ADOBE AFTER EFFECTS: Forum Expressions Tutorials Creative Cloud

after effects cs5.5 camera 3D rig ghosting

COW Forums : Adobe After Effects

<< PREVIOUS   •   FAQ   •   VIEW ALL   •   PRINT   •   NEXT >>
Mina Nicolasafter effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 12:06:45 am

well I used the new 3D rig and tweaked the settings and created a nice looking scene.

the problem is, when I view the video (or still frames) in the composition window , the 3D effect is great and there's no ghosting.
but when I render the video, it becomes very blurry with a huge amount of ghosting that ruins it all.
rendered still frames are great but still has a small amount of ghosting but the depth is reserved.

I'm rendering it in H.264 with a bitrate of 9.4 Mbps

am I doing something wrong? or is there something I should do?
it's my first project using the 3D camera rig so any tips (especially for rendering) are highly appreciated.


thanks


Return to posts index

Victor NguyenRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 4:51:16 am

could you give me a screenshot of your problem? and also a screenshot of your settings when you export?


Return to posts index

Mina NicolasRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 12:07:13 pm

sure

here's a rendered frame


and that's a sample rendered video:
http://www.mediafire.com/?0zzz5grou3uy9hp

and these are the settings:




Return to posts index


Michael SzalapskiRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 6:02:56 pm

One thing that may (or may not) help would be to render a lossless file out of AE to test it first. Then use that lossless file to compress.
AE is pretty bad at compression, especially stuff like H.264. I'll Let Dave LaRonde explain:
Dave's Stock Answer #3:

Don't use AE to compress files for final delivery. The various compressors are there only to make quick 'n dirty files showing a project's progress to producers, clients, the kids, etc. AE is incapable of doing multipass encoding, a crucial feature that greatly improves the image quality of H.264 and MPEG-type files in particular.

Render a high-quality file from AE, and use a different application to do the compression. Popular ones are Adobe Media Encoder, Sorenson Squeeze and Apple's Compressor, which comes bundled with Final Cut Suite. Even compressing in Quicktime Pro is better than compressing in AE.

Making good-looking compressed files is almost as much an art as it is a science. It is NOT straightforward at all.
The COW even has a Compression Techniques forum.

Rendering a lossless file out of AE and using the Adobe Media Encoder or some other compression app for your final compression offers another benefit: you can test different settings without having to rerender your AE comp each time.

- The Great Szalam
(The 'Great' stands for 'Not So Great, in fact, Extremely Humble')

No trees were harmed in the creation of this message, but several thousand electrons were mildly inconvenienced.


Return to posts index

Mina NicolasRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 6:25:43 pm

thanks for your reply

I tried it but unfortunately it gave me the same results :(

any other solutions?


Return to posts index

Michael SzalapskiRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 14, 2011 at 7:16:28 pm

The info on the following pages should help you out a lot. There is talk of ghosting and a whole lot more.

[Link 1] - this is the biggie
[Link 2]

- The Great Szalam
(The 'Great' stands for 'Not So Great, in fact, Extremely Humble')

No trees were harmed in the creation of this message, but several thousand electrons were mildly inconvenienced.


Return to posts index


Mina NicolasRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 16, 2011 at 3:58:15 pm

thanks a lot I'll check them :)


Return to posts index

Victor NguyenRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 15, 2011 at 3:49:58 am

where can I find some more of Dave Larode stock's answer? I been looking for those since forever.


Return to posts index

Michael SzalapskiRe: after effects cs5.5 camera 3D rig ghosting
by on Sep 15, 2011 at 1:31:26 pm

You want Dave LaRonde's stock answers? I've GOT Dave LaRonde's stock answers!

Dave's Stock Answer #1:

If the footage you imported into AE is any kind of the following -- footage in an HDV acquisition codec, MPEG1, MPEG2, mp4, m2t, H.261 or H.264 -- you need to convert it to a different codec.

These kinds of footage use temporal, or interframe compression. They have keyframes at regular intervals, containing complete frame information. However, the frames in between do NOT have complete information. Interframe codecs toss out duplicated information.

In order to maintain peak rendering efficiency, AE needs complete information for each and every frame. But because these kinds of footage contain only partial information, AE freaks out, resulting in a wide variety of problems.



Dave's Stock Answer #2:

When you're out on a shoot, and you say, "we'll fix this in post" without knowing PRECISELY HOW you're going to fix it in post, don't shoot it! You'll only end up shooting it over again.

Since post typically costs three times the cost of production, fixing something in post is not a way to save money, but rather a way to spend more of it.

