I have footage of a light bulb swinging around in a circle at 60fps. I want to create a long shutter effect that connects and blends all of the frames into a smooth light trail.
The first step I did was to increase the temporal resolution by doing a time stretch of 500% and setting After Effects to interpolate new frames. This looks terrific and smooth, especially with something as simple as a lightbulb, but now my clip is around 11 minutes long.
Next I tried the Echo Effect, which is EXACTLY what I want to achieve, but will take far too long for me to render, since I needed to use 1000 echoes.
The Echo settings were:
Echo Time (seconds): -.0417 (which is 1 frame at 23.976fps)
Number of Echoes: 1000
Starting Intensity: 1
Decay: .995
Echo Operator: Maximum
Since this many echoes wasn't an option to render, I decided that perhaps I could recreate the effect manually by stacking layers, offset by 1 frame, then pre-rendering some of them together and repeating. My logic was that if I did 10 layers, rendered that out, then did 10 of those (which would be 100 "layers" at this point), rendered that, and finally 10 of those, I'd have the equivalent of 1000 echoes.
So I took a layer of the footage to represent the current frame and duplicated it. To create the decay for each, I wrote an expression that basically looks at the layer above (the previous echo) and sets its opacity to 99.5% of that layer's opacity, which is what the Decay ratio indicates within the Echo effect.
Here's the expression: thisComp.layer(index-1).opacity*.995;
I set the Blend Mode of all the layers to Lighten (or Lighter Color) which I believe is performing the same function at the Maximum operator in the Echo Effect.
For these 10 layers, the effects were IDENTICAL to the original Echo Effect. However, when I rendered out 10 layers, and tried to continue the effect, things started looking different.
The result were similar, but not satisfying!
One thing I realized was an issue was that my pre-comp of 10 layers was going from a range of 100% opacity (for the original frame) down to 96.069% opacity for the 10th layer, so when I re-imported this new file, I would need to adjust it to match what should be the opacity for the 11th layer, which would be 95.589%. But if I set the opacity of that 10 layer clip to 95.589%, I'm disproportionately affecting the opacity ratio between each layer within that clip. The lower I reduce the opacity, the less dynamic range and tonal definition within that clip...
I set up another simplified comp (a 100% white circle animating along a circular path) so that I could compare the differences between the Echo Effect and my method.
With the Echo Effect, you can see a very smooth tonality shift as the echoes decay (even on the wave... The top lines on the waveform represent the overlaps between circles (which admittedly seem brighter than they should be if the operator is simply choosing the Maximum value).
With my method, I create a composition of 10 circles. Within this comp, the Echo Effect and MY METHOD produce identical results. But in order to continue in this fashion, I would have to produce 1000 layers in a single comp, which wouldn't help me any in terms of rendering. In order to make it work, I'll need to render 10 layers at a time, then repeat that process twice more.
When I compare combining these layers with an Echo Effect of the same amount, things fall apart.
Here is the Echo Effect with 40 instances of the circle (1 original, plus 39 echoes).
And here is my method with two layers of a 20 Circle Pre-comp.
And here it is again with the second layer being properly adjusted by a subtraction layer to fall in... You can tell there is a significant loss in tonal range, as well as the overlapping regions do not retain their bright quality. Also, the steps between each echo's opacity are not smooth (or even) using my method, which only gets compounded by repeating. You can also see a distinct shift in tonality between each instance of 20 circles.
Can someone help me understand what is going on here? Is the Echo Effect doing something behind the scenes that I'm not accounting for? Why does it's waveform look like THIS, while my method looks like THIS?
Why are the overlapping regions brighter in the Echo effect? Isn't Maximum the same as Lighten? If not, how can I simulate an identical effect to the Maximum operator in the Echo Effect?
I'd really love to be able to make this work and need some expertise in understanding why these two things are not the same.
Thanks!
AdChoices