1 person found this helpful
Well if you have layers at 25000px from camera, they will of course be really tiny. You'll have to animate the camera closer to see them, or maybe select a longer lens, like 200mm.
There is no limitation in the AE 3D world, but see the Z axis as a "scale" axis. The hight z position, the smaller your object is, and at one point, it's normal to see nothing as the reduction must scal your layer around 1px 1px or less.
If you have a fixed camera, it's really rare that you'll need to go so far with your layers.
Hope that helps
i mean i cannot drag the preview zone (top view) beyond the 25000 pxls by Z axis! Strange but that's it! I know how Z-distance is connected with scale... hmmm
1 person found this helpful
How okay, i understand better now.
Well yes, you cannot move the top view very far up/down/left & right, even at 2 or 3% zoom. So I guess this is rather a design limitation of the "fixed" views than a limitation on the z depth it self.
Well, AE's composition buffer cannot exceed 30000 pixels in either direction. By displaying items in perpendicular views you hit just that limit. AE does not have such a thing as a real 3D space, so in fact the alternate views should be considered kind of a hack way to solve this dilemma within the existing framework. I don't think it would be possible to go beyond that without rewriting some parts of AE considerably.
Sebastien, thanks for explaination. It's exactly my case with camera moving along z axis. I had to reedit this project in Motion, inspite of AE looked better. Thank you all
While it is true that you can't have a footage layer that's more than 30,000 pixels, you can put the camera anywhere you want. Here's the keyframe data from a camera move that I just set up:
Adobe After Effects 8.0 Keyframe Data Units Per Second 24 Source Width 1920 Source Height 1080 Source Pixel Aspect Ratio 1 Comp Pixel Aspect Ratio 1 Transform Position Frame X pixels Y pixels Z pixels 0 960 540 -2666.67 96 960 540 -5e+09 End of Keyframe Data
See the data at frame 96? That's a z position of -5e + 9 or 5,000,000,000 pixels... You can't zoom back in a top comp view and see the camera, but you can easily position the camera there using the values in the time line.
It's easy to let a 3D stage get too complicated and end up with camera moves that are difficult to execute. There are some good practices that will prevent you from running out of room in the first place.
There's a good practice when working strictly in 2D that also applies to 3D. You shouldn't have a footage layer in your comp that's more than about 2 to 4 times the size of the composition. First of all, this unnecessarily ties up ram. Secondly, you end up with really soft wide shots because AE just doesn't handle scaling below about 50% very well. The footage turns soft.
The solution is to open up the original image in Photoshop then crop the image into sections that correspond to camera views at 3 or 4 key points in the move and save each as a copy. There's the original, a copy of the original, then there's a copy cropped to about 2/3 the original size, then a copy at 1/2 the original size, then one at 1/4 the original size. You then resize each of these copies to about 2X the width of your composition, add the appropriate sharpening to bring the quality back, then bring them into AE and start your move. It's really not that hard to match up the shots.
The same technique applies to 3D. When a composition sized layer is the same number of pixels from the camera as the zoom value you're seeing every pixel. It's the same as being at 100% scale. All you have to do to pull off a smooth move is stack the cropped and sized copies up in 3D space and cut between them. This means a lot less render time and a lot less movement with the camera with a higher quality end result.
The other good practice that helps keep 3D compositions under control is to use an appropriate lens. You wouldn't do a dolly move with a 1000 mm lens on a camera, yet there is sometimes a tendency to try and pull this off in the 3D world. The longer the lens, the farther you have to move the lens toward the subject to change the shot. Perspective isn't controlled by the focal length of the lens, it's controlled by camera position. The only thing the focal length does is crop the image (not counting barrel distortion common in "real" wide angle lenses).
I hope these suggestions help.
Thanx, Rick! In my case the lens was set up to obtain a "distance blur" effect, so it caused the layers to be 5000 px far from each other. May be that was not the right decision, or there are some other ways, but... Thank You anyway!
If the images needed to be 5000 pixels apart to get the depth of field effect you wanted then your aperture was too small or your lens was too wide and your layers too large.
Just as with a real camera depth of field depends on focal length and aperture. Unlike a real camera, you can set the aperture to any value you like so even a wide lens can have an extremely shallow depth of field. This image is a DV sized comp with 2 layers 400 pixels apart. The camera has a 50MM lens but the aperture is set to 500 pixels so there's a pronounced depth of field effect. Set the aperture to 2000 pixels and the back layer is so out of focus that it is nothing but a blob.
So the good practice that I failed to mention is to make effective use of your aperture and focal length.
Thank you Rick for detailed answer. In my case the comp size was 1920 by 1080, but the fotoes were too large, since they were scanned at 300 dpi, and i put them there, just cropped a bit. It was my first expirience of that kind. Now i see i've done some mistakes due tj lack of time. So i had to change my project from PC + AE to Mac + Motion. (I work both) Because Motion has no limit for scene dimensions. But i like AE best. Thank you once again!