either way =]
Disabling multiprocessing unfortunately didn’t help at all with the rendering times. So far the quickest (albeit still very slow) was to allocate 1.5gb across 3 cores. I’m still working with that configuration.
I am in the process of rendering out most of the passive precomps as lossless then bringing them back into the main composition. I still have a couple more renders to go but it appears at least the RAM previews are getting a tad more responsive.
Once I’m finished with those renders I’ll do a full render and post back with the estimated completion time. Of course, any more tips always help! Learning a lot trying to conquer these issues. Tya.
edit: also just looking for some opinions; if you watched the vid I posted you may have noticed it’s not playing nice with YouTube and has shrunk it down to 480p. I tried using a similar render process that I used to use with Vegas, so that may be acting up in AE. I’m going to start rendering out my final comps as QT->H264. Good idea? I have no other codecs other then DivX, CCCP codec pack, and whatever came standard in the installation.
-
[Rahim Kanji] “My comp was created at 1920*2×1080*2. Rendered out at 1920×1080, but will probably settle on 1280×720. Also, rendering at 1280×720 saved no noticeable time.”
By working at 3840×2160, you have quadrupled the number of pixels your computer must render (and quadrupled your memory requirements during render) for no reason.
If you are going to deliver at 1280×720, work at 1280×720 to save render time. You can use the “Scale compositions.jsx” script (located under File > Scripts) to scale down your existing comps to a more reasonable size.
Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog – What I’m thinking when my workstation’s thinking
Creative Cow Forum Host: Live & Stage Events -
YT tut about making text animations. The theory ofcourse is to start large then shrink,
——-
That may have been good advice in 1990. These days, software and programmers know how to anti-alias text pretty well.
Work at your deliverable size to speed things up.-Jeff
-
@Jeff any tips to help with AA? Everything I’m doing is by trial and error but this one I couldn’t figure out. In the end with the grid effect I placed on the text, AA wasn’t an issue so making the comp at such a high-res proved to be inconsequential.
@Walter I had already shrunk the comp down to 1920×1080 then I used a null layer to scale everything down. Think it would make a difference reducing it further to 1280×720? Will try it out. And thank you very much for pointing me to that script, should prove useful in my projects.
So here’s the final render. It’s a .FLV @1280×720, 5mbps. Quality turned out the way I wanted but it still took 6 1/2 hours to render. Yikes.. The only layers in the video that are dynamic are the left and right audioforms respectively with their reflections, and the two marquee texts at the bottom. The main text, the subscribe text, and the clock are all pre-renders in lossless. Should it really take this long? I’ve been trying very hard not to use this word as it’s far more rewarding to do things on my own and experiment, but now, I kind of need ‘help’. I need a shower after saying that..
But thanks a ton guys I assure you I’m learning a lot. =]
‘Ere’s the final render. I won’t be adding or removing anything else, may just play with the midtones of the subscribe text but otherwise it’s a finished template. (Warning, the video opens ON a beat so adjust your sound levels before clicky.
https://www.youtube.com/watch?v=kRs56mnKXdM&hd=1
Some contents or functionalities here are not available due to your cookie preferences!This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.
Reply to this Discussion! Login or Sign Up