mahjqa has again provided us with more details on some of his animation techniques - this one as a prelude for turning it into a tutorial for the GN:
quote:(before you start, make sure you have Imageready with your PS. it is included since PS version 5.5 .)
It is the best thing to start with the image you want 'decayed' on one layer, with a transparent background. for best results, the image (text, whatever) should be twice the size of how big you'd want the final animation to be.
It's also a good idea to save often. (CTRL-S)
[introducing: the noise filter]
make a new layer, and fill it with black.
then click [filter>noise>add noise] set it to 400%. gaussian or uniform shouldn't matter, and set it to monochromatic. press OK.
whoo. nice. as if you didn't set up your TV set. If you'd zoom in you'd see a lot of black and white pixels, and some gray pixels in between. Noise simply gives a pixel a random brightness, so there should be about as much black as white pixels.
Now click [image>adjust>treshold]. You'll see some box with lines. Treshold makes pixels either 100% black or 100% white.when you move the slider, more pixels will become black or white (depending on their brightness) but for now it's okay to set it precisely in the middle. Press OK.
If you zoom in now, you'll see that every pixel is indeed pure black or white.
Delete the layer with the noise. now press Q. you'll enter what they call 'quickmask mode' to put it simple, it makes selections. If you whip out a drawing tool (say, an airsbrush) you can draw with black or white (or grey colors). now it you draw with black, you'll see a slightly transparent color appear (most likely red). now, if you press Q again, you'll exit quick mask mode. now everything that you painted black is unselected, and the rest is selected. Press Q once again to get back to quickmask mode again, and your colored spot will be back.
Now, repeat the process what you did with the layer you first created. fill it completely with black. Run the noise filter, then use treshold. Then press Q again.
You'll propably have a lot of blinking mess on the screen; 50% of the pixels is selected, 50% is not. Now move to the layer with your picture on it, and select one of the selection tools in the toolbar. right-click, somewhere in the middle of your picture, and a menu should come up. (you might lose your selection. if that happens, click [select>reselect], it gets back the last selection you had.) when you get the menu, click [layer via cut]. now, everything that was selected is cut away to a new layer.
you now have two layers, most likely named [layer1] and [layer2]. if you click on the eye icon next to them, you'll see that half of the pixels are on one layer, and half of them are on the other. You'll have to repeat this procedure (quick mask, noise, treshold, exit quickmask, layer via cut) again on layer 1 and layer 2, so you'll have layer 1 'til 4 when you're done, each with 1/4 of all pixels. repeat the procedure again, now you'll have layers 1-8. the whole deal of this is that the effect would look best if every pixel fell down separately, but but what we're gonna do is letting 1/8th of the pixels (so, one layer) at a time.
Now comes the tedious part, animating. click on the downmost button of the toolbar to go to imageready. It has an arrow on it. If you're not sure, hover with your mouse over it to see if you've got the right one.
This might also be a good time to save your work again.
Imageready has a layout much the same as photoshop. The main difference is that it has an animation palette. If you can't see it, click [window>animation] to make it appear. right now you can only see one frame, it's a miniatrue version of the picture you have at the moment. If you click at the 'new frame' icon (it looks just like the 'new layer' icon in the layers palette) you see it creates a new frame, which would be next frame in the animation. Imageready remebers for every frame the position and the opacity of a layer. If you change a layer however (IE; you draw on it) the layer will change in every frame. What we now need to do is to move the 8 layers down, and reduce the opacity in steps. Right now, you have two frames. select layer 1 in the layers palette, select the move tool, and press the down arrow on your keyboard once. layer one has now moved down one pixel. now, make a new frame in the animation palette. with the move tool and layer 1 still selected, press the down arrow twice. Also, move layer 2 down one pixel. For the next step, create a new frame, move layer 1 down three pixels, layer 2 two pixels, layer 3 one pixel. See the pattern here? All the layers start falling, first slowly, then faster. You'll still need to do the rest of the work yourself, though. next step; move layer 1 down four pixels, layer 2 three pixels, layer 3 two pixels, layer 4 one pixel. Also, make the opacity of layer 1 80%. next frame, 60%, next frame 40% , then 20%, then make it invisible.
Go on with every layer ontil all layers are invisible. Press the 'play' button in the animation palette, and marvel at your work.
Edit: the exporting stuff. Yes.
If you want a looping animation (so the sig falls apart and builds back up) you'll need to copy and reverse the frames.
select all the frames, and drag them to the 'new frame' icon. (all frames will be duplicated) Now click on the encircled arrow in the animation palette, (or an arrow button, it should give a dropdown menu) there should be an option 'reverse frames'.
If you're happy with the animation as it is now, specify a background color in the optimize window (if you can't see it, click [window>optimize]). It is called 'matte color' there. GIF doesn't support half-transparency, so a pixel is either transparent or not. the pixels that are not completely transparent are blended agains this color, so it shows up right on, say, a webpage. to save the animation, click [save>save optimized].
If you want to shrink down the animation to 50% of it's size (to make it look smoother or to preserve filesize) make your canvas a little bigger on the sides (click [image>canvas size]) and add a new layer above all others. now, make a border around the animation. the color or size doesn't matter, just make sure you can crop it out later on. (if you don't know the crop tool, refer to your manual or access the help files that came with your program) make sure the border is visible in all frames of the animation. Now, export it like you would export a regular animation (see above) save your PSD again, and close it. Now, open the GIF file of the animation. Every frame of the animation is now on a separate layer. The reason why I told you to make the border is this: if you shrink a layered image, sometimes inconsistencies appear. Right now every layer has a border, so every frame has the same heighth and width, so every frame will be resized in precisely the same way. Now, click [image>image size] and specify the size you want your shrunken image to be. 50% works best. Click [ok].
Grab te crop tool (or press C) and crop out the border, as we don't need it anymore.
Now, once again, specify the matte color, test your animation and save as optimized again. If you try to close the image, photoshop will ask if you want to save the file. If you click [yes] photoshop will save the layered PSD file of the animation.
