After Effects Export Guide: ProRes, H.264, and Clean Deliverables

Why Your Export Settings Decide How Good Your Edit Looks

You can spend a week on a composition and lose most of that work in the last ten minutes. Export is where an edit either survives intact or quietly degrades — banding in gradients, mushy grain, chattering motion, audio drifting a few frames out of sync. None of these problems are visible in the timeline. They only appear after compression, and by then the file may already be on a client's desk.

The core problem is that delivery formats are lossy by design. H.264 and H.265 are built to throw away detail the eye is unlikely to notice, and they make that judgement frame by frame under a fixed bitrate budget. When your source is clean and your settings are sensible, the result looks great. When your source is a heavily compressed screen recording, or your bitrate is too low for the motion, or your color pipeline was never set up properly, the encoder discards the wrong things.

The fix is not a secret preset. It is a repeatable pipeline with two distinct outputs: a high-quality master that preserves everything, and a lean delivery file tuned for each destination. This guide walks through codec selection, render settings that matter, encoder troubleshooting, and a platform checklist you can reuse on every project.

The Two-File Rule: Master First, Delivery Second

Almost every export headache traces back to skipping the master file. Editors try to render directly to a small H.264 file and then discover the client wants a slightly different crop, a textless version, or a version with a different end card. Re-rendering from the compressed file guarantees generational quality loss.

A master file is an intermediate, high-bitrate export that is visually indistinguishable from your composition. It is large, it is not meant for upload, and it exists so every downstream version can be made from the best available source. From one master you can create:

A platform upload in H.264 at the right resolution and bitrate.
A vertical crop for short-form feeds.
A textless or alternate-audio version.
A broadcast or festival file with different color tagging.
Frame grabs and thumbnails at full resolution.

A practical workflow looks like this: finish your composition, export a master at full resolution with no scaling applied, archive that file with the project, then use a dedicated encoder — Adobe Media Encoder, HandBrake, or FFmpeg — to produce delivery versions from the master. That last step decouples creative work from distribution demands. When a client asks for a new aspect ratio, you are transcoding, not re-rendering.

Name your files so this system stays usable. Something like projectname_master_prores_4444_v03.mov and projectname_delivery_h264_1080p_v03.mp4 tells you at a glance which file is which, and version numbers prevent the classic mistake of delivering an old render.

Choosing the Right Master Codec

ProRes 422 HQ and ProRes 4444

ProRes is the default professional intermediate for good reason: it is fast to encode, broadly supported on macOS and in most post tools, and visually transparent at high enough data rates. ProRes 422 HQ is the workhorse for standard footage — plenty of headroom for grading, comfortable file sizes, and reliable playback. ProRes 4444 is the right pick when you need an alpha channel (motion graphics, logo stings, overlays) or when you are preserving source material with heavy grading potential. ProRes 4444 XQ exists for the extreme cases where every bit matters, but for most projects the file size penalty is not worth it.

One caveat: ProRes 422 is chroma subsampled. If your composition contains fine colored text, thin saturated lines, or heavy keying work, 4444 avoids the small color smearing that subsampling can introduce.

DNxHR for Windows-Centric Workflows

DNxHR is the Avid-family equivalent and behaves admirably across platforms. If your team is Windows-first and your finishing tools handle DNxHR cleanly, use it. The quality tiers map roughly onto ProRes: SQ for lightweight work, HQ and HQX for standard finishing, 4444 for alpha and maximum fidelity. Mixing ProRes and DNxHR in one pipeline is possible but adds unnecessary transcode steps, so pick one for the whole project.

Image Sequences and Lossless Archives

When a shot absolutely cannot lose a pixel — VFX handoffs, restoration, or material headed to a colorist who will push it hard — an image sequence is the safest option. PNG or TIFF sequences are frame-independent, so a corrupt file costs you one frame instead of the whole render. TIFF at 16-bit is heavy but effectively archival. The tradeoff is that sequences are awkward to review and take forever to upload, so they belong in the archive, not in the delivery folder.

Matching Codec to Source Material

Here is the rule most people miss: never use a heavy lossy codec as your master. If your source is AI-generated video, screen capture, or phone footage that already went through compression, exporting to another lossy format compounds the damage. Go up, not sideways. Transcode compressed sources into a high-bitrate intermediate and do all further work there.

Render Settings That Quietly Ruin Quality

Bit Depth and Color Management

Set your project to the highest bit depth your workflow tolerates — 16-bit or 32-bit float per channel — before you export. Eight-bit processing creates visible banding in skies, gradients, and glow effects, and no export setting can recover information the composition never carried. Equally important, decide your color pipeline early: if you are working in a wide-gamut space, make sure the master carries the correct color primaries and transfer characteristics, and that your delivery export converts rather than strips them. Untagged files get interpreted differently by every platform, which is how edits end up looking washed out after upload.

