
Anyone who has finished a comp, hit the render button, and gone looking for H.264 in the output module list has met the same wall: After Effects simply does not write MP4 files on its own. That is not an oversight, and it is not a limitation you need third-party plugins to fix. It is a deliberate division of labor inside the Adobe ecosystem.
After Effects is a compositing and motion graphics application. Its render engine is designed to push uncompressed or lightly compressed frames into formats that preserve maximum latitude: QuickTime with ProRes, DNxHD and DNxHR, TIFF and PNG sequences, or the AE-native lossless codec. These formats are large, but they are also predictable. Every frame carries enough data that a colorist, an editor, or a VFX artist downstream can still push the image around without banding, blocking, or mosquito noise creeping in.
H.264, the codec that lives inside an MP4 container, is a delivery format. It discards information aggressively using spatial and temporal compression, and its quality depends heavily on rate-control decisions made across groups of frames. That is a job for an encoder, not a compositor. In practice, the split makes sense: After Effects decides what the image looks like, and the encoder decides how small it can be made without ruining it.
Use MP4 when the file is going straight to a viewer: a client review link, a website hero video, a social post, a slide deck, or an internal knowledge base. Use a mezzanine format when the file is going to another professional tool. If your MP4 will be re-edited, re-graded, or recompressed a second time, you are stacking two generations of lossy compression on top of each other, and the result will show it in the shadows and in any large flat gradient.
A simple rule: preview in MP4, edit and archive in ProRes or DNxHR, and only produce the MP4 at the very last step of the pipeline.
There are two workflows that consistently produce clean MP4s from After Effects, and they exist for different situations rather than competing with each other.
The fastest route for a one-off deliverable is to send the composition directly to a dedicated encoding application while the project is still open. The encoder reads the composition, renders it, and writes the MP4 in a single pass. Because it understands the project file, it can render straight from the comp rather than from a pre-rendered intermediate, which means one generation of compression instead of two, no giant temp file, and no guesswork about whether the master was written correctly.
For anything client-facing, for anything long, or for anything that might need a fix, render a master first. Export from After Effects as ProRes 422 HQ or DNxHR HQX, then transcode that master to MP4. This costs disk space and one extra step, but it buys you the ability to re-encode with different settings, cut a shorter version, or fix a typo in the end card without re-rendering the entire composition from scratch.
The direct handoff is where people lose time, not because it is complicated but because a few defaults quietly work against them.
Select the composition in the Project panel, then use the queue command. The comp appears as a job with its own output settings and its own file destination. Two things need checking immediately: that the correct composition was queued, not a precomp several levels deep inside the project, and that the output filename is meaningful. Naming files Comp 1.mp4 costs ten minutes every single time a client asks for a revision, because nobody can tell which file is which.
Every queued job has two halves that people constantly confuse. Render settings control which frames get rendered: the work area, a specific frame range, resolution, and frame rate. Output settings control the container, codec, bitrate, and audio. Mixing them up is the source of the classic complaints about cropped exports and half-resolution files. If a comp is 3840x2160 but the render settings are set to half resolution, you will get a 1920x1080 file that looks soft, and you will blame the codec when the codec did exactly what it was told.
If you regularly export the same kind of deliverable, say thirty 15-second social cuts, set up a watch folder. Drop exported masters into a folder and the encoder picks them up automatically, applies a saved preset, and writes finished MP4s to a destination folder. This turns a repetitive manual task into a background process, and it is the single biggest time saver in a high-volume motion graphics pipeline. The trick is to save the preset once, verify it on a test file, and then never edit it mid-project.
The default H.264 preset is built for compatibility, not for quality. A handful of adjustments changes the output dramatically.
Use High profile for anything beyond the most basic needs, since it enables more efficient compression tools. The level should match your resolution and frame rate, and the encoder will warn you if you choose one that cannot carry the stream. Match Source is usually the safe choice, and it removes one more decision from a long list.
Variable bitrate spends more data on complex frames and less on simple ones. For most web delivery, VBR with two passes and a target around 10 to 16 Mbps at 1080p, and 40 to 60 Mbps at 4K, produces an excellent result. Constant bitrate exists mainly for streaming and hardware-constrained playback, where predictable data rates matter more than peak quality. Unconstrained VBR is best reserved for archival work where file size is irrelevant.
These are starting points, not laws. Grain, fast motion, confetti, and hard cuts all consume bitrate. A talking head at 8 Mbps can look cleaner than a particle-heavy title sequence at 30 Mbps, because the second one gives the encoder far more to describe.
AAC at 320 kbps, 48 kHz, stereo is a safe default for web delivery. If the source audio was recorded at 44.1 kHz, resampling is usually harmless, but keeping sample rates consistent across a project avoids tiny sync drifts in long timelines. For anything driven by dialogue, listen to the exported file on headphones rather than trusting the timeline preview, because the codec can do surprising things to sibilance and room tone.
Understanding the relationship between bitrate and file size removes the guesswork about whether a file will fit an upload limit.
File size in megabytes is approximately bitrate in Mbps multiplied by duration in seconds, divided by eight. A 60-second clip at 12 Mbps is roughly 90 MB; the same clip at 50 Mbps is roughly 375 MB. Video and audio bitrates add together, so a 320 kbps audio track adds about 2.4 MB per minute on top of the video.
That arithmetic tells you two useful things. First, if a platform enforces a hard file size limit, you can solve for the bitrate you can actually afford. Second, if you are uploading somewhere that will re-encode your file anyway, pushing an enormous bitrate is mostly wasted effort that only slows the upload. Most platforms transcode to their own ladder, so the goal is to hand them a clean, artifact-free source rather than the largest possible number.
