Time Remapping: The Decisions That Matter

Time remapping is the technique of changing the speed of a composition or clip over time: slowing a moment down, speeding through another, holding a single frame, or running a section backward, all driven by a curve you can edit rather than by a single fixed speed setting. Under the hood, every remapping tool does the same thing. It maps playback time, the frame the viewer is watching, to source time, the frame of the original animation or footage that gets shown at that instant. Change that mapping and you change the rhythm of the piece without touching the animation inside it.

That makes time remapping one of the most useful tools in motion graphics, and one of the easiest to misuse. It lets a studio reuse finished animation in new ways: a cutdown that runs tighter, a hero moment that lingers, a freeze that gives a caption room to breathe. It also exposes whatever the source was hiding. Slow down a sequence far enough and the viewer starts to see the gaps between frames, the stepped timing of hand-drawn work and the smear of frame blending trying to invent motion that was never rendered. Marketers approving a late speed change, producers trying to squeeze a 30-second spot into 15 seconds, and animators deciding whether to remap or re-animate all face the same set of decisions.

Rather than list those decisions in the abstract, this article follows one illustrative project from brief to delivery. The project is a composite built for teaching purposes, not a client story, and its numbers are realistic but invented. At each stage we explain what was decided, why, and which trap it avoided, so you can apply the same reasoning to your own work.

The Illustrative Project: A Launch Film That Needed a New Tempo

Here is the setup. A software company commissions a 45-second product launch animation: flat vector illustration, kinetic typography, a short character sequence of a user working at a desk, and a logo reveal at the end. It is built in After Effects at 1920 by 1080, 24 frames per second, with the character sequence drawn frame by frame on twos, meaning each drawing is held for two frames, which gives 12 unique drawings per second. The 45-second master is approved on schedule.

Two weeks after approval, the client comes back with three requests. First, a 15-second cutdown for paid social that keeps the same story beats. Second, a dramatic slow-motion moment in the master, where the product interface "snaps" into place, to be stretched from one second to four seconds. Third, a freeze on the character's reaction shot so a new testimonial-style caption can sit on screen long enough to read. None of these requests changes the design. All of them change time.

This is the classic brief for time remapping, and also the classic moment when projects go wrong, because timing changes arriving late in a project tend to be treated as small edits when they can touch every layer in the build.

Those four numbers already frame the problem. The slowdown needs 96 output frames from a source that has 24. The character freeze needs a hold, not a slowdown. The cutdown has to lose 720 frames without the story feeling rushed. Each request calls for a different technique, and the first real decision is recognizing that.

How Time Remapping Actually Works Before You Touch a Keyframe

Every remapping decision in the project rests on a clear mental model of what the tools do, so it is worth setting that out before walking through the work.

The mapping curve

In After Effects, enabling time remapping on a layer (Layer, then Time, then Enable Time Remapping) adds a Time Remap property with two keyframes: one at the layer's first frame holding source time zero, and one at its last frame holding the source's final time. Between them, the value rises in a straight line, which is normal speed. Every change you make is a change to that line. A steeper slope plays faster than normal, a shallower slope plays slower, a flat segment is a freeze, and a downward slope plays in reverse. The value graph in the Graph Editor shows this literally: the vertical axis is source time, the horizontal axis is composition time. Adobe's After Effects User Guide documents the Time Remap property alongside the related Time Stretch, Time Reverse Layer and Freeze Frame commands.

DaVinci Resolve uses the same idea with different vocabulary. Its retime controls let you set speed points along a clip and ramp between them, and its retime curve exposes the frame mapping as an editable graph. The DaVinci Resolve Reference Manual from Blackmagic Design covers speed changes, retime curves and the choice of retime processing, which is where Resolve decides how to fill frames that do not exist in the source.

Time remapping versus time stretch

Time stretch applies a single constant speed to a whole layer. It is fine for "make this 10 percent faster" and useless for "hold here, then ramp." Time remapping is the tool whenever the speed needs to change within the layer, whenever you need a freeze, or whenever you want to reverse a segment and then play forward again. In the project, all three requests involve speed that varies across the timeline, so time stretch was never the answer.

