Steve Renard
Forum Replies Created
-
Looks to me like 3D Stroke could do it. Try playing with multiple layers and composite modes (Add, Subtract, etc) then adding a glow to the whole thing with an adjustment layer.
If you have Particular, check out this tutorial that does something similar using that effect: https://library.creativecow.net/articles/kramer_andrew/nano.php
Good luck,
Steve -
I think you’ve already hit it, at least partly, with the precomp, but I’m not sure if I completely understand what you’re suggesting. Here’s what I would do, based one what you said: put two layers of your RGB pass in a timeline, then go to Effect => Color Correction => Channel Mixer and put everything to 0 except Blue-Blue, for example, to pull out the blue blue channel on one, and then on the other pull out the red channel using the same effect. Once you’ve isolated your channels, use the Tint effect to make them black and white, then precomp them together and use that comp as a Luma Matte for your source, which should reveal what you want to reveal.
If this doesn’t make any sense, post a video or still so we can see what you’re talking about.
Cheers,
Steve -
There are two elements to the animation happening here, but I think both could be taken care of by making the text as a still in Photoshop (with each character on a separate layer), then animating those still elements in AE. The first part, which they used for most of the text, uses the Write On effect in “reveal original” mode. There are plenty of tutorials out there that show you how to do this, including a couple on the Cow, I think.
The other part of the animation is a few characters (the “m” in comcast, the “y” in security, etc.) that have arms moving around or spinning or growing in or whatever. You could tackle this a couple different ways. The first would be to go back to your still of the text and cut those pieces of those characters apart into separate layers before bringing the whole thing into AE. Then you could animate those on their own. The second way to do it would be to use the Puppet Tool on just those characters, and figure out a way to hold the rest of that character still while you animate just that arm in. Guessing again here, but I’d say they used option 1 for the second “hump” of the “m” in comcast, and the Puppet Tool for the “y” in security.
Good luck on it,
Steve -
Not sure if this will work, but maybe come at it from the other direction. Instead of trying to stabilize the footage, do a motion track on the corners of the sign in the original shot, then use an expression to tie the corners of your corner pin to your tracking points. If you’re not familiar with expressions, check out Dan Ebberts’ site or the Cow AE Podcast for the basics. Really all you’ll be doing in this case is Alt-clicking on the stopwatch to turn expressions on for each of your corner pin points, then using the pickwhip to link that element to your tracking points. If the corners of the sign travel into or out of the shot, check out the tutorial on 3D Set Extensions at videocopilot.net for a great demonstration of how to use null objects to get around that particular problem.
Good luck!
Steve -
Looks to me like you’ve turned on the aspect ratio correction. Turn it off, or make yourself a square-pixel comp, and that should sort it out.
-
Thanks to everyone who’s offered their thoughts and I appreciate all the input. I did finally get a chance to talk to the original shooter and it was a creative decision on that end (the footage was not shot for this project originally). I’ll see what I can do with it, keeping in mind the soft look they were going for, and hopefully come up with something good.
Steve
-
Yeah, I got there in the end, but it looks the same. The best looking result I’ve had so far is the 24fps conversion, but clearly I need to talk to the folks who shot it to figure out what to do with this stuff.
Thanks for your help!
-
When you slow down footage in FCP, it gives you the option to use frame blending, which (for example) gives you a half-and-half dissolve between each frame in order to fill the extra time. When you take it over to Motion, you can keep with the frame blending, or use Optical Flow, which as I understand it uses some sort of algorithm to interpret new frames in between the orignals. So if frame blending would give you A A/B B B/C C, optical flow will give you A A+1 B B+1 C at 50% or A A+1+1 B B+1+1 C C+1+1 D at 25% (I think). Someone with more technical knowledge might be able to explain better.
What I’ve noticed when I’ve worked with optical flow is that you have to have really clearly defined shapes in order for it to work well. So, a dancer in a white leotard on a black background looks good. A gray car driving down the road on a rainy day not so much.
-
The timeline is 60fps. I have not run it through the Panasonic Frame Rate Converter – to be perfectly honest I don’t know what that is.
As far as I know, it was shot only to look a bit rough, not to look sped up.