And, before you say "we'll fix it in post," always consider who's doing the work, especially if you're the one doing the editing.



Dave's Stock Answer #3:

Don't use AE to compress files for final delivery. The various compressors are there only to make quick 'n dirty files showing a project's progress to producers, clients, the kids, etc. AE is incapable of doing multipass encoding, a crucial feature that greatly improves the image quality of H.264 and MPEG-type files in particular.

Render a high-quality file from AE, and use a different application to do the compression. Popular ones are Adobe Media Encoder, Sorenson Squeeze and Apple's Compressor, which comes bundled with Final Cut Suite. Even compressing in Quicktime Pro is better than compressing in AE.

Making good-looking compressed files is almost as much an art as it is a science. It is NOT straightforward at all. I recommend asking a few questions at the COW's Compression Techniques forum.


I also have some "stock answers" from others:

Walter Soyka's post on improving performance:

If you're used to working with SD or even HD comps on modern computers, the render times for larger comps can be stunning. I frequently work with large comp sizes (4K and up), and I've found it's very helpful to adapt my workflow a bit to stay productive despite the lengthy render times.

There's an outstanding page on Improving performance in the After Effects help system, which includes all the features I'll mention below and many others.

Here are some After Effects features—some common, some often-overlooked—that I rely on when working with larger comps:

  • Multiprocessing: if you have a multi-core system and at least 2-4 GB of RAM per core, you can enable multiprocessing. After Effects will launch multiple copies of its renderer to processing multiple frames simultaneously.

  • Zoom and resolution: these go hand-in-hand. You can reduce your preview resolution to save processing time and RAM.

  • Caps Lock: when on, Caps Lock disables rendering for preview.

  • Draft 3D, frame blending, and motion blur: toggle these to speed up previews. For render, you'll probably want Draft 3D off, and frame blending and motion blur on, but they are all render-intensive, so you can save time by toggling them for working and preview.

  • Proxies and pre-rendering: allow you to render out a complex pre-comp, save it to disk, and refer to the footage on disk instead of re-rendering. Once I get a pre-comp working well, I'll proxy it so I don't have to continually re-render it during previews while working on other elements. You can use proxies with a wide variety of settings: simple stills, draft-quality renders, or high-quality renders. If you use proxies instead of pre-renders, make sure you change your render settings in the render queue to use proxies.

  • Region of Interest: allows you to select a smaller section of the comp to preview; saves on processing time and RAM usage.

  • Shift-RAM preview: allows you to choose different setting for RAM preview (0 on the numeric keypad) and Shift-RAM preview (Shift-0 on the numeric keypad). I use regular RAM preview for high quality previews so I can see detail in my work, but I use Shift-RAM preview at quarter or eighth resolution, skipping at least every other frame, so I can quickly see the overall effect.




Rob Neal in the VideoCopilot.net forums on Slow motion with "daddy's camcorder that I borrow".
Q: How to get clean, crisp, smooth slow-mo?
A: Use a high speed camera with a decent lens and a fluid head tripod. Oh, and TONS of lights.

Without this it's like asking "How do I get 200mph out of my car? - I own a VW Golf".
Well it's simply not going to happen. You have neither the resources or the money to do what is really needed. Yes, there are fudges, fixes and workarounds to create fake slo-mo, but you seriously can't expect high-end results from a cheap consumer camera, no matter what tricks you chuck at it. You can't make a silk purse out of a sow's ear.

Some things were never to be. Sorry.


Rick Gerard from the Adobe After Effects forum on why AE is slower than NLE's to Preview

Sony Vegas, Premiere Pro, Final Cut, Avid are all NLE's (Non Linear Editors) and they are specifically designed to playback a video stream. With any of them, if you stack enough layers or effects on the video they will have to render a new video stream based on pixel based calculations for every pixel in the stack. This rendering, especially for HD sources or for complex plug-ins, will take quite a bit of time.

After Effects, Flame, Fusion, Shake -- are all pixel based image processing applications that act very much like Photoshop. They calculate the values of every pixel in every frame, come up with a new pixel, and then play those pixels back as a video stream. More importantly, AE and all the other pixel based compositing apps, always work internally with completely uncompressed pixel data. NLE's rely on codecs and in some cases, hardware, to playback the video. It's an entirely different way of working with moving images.