Resolution, Scaling, and Field Order

Export at the full resolution of your composition. Scaling up in the output module does not add detail and scaling down loses it. If you must resize, do it once in an intermediate step with a decent resampling filter rather than combining it with compression in a single pass. For interlaced broadcast deliverables, set field order explicitly and verify it — a wrong field order produces a combing artifact that is obvious on any movement. For everything web-facing, progressive is correct.

Alpha Channels and Edge Quality

If your deliverable needs transparency, confirm both the codec and the output module support it. ProRes 4444 and DNxHR 4444 handle alpha; H.264 does not, despite what some presets suggest. Straight versus premultiplied alpha matters too: graphics with soft edges look correct when premultiplied against the intended background, but hard-edged or keyed elements sometimes need the straight interpretation. The failure mode is a thin light or dark fringe around your subject, which is easy to spot against a contrasting background but easy to miss on a busy one. Always test against two backgrounds.

Motion Blur, Frame Blending, and Effects That Break

Some effects behave differently at render time than in preview. Check motion blur settings, verify that any frame blending you relied on is enabled in the render, and be careful with effects that depend on the composition's current time. Third-party plugins can also render differently in the background than in the interactive viewer — if a plugin has a draft/production quality toggle, set it to production before the final pass.

Delivery Encoding: H.264, H.265, and Platform Targets

Bitrate Ladders That Survive Re-Encoding

Every platform re-encodes your upload. That second compression is where quality really gets decided, and the best defense is to give the platform a clean, generously encoded file. As a starting point for 1080p H.264 at 24–30 frames per second, a target bitrate of 12–16 Mbps with a maximum around 20 Mbps looks good on nearly any service. For 4K at 30 fps, 45–60 Mbps is a sensible range; for 4K at 60 fps, push toward 60–80 Mbps. Fast-moving content, grain, confetti, and rain need more headroom than slow dialogue.

Use two-pass encoding when the tool offers it — the encoder allocates bits more intelligently across the whole file. Keep your keyframe interval around two seconds, use the High profile at level 4.2 or above for 1080p, and avoid dropping below 8 Mbps for anything with motion. If your file is too large for an upload limit, lower the bitrate rather than lowering the resolution; a crisp 1080p file beats a soft 4K one on almost any display.

H.265 and When It Is Worth It

H.265 (HEVC) delivers comparable quality to H.264 at roughly half the bitrate, which is genuinely useful for high-resolution uploads and for archiving delivery files. The tradeoffs are compatibility — older devices and some web players struggle with it — and encode time. If you control playback (an app, a kiosk, a private page), H.265 is a clear win. If a general audience is watching, H.264 remains the safe default, with H.265 as the archival or premium variant.

Audio Settings People Forget

Video quality gets all the attention, but audio problems are more noticeable. Use AAC at 320 kbps for stereo delivery, and keep 48 kHz sample rate end to end — mixing 44.1 kHz source audio into a 48 kHz timeline without resampling is a classic source of subtle pitch issues. Enable loudness normalization if the platform expects it, but never rely on the platform to fix a mix. Check that your master includes a safety margin below clipping and that your delivery file has the right channel layout: stereo for web, 5.1 only when the destination explicitly supports it. A mono track exported to a stereo container without duplication plays in one ear.

Frame Rate Decisions: 24, 30, or 60

Frame rate is a creative decision with technical consequences, and it is easiest to resolve it before you animate anything.

24 fps reads as cinematic. It is right for narrative, mood pieces, and anything with film-like motion blur. Avoid it for crisp UI animation or screen content.
30 fps is the safe generalist. It matches most screen content, works for social, and sits comfortably in both 24-heavy and 60-heavy ecosystems.
60 fps suits gameplay, sports, fast camera moves, and any deliverable that will be slowed down in a later edit. It doubles data rate and render time.

Export at your composition's native frame rate whenever possible. Converting between rates requires motion interpolation, which invents frames and produces artifacts on complex motion. If you genuinely need a different rate, do it as a separate, deliberate step from the master and review the result frame by frame at the moments with the most movement.