Where the in-between frames come from

This is the part most people skip. When you slow a source down, the output needs more frames than the source contains, and something has to decide what goes in the gaps. There are three possibilities. The tool can repeat the nearest source frame, which keeps images crisp but creates visible stepping. It can blend neighboring frames, which smooths playback at the cost of ghosting. Or it can estimate motion between frames and synthesize new pixels, which can look remarkably smooth or remarkably broken depending on the content. In After Effects these correspond to frame blending off, Frame Mix and Pixel Motion; in Resolve they appear as retime process options such as Nearest, Frame Blend and Optical Flow.

There is a fourth possibility that only applies to animation built inside the compositing tool: if the remapped layer is a precomposition made of keyframed vector layers, the software can evaluate that animation at the in-between moments directly, because keyframe interpolation is a continuous function, not a set of stored frames. That distinction between rendered pixels and live keyframes drives most of the decisions in this walkthrough.

Day One of the Change Request: Triage Before Touching the Timeline

The temptation with a late timing change is to open the project and start dragging keyframes. The studio in this example started with a spreadsheet instead, listing each request, the layers it touched, the technique it implied and the risk it carried. That triage took about an hour and saved several days.

Three insights came out of the triage. The slowdown was really two problems, because the glow burst used a particle simulation that had been cached and imported as a rendered pass, so it behaved like footage. The freeze was straightforward, but only if it landed on a key drawing rather than an in-between. And the cutdown, which looked like the biggest time remapping job, turned out not to be a remapping job at all for most of its length.

The Slow-Motion Moment: Remapping a Precomp Instead of the Layers Inside It

The interface snap was the showpiece of the change request, so it got the most care. In the master it lasted one second, 24 frames: three interface panels slide in from off-screen, overshoot slightly, and lock into a grid while a cursor clicks and a glow burst radiates outward. The client wanted it stretched to four seconds, with the slow section ramping in and out rather than cutting abruptly to slow motion.

Step one: isolate the moment

The snap already lived in its own precomposition, because the original build followed a disciplined project structure. That mattered: time remapping applies to a layer, and if the snap had been spread across twenty layers in the main composition, each would have needed identical remapping, or the team would have had to precompose it late and risk breaking parenting and expressions. The logic behind that kind of structure is covered in our guide to precomposing and project structure in motion projects, and this project is a good example of why it pays off later.

Step two: shape the ramp, not just the speed

A four-times slowdown applied instantly looks like a mistake. The team designed the curve in three parts: normal speed approaching the snap, a ramp down to 25 percent speed across about eight frames, the slow section itself, then a ramp back up to normal speed across about twelve frames so the logo transition did not feel sluggish. On the time remap value graph, that reads as a line that flattens smoothly, runs shallow, then steepens again. Bezier handles on the speed graph control how gradual those ramps feel, and the same principles that govern property animation apply here. If you want the underlying theory, our article on easing and timing curves explains why a gradual deceleration reads as weight while an abrupt one reads as a glitch.

Step three: account for the added duration

The slowdown added three seconds to the master, turning 45 seconds into 48. Every layer after the snap in the main composition had to shift three seconds later: the logo reveal, the end card, the music edit point and the legal line. The team extended the snap precomp's layer duration in the main timeline, then moved everything downstream as a block. The music was the harder part. The composer's track had a hit on the snap, so the audio editor rebuilt the bars around the slow section rather than stretching audio, which would have shifted pitch or introduced artifacts.

Step four: check what the slowdown revealed

Because the panels, cursor and grid were live keyframed shape layers, the remapped precomp evaluated them at in-between moments and the slow section looked smooth. The glow burst did not. It had been rendered as a 24-frame image sequence, so at 25 percent speed each particle frame was on screen for four output frames. Played back, it stepped visibly, a stutter that was especially obvious because everything around it moved smoothly.

That choice follows directly from the reference guidance: for extreme slowdowns, re-render the source animation rather than asking the software to invent frames. Frame blending was tested first and rejected. Frame Mix produced soft double images of each particle, which looked like motion blur applied in the wrong direction. Pixel Motion did better on the large shapes but tore the small particles into smeared fragments, because motion estimation struggles with many small elements moving in different directions. A 20-minute re-cache beat both.

The Freeze Frame: Holding the Right Drawing for the Caption

The second request was a freeze on the character's reaction, a moment where the user at the desk looks up and smiles, so that a new caption could sit on screen for two and a half seconds. This is the textbook use of time remapping: a freeze holds a moment without duplicating layers or exporting a still, and it keeps the frame live in the project so later changes to the character precomp still flow through.