In After Effects you enable the preview by loading a bunch of frames into RAM then the video stream is played back. You start the process with the 0 key on the numeric keypad and not with the Space Bar as you do in nearly every NLE ever created. The length of the preview depends entirely on how much free ram you have available and it takes some time to generate these new pixels. The more layers, the more effects, the more calculations that need to be performed the longer it will take to process the RAM preview. There's currently no way around this rendering time. A modern NLE will handle an amazing number of video streams simultaneously, but as soon as you exceed the capability of the system you're stuck with a render. Most NLE's, given the same number of calculations, actually take a little longer than After Effects to do the same kind of effects. Open GL, and other GPU acceleration helps many NLE's achieve higher performance but it has yet to be implemented into a pixel based compositing app. The sad truth of the matter is that if you want to do compositing in any of the available compositing apps, you have to wait for renders. They are getting better. Memory management and efficiency is improving. GPU accelerated effects are being added, but for now, that's about as good as it gets.

I hope this helps. As long as you use After Effects to create shots and don't try to make it do the work of a NLE you should be fine. Movies come from NLE's, amazing shots come from AE.


Rotoscoping tips from Jeff Kaplan at http://www.point360west.com, Adobe's help files, and me:
Unless you have CS5 which features the new rotobrush tool - the only fast way to rotoscope is to do so EFFICIENTLY by "keeping your points to the joints."

In-Depth Tips on Rotoscoping:

Use as _few_ spline curve points as possible to mask the shape -- the less points you have to move, the less time the tedious process will take.

Create as few keyframes as possible to minimize edge jitter and retain fluid motion. If you set keys every frame, it's likely that you'll run in to jittery edges, though sometimes you _have_ to set keys every frame with fast motion. Depending on the source material, you may have a 30 frame clip to roto ... It may be conducive to start with a mask on Frame 0, then one on Frame 30. From that point you could add more keys at Frame 15. Then Frames 7 and 22, etc. (See Point 5 before you do this technique, though.)

Try to keep specific points in your mask corresponding to specific parts of your source material. For example, if you have three points that make up the armpit, elbow, and wrist joint of the edge of someones arm, make sure that you keep the armpit point going to the armpit in each frame. This can be difficult sometimes if there is a lot of movement, but if you don't stick to it, your points will "crawl" around the edges of the object and will cause funky edges and make you have to create more keyframes than is neccesary.

Use multiple masks to generate a matte. There is no reason to try and get the outside edges of a complex shape (such as a person) to fit all in one mask. Make seperate masks for the head, arms, torso, etc. as appropriate and add them together.

Try to pick "natural" keyframe points from the source material and fill add detail as needed. A basic example of this would be if you needed to cut out a person jumping on a trampoline. Obviously, your starting point for keyframes are at the apex and bottom of their jumps. Then you fill in more keyframes in between these points to make the mask more accurate. This is probably a better way to deal with roto than the aforementioned process of just cutting the clip in half over and over (the 0, 30, 15, 7, 22, etc. business) because it pays attention to the natural points of the clip that require keyframes.

Get some good music to listen to. Use the force. Try not to go crazy if you do it all day.


Some more AE specific tips on rotoscoping available here
Such as:
  • Immediately after beginning to draw a mask, press Alt+Shift+M (Windows) or Option+Shift+M (Mac OS) to turn on keyframing for that mask and set a keyframe. This way, you won’t edit a mask frame-by-frame for several minutes (or longer) and then realize that you lost all of your work on previous frames because you forgot to click the stopwatch to make the mask path animated.
  • Draw your masks on a white solid layer with its Video (eyeball) switch off, above the (locked) footage layer. This way, you run no risk of accidentally moving the footage layer when you manipulate the mask, and you can also much more easily apply tracking data to the mask. (You apply the tracking data to the invisible solid layer that holds the mask.) This also means that you don't lose your cached RAM preview frames each time you manipulate the mask. (See Toggle visibility or influence of a layer or property group and Lock or unlock a layer.)
  • Turn on the Preserve Constant Vertex Count preference. (See Designate the first vertex for a Bezier path.)
  • When possible, transform (rotate, scale, move) the whole mask or a subset of the mask vertices instead of moving the vertices individually. This is both for efficiency and to avoid the chatter that comes from inconsistent movement across frames. (See Move vertices in free-transform mode.)
  • Manual motion tracking is less time-consuming than manual rotoscoping. The more effort you spend getting good tracking data for various parts of your scene and object, the less time you'll spend drawing and fine-tuning masks. (See Tracking and stabilizing motion.)


- The Great Szalam
(The 'Great' stands for 'Not So Great, in fact, Extremely Humble')

No trees were harmed in the creation of this message, but several thousand electrons were mildly inconvenienced.


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]