[Darby Edelen]

Less lower saturation? Starting the reduction later? The complaint you have is that the color channels are being clipped differently so the only solution is to make the color channels more uniform, translation: desaturation. You might try using the Tint effect for a slightly different result. I’d also look into desaturating before the levels effect and and after the levels effect to see which you prefer.

Darby Edelen

[Darby Edelen]

I would recommend keyframing a reduction in the saturation as the white level is adjusted. There are other options but that would be my first thought.

Darby Edelen

[Darby Edelen]

When you say ‘between the most recent two keyframes’ do you mean the maximum speed that is reached at any point in between those keyframes or the maximum speed at one of those two keyframes?

Darby Edelen

[Darby Edelen]

Just because a technique doesn’t do exactly what you want doesn’t mean it can’t be useful if modified to fit your needs. You obviously know what you need better than I do which is why I wasn’t initially prescribing an exact solution. I don’t know exactly what your requirements are but the outline for a solution that I described may still be useful to you if you’re flexible in the way you think about it.

For example: once you have Time Difference setup in a meaningful way you could use a sampleImage() expression to sample the average difference between frames. You could then convert that sampleImage() expression to keyframes and then view the value graph in the Graph Editor to see a representation of differences across time through the entire clip. In fact, that may be more useful than markers as it can provide information about the amount of motion across time instead of a binary ‘motion/no motion’ marker.

Darby Edelen

[Darby Edelen]

You’d need an expression on the audio levels. Something like this:

min = -64;
transition = 0.1;
s = Math.abs(timeRemap.speed - 1);
linear(s, 0, Math.max(0.001, transition), [value, value], [min, min]);

This should lower the audio levels from the keyframed values to -64 dB as the speed of the timeRemap property changes from 1 to 0.9 or 1.1.

You can change the target audio level by changing the ‘min’ variable in the expression. If you want to make the drop in volume less sudden you’d increase the value of the ‘transition’ variable. If you want it more sudden you can decrease it.

I haven’t tested this expression, writing from memory on my mobile so let me know if anything doesn’t work.

Darby Edelen

[Darby Edelen]

You’ll definitely want to make sure that the First Vertex of each path keyframe is aligned as the first vertex at each keyframe will always move to the next keyframe’s first vertex.

Beyond that you might look into the Window > Mask Interpolation tool. This can give you more pleasing transitions from shape to shape at the cost of flexibility. It creates a keyframe for every frame of the mask path animation.

Darby Edelen

[Darby Edelen]

After Effects isn’t particularly well suited for this sort of thing, but you could use the Time > Time Difference effect to get the difference between the current frame and the next (or previous) frame and then adjust those results to create something that could work for your purposes.

I’d apply this expression to the Time Difference’s Time Offset property:

source.frameDuration
So that it will find the difference between the current and next frames. I would also enable ‘Absolute Difference’ to get a more meaningful result for your purposes. After that I’d apply a Gaussian Blur to reduce the affect that noise and small changes might have on your result. I’d then apply a Levels effect and increase the Input Black to remove those small differences and decrease the Input White to boost the more apparent differences.

There are a couple of other things I’d probably do but let’s see if the above is at all helpful before we dive any farther in 🙂

Darby Edelen

[Darby Edelen]

It’s just another transform that gets concatenated with the other transforms. Which one you use depends entirely on what you’re trying to achieve.

There’s a Position property for the procedural path types like Rectangle, Ellipse and Polygon. Above that there’s usually at least one ‘Group’ Transform as well where the Position, Anchor Point, Rotation, Scale, Skew and Opacity can be adjusted. Whenever you ‘Group’ paths you’ll get another one of these transforms. Then at the highest level there is the standard Layer Transform.

These are applied hierarchically. For example, if you have multiple paths within a single Group and you want them to move independently then you’ll need to animate each path’s Position property. If there are multiple Groups and you want all of the paths in that single Group to be transformed together then you’d modify that Group’s Position property. If you wanted all paths on the layer to be transformed together then you’d modify the Layer Position property.

Darby Edelen

[Darby Edelen]

First I’d recommend purging your disk cache, just in case there are some frames hanging around in there. Edit > Purge > All Memory & Disk Cache.

You could try using the Temporal Smoothing feature of the Auto Color and Auto Levels effect in case what you’re seeing is the result of changing color levels in the source footage. Unfortunately, that tends to slow the effect down significantly.

If changing color levels in the source is the problem then I’d suggest going with the other suggested solution of using the (non-auto) Levels effect.

Darby Edelen

[Darby Edelen]

Here’s an expression you could apply to an Expression Selector on a Text Animator that drops Opacity to 0%.

if(numKeys > 0){
nk = nearestKey(time);
i = nk.index;
if(nk.time > time) i--;
100 * Math.max(0,textIndex-i)
}
else 100;

You can then add keyframes on the Amount property of the Expression Selector. The first word will reveal itself at the first keyframe, the second at the second keyframe, etc… It’s not exactly what you’re asking for in that you’d still need to do a fair amount of clicking to add and move keyframes but hopefully this will simplify things a bit for you and may give you some other ideas of how to approach it. The keyframes you add aren’t used for anything but timing so no need to change their values.

You could also do this with layer markers but I find those more difficult to manage (you can’t select multiple markers and move them together). The one upside of layer markers is that you can add them with the layer selected by tapping the * key on the numpad as you preview audio. If you’re fast on that asterisk then you might be able to rough out the timing for an entire subtitle in realtime and only using the keyboard. If that’s a worthy enough trade off then this is the modified expression to look for layer markers:

markers = thisLayer.marker;
if(markers.numKeys > 0){
nk = thisLayer.marker.nearestKey(time);
i = nk.index;
if(nk.time > time) i--;
100 * Math.max(0,textIndex-i)
}
else 100;

Darby Edelen