Choosing the frame to hold

In a sequence animated on twos, not every frame is equally good as a hold. Key drawings, the strongest poses, are designed to be read. In-between drawings are designed to be seen for a twelfth of a second while in motion, and often look awkward when frozen: a half-closed eye, a mouth mid-shape, a hand smeared for speed. The animator scrubbed through the reaction and picked the key drawing where the smile peaks, then set the hold on the first frame of that drawing's two-frame pair so the freeze landed cleanly.

Building the hold

With time remapping enabled on the character precomp, the team placed a keyframe at the chosen source time, a second keyframe 60 frames later holding the same value, and then let the curve resume at normal speed. Two and a half seconds at 24 fps is 60 frames. Both keyframes use hold interpolation on the way into the freeze so that no interpolation creeps in. After Effects also offers a Freeze Frame command that enables time remapping and holds the current frame for the whole layer, which is useful for a permanent still, but a mid-layer hold is cleaner when you want the animation to continue after the freeze.

Keeping the caption out of the freeze

The new caption had to animate in and out while the character was frozen. Had it been placed inside the character precomp, it would have frozen too. So the caption was built as its own layer in the main composition, above the remapped precomp, animating on normal time. This is a small point that catches people constantly: anything inside a remapped precomp inherits its timing, so design elements that must move on their own clock belong outside it.

The caption also needed to be readable in the time available. The team set type size, weight and on-screen duration using the same checks described in our piece on text legibility in motion, and the two-and-a-half-second hold was chosen by timing a slow read of the caption aloud and adding a margin, not by guessing.

The 15-Second Cutdown: When Re-Timing Keyframes Beats Remapping

The cutdown was where the project's most important decision lay, and it was a decision not to use time remapping for most of the work. It sounds paradoxical in an article about remapping, but the reference guidance is explicit on this point: remapping instead of re-timing keyframes is a mistake, and the cutdown shows why.

Why remapping the whole master fails

Speeding the master to three times speed would have given a 15-second film in minutes. It would also have been unwatchable. Every ease would be compressed to a third of its duration, every hold on a line of text cut to a third, every character pose flashed past. Remapping changes when source frames appear; it cannot change the internal proportions of the animation. A text line that needed 1.5 seconds to read still needs 1.5 seconds, not half a second.

What the team did instead

The cutdown was treated as an edit first and a re-time second. The team chose five of the master's nine scenes, keeping the problem statement, the interface snap, one feature, the character reaction and the logo. Within those scenes, animators went into each precomp and re-timed keyframes directly: tightening entrance animations from 18 frames to 12, shortening holds between typographic beats, and removing secondary motion that the shorter format could not support. The easing curves were kept in proportion, so motion still felt designed rather than hurried.

Because the master was built from reusable precomps with shared controls, the cutdown versions of those scenes were duplicates with adjusted timing rather than rebuilt animation. Teams that produce many versions of the same spot should think about this from the start, and our guide to reusing rigs and templates explains how to set up a build so that variants like this are cheap.

Where remapping still earned its place

Remapping was used in the cutdown in exactly two places. The interface snap was included without the slow-motion section, played at normal speed, which meant using the original precomp rather than the remapped one. And the transitions between the five scenes were built with short speed ramps: the last eight frames of each outgoing scene accelerated to roughly 200 percent before a cut, a common trick that lets a scene exit with energy instead of waiting for its full animation to complete. Those ramps were short enough, and applied to live keyframed content, that no artifacts appeared.

Frame Rates, Frame Blending and the Settings That Decide Quality

Midway through the project, an issue surfaced that had nothing to do with the client's requests. The two rendered 3D product shots had been delivered by a 3D artist at 30 fps, and imported into the 24 fps master without anyone checking. At normal speed, After Effects was quietly skipping and repeating frames to fit 30 into 24, and a slight judder on the rotating product had been approved as "fine." Once one of those shots sat near the slow-motion section and was slowed to half speed as part of the ramp, the judder became obvious.