Troubleshooting Common Export Failures

Render Queue Stalls, Timeouts, and Memory Errors

A render that dies at 60 percent is usually a resource problem, not a bug. Purge the disk cache, empty the media cache database, and check that your scratch disk has several times the expected output size free. Long compositions with heavy effects can exhaust memory; render in segments using the work area, then stitch the parts together, or pre-render complex nested compositions into intermediates. If a specific frame consistently kills the render, isolate it with a short work area and inspect the layers active at that moment — one bad plugin invocation is a common culprit.

Codec Not Found and Silent Export Failures

"Codec not found" or an output module that quietly reverts to a default format usually means a missing component: a 64-bit versus 32-bit mismatch, a plugin that is not installed on the render machine, or a checkpoint build that removed a legacy codec. Render through a dedicated encoder rather than the internal queue when possible, and export to a format you control rather than relying on a format list that may change between versions.

Black Frames, Missing Audio, and Out-of-Sync Exports

A black frame at the start or end is often a composition duration set slightly longer than the content. Trim the work area to actual content. Missing audio usually means the track was muted, soloed elsewhere, or routed to an output that the module does not include — check the audio output mapping, not just the mix levels.

Audio drift on long exports tends to come from sample rate mismatches or variable frame rate sources. Convert variable frame rate footage to constant frame rate before the final export, and keep one sample rate throughout. If drift persists, export the audio separately as a WAV file and mux it with the video in your encoder; this removes any chance of the video pipeline touching the timing.

Banding, Blocking, and Over-Compression

Visible blocking or macroblocking in dark areas means your bitrate is too low for the content, or your source was already compressed and you are asking the encoder to preserve artifacts. Banding in gradients usually begins as an 8-bit color problem and gets worse with compression. Fix the source bit depth first, then raise the bitrate, then consider adding a very subtle dither or grain layer — a tiny amount of noise gives the encoder something to chew on and measurably reduces banding in flat areas.

A Platform Deliverable Checklist

Before any upload, run through this list:

Container and codec — MP4 with H.264 video and AAC audio is the universally safe combination.
Resolution — match the platform's preferred size; never upscale a smaller master to hit a number.
Frame rate — native to your composition, constant, not variable.
Bitrate — generous enough to survive re-encoding; check the platform's published recommendation and go slightly above it.
Color — correct primaries, transfer function, and matrix tagged in the file.
Audio — 48 kHz, 320 kbps AAC, correct channel layout, no clipping, sensible loudness.
First three seconds — clean frame, no black leader, hook visible immediately.
Aspect ratio — verify any letterboxing or cropping was intentional, not accidental.
File name — descriptive, versioned, no spaces if your pipeline dislikes them.
Playback test — watch the exported file end to end on a phone, a laptop, and a TV if possible.

That final test catches more problems than any settings panel ever will.

Common Mistakes and How to Avoid Them

The most expensive mistakes are the boring ones. Exporting straight to H.264 and deleting the master. Relying on a preset name instead of checking the actual bitrate and color tags. Assuming the platform will fix your audio. Rendering once and uploading without watching the result. Forgetting that vertical crops need their own composition rather than a scaled-down landscape export.

Less obvious but equally damaging: over-sharpening before compression (the encoder amplifies edge artifacts), adding heavy grain to a low-bitrate delivery (the encoder either destroys it or burns all its bits on it), and stacking multiple lossy generations when a single high-quality intermediate would have kept everything intact.

Build the two-file habit, verify your color and audio once per project rather than once per export, and keep a checklist taped to the wall. Most export quality problems are solved before you ever hit render — they are decided by how you set up the composition and what you choose to preserve.

FAQ

Should I export directly from After Effects or through a separate encoder?
Use After Effects to produce a high-quality master, then use a dedicated encoder for delivery files. Dedicated encoders give you better bitrate control, two-pass encoding, and batch processing across multiple versions.

Is ProRes worth the file size?
Yes, as a master. A ProRes master lets you create every downstream deliverable without another generation of compression, which matters most when the client requests changes later.

Why does my exported video look worse than the preview?
Almost always one of three causes: insufficient bitrate for the amount of motion, an 8-bit color pipeline causing banding, or a platform re-encode compressing your upload further. Fix the source first, then raise the bitrate.

Can I export H.264 with transparency?
No. H.264 does not carry an alpha channel. For transparent deliverables, use ProRes 4444, DNxHR 4444, or a PNG image sequence.

How do I stop audio drift on long exports?
Convert variable frame rate sources to constant frame rate, keep a single 48 kHz sample rate throughout, and if drift persists, export the audio as WAV and combine it with the video during encoding.

What is the safest delivery format for an unknown platform?
An MP4 container with H.264 High profile video, AAC audio at 320 kbps, progressive scan, and 1080p resolution. It plays essentially everywhere and uploads cleanly.