Two-pass encoding analyzes the whole clip before writing it, allocating data where it matters most. It roughly doubles encode time and costs nothing in quality terms. For short deliverables, always use two-pass. For hour-long footage, one-pass VBR at a generous target is a reasonable compromise between time and polish.
Modern GPUs can encode H.264 far faster than a CPU can. The trade-off is efficiency: hardware encoders typically need a higher bitrate to match the visual quality of a strong software encoder, especially in dark scenes, grainy footage, and broad gradients. If your comp is full of subtle gradients or simulated film grain, software encoding at a moderate bitrate often beats hardware encoding at the same setting. If you are producing dozens of internal cuts against a deadline measured in hours, hardware is the pragmatic choice and nobody will notice the difference on a phone screen.
Three technical details cause most of the "the export looks wrong" messages that land in a motion designer's inbox.
After Effects is time-based rather than frame-based, so a 23.976 fps comp placed into a 29.97 fps timeline will be conformed, not duplicated, and motion will look subtly wrong. Decide the delivery frame rate before you animate. Converting 60 fps footage to 24 fps requires either frame blending or optical flow, and neither is free: both can produce warping on text, logos, and any thin vertical line that moves horizontally.
Almost nothing delivered today should be interlaced. Set field order to Progressive and verify it in the output settings before you render. If you inherit an interlaced source, deinterlace it deliberately with a proper method rather than letting the encoder guess, or you will see combing on every frame with horizontal motion.
This is the most common complaint about MP4 exports: the file looks washed out, flat, or darker than the timeline. The cause is usually a mismatch between the working color space, the export metadata, and the player used to view the result. Tagging the output with the correct color primaries and transfer characteristics, then reviewing in a player that respects those tags, prevents most surprises. Do not start adding contrast inside the comp until you have confirmed whether the problem lives in the file or in the viewer.
Upload the highest practical quality, typically a clean 4K or 1080p master with moderate grain, and let the platform build its own encoding ladder. Avoid heavy sharpening, because platform compression amplifies it into visible halos.
Deliver native 1080x1920 rather than cropping a landscape export, which throws away more than half your pixels. Safe zones matter: keep text and logos away from the outer edges where interface overlays sit. Keep the bitrate generous, because phone screens are viewed close-up and compression artifacts are far more obvious at that distance.
Never deliver H.264 to broadcast. Deliver ProRes or an equivalent mezzanine and keep the MP4 as a review copy. For archival, store both the project file with collected assets and the master render, since an MP4 alone is a delivery artifact, not an archive.
Precompose and prerender particle systems, 3D layers, and anything with expressions that recalculate on every frame. A pre-render turns hours of repeated calculation into minutes, and it makes the final render dramatically more stable.
Give After Effects a generous disk cache on a fast SSD and clear it before a big render, since a corrupted cache is a common source of inexplicable slowdowns. Close other memory-hungry applications. Purging before a final render is cheap insurance.
Rendering a ten-minute comp as five two-minute comps and concatenating them afterward is faster, more crash-resilient, and easier to revise. It also lets you render overnight in segments and resume after a failure instead of losing the entire job.
Export stalls at zero percent. Usually a missing linked asset, an offline font, or an effect that depends on a plugin you no longer have installed. Check the comp for missing-footage warnings before touching any settings.
Audio and video drift apart. Mixed sample rates or a variable frame rate screen recording in the source. Conform the source to a constant frame rate before importing it.
The file plays in one player but not another. Container and profile mismatch. Re-export with High profile and a standard level, and confirm the audio codec is AAC.
Colors look wrong only in a browser. The browser is applying a different color transform than the reference player. Compare in a dedicated player before changing your export settings, because changing the export to fix a browser quirk usually breaks the file everywhere else.
Renders are far slower than expected. Check whether hardware encoding is enabled, whether the disk cache sits on a slow drive, and whether motion blur or depth of field is set higher than the shot actually needs.
The output is letterboxed or cropped unexpectedly. The composition size and the output size do not match. Set the output to Match Source, or resize the comp deliberately.
File size balloons far beyond expectations. Unconstrained VBR with a very high ceiling, or a very long duration. Switch to two-pass VBR with a defined target and maximum.
Can After Effects export MP4 directly? Not with the built-in output module. Use a dedicated encoder or render a master and transcode it.
Is H.264 good enough for a client deliverable? Yes for review and final web delivery, provided the bitrate is sensible and the source is clean.
Should I export at the comp frame rate or the delivery frame rate? The delivery frame rate. Decide before you animate, not after.
What is the best bitrate for 1080p? Around 10 to 12 Mbps at 24 to 30 fps and 16 to 20 Mbps at 60 fps, adjusted upward for grain and fast motion.
Why does my MP4 look softer than the preview? Usually insufficient bitrate on detailed footage, or render settings left at half resolution.
Can I queue multiple compositions at once? Yes. Select them all, queue them, and set output paths and filenames individually before starting the batch.
Does two-pass encoding really matter? For short clips it costs little time and provides more consistent quality, so it is worth leaving on.
Before every render, confirm the composition, the work area, the frame rate, and the resolution. Choose the path: direct handoff or master plus transcode. Set the codec, profile, rate control, and target bitrate. Set the audio codec and sample rate. Pick a meaningful filename and a destination that will not be overwritten. Render. Then watch the actual output file end to end on the device your audience will use.
That last step is the one people skip, and it is the only one that reliably catches the problems no settings panel can show you.

Alexander
Guida pratica al montaggio video rapido con AI: transizioni fluide, software, workflow, coerenza visiva, errori da evitare e FAQ operative.
A practical guide to planning, generating, editing, and finishing AI-assisted videos with consistent characters, clear shot lists, and reliable review loops.
Build a repeatable AI video post-production pipeline: organize assets, cut story-first, design transitions, refine audio, and review versions calmly.