Check the source frame rate before any speed change

Every remapping calculation depends on knowing how many unique frames per second the source actually contains. In this project there were four answers in one composition: vector animation with effectively unlimited temporal resolution, a character sequence with 12 unique drawings per second, a particle pass rendered at 24, and 3D renders at 30. Keeping source frame rates in mind is one of the reference practices for good reason. The team interpreted the 3D footage correctly, conformed it deliberately, and requested a re-render at 24 fps for the shots near the ramp so that the math was clean.

When frame blending helps

Frame blending smooths playback at changed speeds, and it is genuinely useful in the right place. The team used Frame Mix on a rendered background texture loop that was slowed modestly, to about 80 percent, where the blend was invisible and removed a slight stutter. They avoided it on the character sequence entirely, because blending two hand-drawn frames produces a double image that destroys the line quality of the drawing. And they avoided Pixel Motion on the particles for the reasons described above.

Two practical points matter in After Effects. Frame blending has to be enabled on the layer and also enabled for the composition, or it will not appear in previews or renders; forgetting the composition switch is a common source of "it worked in my preview but not in the export" confusion. And Pixel Motion is considerably slower to render than Frame Mix, which matters when you are planning a delivery schedule, a subject covered in depth in our practical guide to planning render time.

The table above reflects the studio's working rules for this project, not universal thresholds. The point is the pattern: the lower the number of unique frames in the source, the less slowdown it will tolerate before you need new frames rather than smarter interpolation.

Expressions, Audio and the Hidden Dependencies of a Remapped Layer

Time remapping changes more than pictures. In the launch film, three hidden dependencies surfaced during the change request, and each is worth knowing about before you start.

Expressions that read time

Several layers inside the interface precomp used expressions driven by time: a subtle wiggle on the cursor, and a looping pulse on a notification badge. Expressions evaluate against the time of the composition they sit in, and inside a remapped precomp that is the remapped source time. So during the slow section, the wiggle slowed down with everything else, which was correct. But one expression used a frame-based calculation to trigger a flash every 12 frames, and at quarter speed the flash timing drifted in a way that looked like a bug. The fix was to rewrite that expression in seconds rather than frames and check it again at the remapped speed.

Audio on remapped layers

Remapping a layer that carries audio will also remap the audio, with results that are rarely usable: pitch changes, stretched transients, and for reversed sections, backward sound. In this project the precomps carried no audio, because sound design lived on its own tracks in the main composition. That separation is the right default. If a remapped layer does contain audio, disable it and rebuild the sound for the new timing, as the audio editor did for the snap.

Layer duration and the last frame

When you extend a remapped layer beyond its original length, most tools hold the last source frame. That is useful for a deliberate end hold and dangerous when it happens by accident. In the project, the first render of the slowed master had a precomp whose remapped curve reached the end of its source a few frames early, leaving a frozen interface for eight frames before the next scene. It was only caught because the team watched the render at full resolution in real time, rather than scrubbing.

That near miss led to a small addition to every remapping task in the project: after changing any time remap curve, confirm the final keyframe lands at or just before the end of the source, and confirm no layer downstream inherits an accidental hold. It took seconds per shot and caught two more issues before delivery.

Previewing and Reviewing Speed Changes Without Fooling Yourself

Unpreviewed speed changes are on the list of mistakes to avoid, and the reason is that low-resolution previews lie about exactly the problems time remapping creates. A preview at half or quarter resolution softens the image, which hides the ghosting of frame blending and the tearing of motion estimation. A preview that cannot play in real time drops frames, which makes a smooth ramp look stuttery or a stuttery one look acceptable.

How the team reviewed

Every remapped section was previewed at full resolution and full quality, cached until it played back in real time, and watched at least twice: once looking at the whole frame and once looking only at the fastest-moving or smallest elements, because that is where artifacts live. For the slow-motion snap, the animator also stepped through the section frame by frame, looking specifically for any frame that did not change from the one before it, a sign of repeated source frames.

Client review happened on a rendered file, not a screen recording or a browser preview of the project. Review platforms that transcode uploads can introduce their own frame rate conversions, so the team uploaded a file at the delivery frame rate and noted in the review message that the slow section should be judged on smooth playback, not on a paused frame.

What to look for

• Repeated frames in slow sections, visible as brief stalls or stepping.
• Double images or ghosting on edges, the signature of frame mixing.
• Warping, tearing or smeared detail around small or overlapping elements, the signature of motion estimation.
• Ramps that start or end abruptly, which read as a dropped frame rather than a design choice.
• Text that is on screen for less time than it takes to read aloud slowly.
• Accidental holds at the end of remapped layers.

The Delivery Timeline and What It Cost in Time

Here is how the change request unfolded, day by day, in this illustrative example. The schedule assumes one senior animator, a character animator for part of one day, an audio editor for half a day, and a producer coordinating.

Five working days for what the client described as "a couple of timing tweaks" is typical once you count the dependencies. The time was not spent dragging keyframes. It was spent on triage, on re-rendering sources that could not survive the slowdown, on audio, and on review. A team that skipped triage and remapped everything would have delivered faster and then spent the same time again fixing artifacts the client spotted.

If your organization produces many versions of the same animation, for instance seasonal edits or regional variants, it is worth building timing flexibility in from the start. Our article on seasonal campaign variants covers how to plan a master so that later versions do not trigger this kind of scramble.

Remap, Re-Time or Re-Render: The Decision Framework the Project Produced

Stripped of its details, the launch film project produced a simple framework that applies to most time remapping decisions. It starts with one question: does the change alter the rhythm of existing motion, or the internal timing of the motion itself?

Use time remapping when

• You need a freeze, a reversal, or a speed ramp across an existing section.
• The source is live keyframed animation in a precomp, or footage with enough frames for the speed you want.
• The change applies to a whole moment and all its contents should share the new timing.
• You want the change to stay editable and non-destructive, so later changes to the source still flow through.

Re-time keyframes when

• The overall duration is changing substantially, as in a cutdown.
• Different elements within a moment need different amounts of change: text holds that must stay long, entrances that can tighten.
• Easing must keep its designed proportions.
• Readability sets a minimum duration that a uniform speed-up would violate.

Re-render or re-animate when

• The slowdown is extreme and the source is baked at a fixed frame rate.
• The source is frame-by-frame drawn animation, where blending damages the drawing.
• Frame blending and motion estimation both produce visible artifacts at full-quality preview.
• Frame rates in the composition are mixed and the slowed section includes the mismatched source.

The framework also clarifies when outside help is worth it. Timing changes that arrive late in a project are exactly when an experienced motion team earns its fee, because the risk is not in the remapping itself but in the dependencies it touches: audio, expressions, source frame rates, and rendered passes that need regenerating. If your in-house team is facing a late change like this, the animation and motion graphics service is built to handle revisions of this kind alongside new production.

Practical Settings and Habits to Carry Into Your Next Project

A few habits from the project are worth making standard practice, whether you work in After Effects, Resolve or another tool.

Build for remapping from the start

Keep distinct moments in their own precomps, keep audio off animation layers, and write time-based expressions in seconds rather than frames. None of this costs anything during the original build and all of it makes later timing changes safer.

Record source frame rates in the project

Label every imported sequence or rendered pass with its native frame rate in the layer name or a project note. When a slowdown request arrives, you can see immediately which layers will hold up and which will not.

Leave headroom on key moments

If a moment is likely to be featured in slow motion later, such as a product reveal or logo resolve, consider rendering any simulation or 3D passes for that moment at a higher frame rate from the outset. Rendering a short section at two or four times the delivery rate is often cheap insurance, though the render cost depends on the complexity of the pass.

Keep ramps proportionate

Ramps into and out of slow motion need enough frames to read as intentional. In the project, eight frames in and twelve out worked at 24 fps for a 4x slowdown. Shorter ramps looked like glitches; much longer ones diluted the moment. These are starting points to test, not rules.

Treat review as part of the technique

A remap is not finished until it has been watched at full resolution, in real time, by someone looking for artifacts. That applies to your own work and to anything you hand to a client. For more on the craft decisions surrounding techniques like this, browse the full set of motion graphics and animation articles.

Where this comes from

• Adobe Help Center — After Effects User Guide
• Blackmagic Design — DaVinci Resolve Reference Manual

The figures and practices above come from the sources listed.

Working on something like this?

We take on Motion Graphics & Animation work for teams who want it done once, properly. Tell us what you are building and we will tell you honestly whether we are the right studio for it. Start a project.

Where to go next

• A Practical Guide to Isometric Animation
• Blending Graphics with Live Footage, Done Properly