What Is an AAF File? The Picture-to-Sound-Mix Handoff, Explained

A re-recording mixer opens a new Pro Tools session and finds eleven mono tracks, two stereo masters, and not one single clip where the editor swore the dialogue was. The editor had sent a single flattened video file and called it the final cut. It was not. An AAF file is a container that hands a locked picture edit's full track structure, clip in and out points, and audio references to a sound mixer so the mix session matches the timeline exactly, instead of forcing the mixer to rebuild it by ear. We have seen what happens when that handoff goes wrong, and it is never a fast Friday.

AAF Definition

What is an AAF file, exactly?

An AAF, short for Advanced Authoring Format, is a structured project file, not a video file. Where an exported MP4 or MOV is a single flattened picture, an AAF carries the sequence itself: every track, every clip's in and out point, basic effects and transitions, and either embedded audio or pointers to the original audio media. Open one in Pro Tools and you see the same multi-track layout the picture editor built in the NLE, ready to mix, not a stereo file waiting to be pulled apart.

That is why AAF is the standard handoff from picture to sound. A news package, a branded documentary, a scripted short: whenever a separate mixer takes over after picture lock, AAF is usually how the session arrives. It is also one of the two or three formats, alongside XML and EDL, that editors reach for when a project needs to leave one application and still make sense in another.

AAF Origin

Where did the AAF format come from?

AAF was published in April 1998 by a group of production toolmakers working together under what was then the Multimedia Task Force. Microsoft's own announcement named Avid, Adobe and Digidesign among the companies releasing the specification together, and Avid led the consortium that developed it. The goal was specific: give post production tools a way to exchange a structured, multi-track project without flattening everything down to a single rendered file first. Before AAF, nothing did that job well across vendors. The specification now lives with the Advanced Media Workflow Association, which also stewards several of the other interchange formats editors run into day to day.

Worth sitting with for a second: before AAF and OMF existed, the handoff from picture to sound was often a literal multitrack audio tape, synced by hand to a work print or video reference. Anyone doing a documentary mix in the 1980s was matching tape counters to timecode burned into a window dub. AAF is the reason that stopped being normal.

AAF Vs OMF

How is AAF different from Avid's older OMF format?

AAF effectively replaced OMF (Open Media Framework Interchange), which Avid used for the same picture to sound handoff before AAF existed, and OMF's design fed directly into how AAF was built. The practical reason post houses moved on is a hard ceiling: OMF tops out at a 2GB file size, a limit Avid's own community forums still field questions about, and that ceiling gets hit fast on a documentary with a dozen audio tracks and embedded media. AAF removed that cap and added richer structural metadata. Most modern facilities default to AAF now and only reach for OMF when an older system on the other end can't read anything newer.

AAF In Practice

What does an AAF handoff actually look like on a real job?

Picture a 22 minute documentary segment, picture locked in DaVinci Resolve, carrying eight audio tracks: two interview mics, a layer of ambient room tone, three music cues, and two sound effects beds. The editor exports an AAF. That file carries every track, every clip's precise in and out point, and references back to the original audio media. The mixer opens it in Pro Tools and sees the identical track layout the editor built, already separated by source, ready to clean up dialogue and balance levels. Nobody is guessing which of the six people talking is on which channel of a collapsed stereo file.

Here is that export in practice, from a locked Resolve timeline into a Pro Tools session a mixer can actually work in:

That difference shows up directly on an invoice. A mixer who has to rebuild track separation from a flattened stereo file is billing for detective work, not mixing. A mixer who opens a clean AAF is billing for the mix itself. On a documentary with a real sound budget, that gap is the difference between finishing in a day and needing three.

Format Comparison

AAF, EDL or XML: which one actually fits the job?

These three formats get confused constantly because they all move a sequence somewhere else, but they are not interchangeable. The table below is the short version; if the EDL side of this needs more depth, this breakdown of what an EDL carries and where it still gets used covers it on its own.

AAF Vs Stems

Why not just send WAV stems instead of an AAF?

Stems (separated audio files per channel or per element, like dialogue, music and effects) solve a narrower problem than AAF does. A folder of stems gives a mixer isolated audio, but not the edit: no in and out points, no track order, no sync relationship to picture unless everyone agrees to rebuild it by hand against a reference video. That rebuild is exactly the kind of guesswork AAF exists to remove. Stems still matter, plenty of final delivery specs require them, but as a mid-process handoff from a locked picture edit to a first mix pass, an AAF gets a mixer into a working session in minutes instead of hours. We'd reach for stems at delivery and AAF at handoff, not the other way around.

Common AAF Mistakes

Where do AAF exports usually go wrong?

The biggest one is timing: exporting the AAF before picture is genuinely locked, then re-exporting after the mixer has already started building their session, which overwrites their work with a new structure they have to reconcile from scratch. AAF handoff should happen once, after lock, not as a rough draft. The second trap is embedded versus referenced audio. An AAF that only references external audio files breaks the moment those files move or get renamed; embedding the audio avoids that but produces a much heavier file to send. The third is mismatched frame rates between the picture edit and the mix session, which causes sync drift that is painfully obvious on dialogue and easy to miss entirely on a music bed until it is too late to fix cheaply.

AAF And The Pre-Edit

Where does AAF fit next to a paper edit and a transcript?

AAF is the last export in a chain, and for anything built around people talking, that chain usually starts with a transcript, not a timeline. An editor working from interview footage typically starts with a paper edit: a transcript with the strongest lines marked and arranged into a narrative before anyone touches an NLE. ScriptCut sits at that front end. Selections marked on a word-level transcript become a frame-accurate AI Select, AI Story and AI Arrange pass, and the resulting timeline exports as FCP 7 XML into DaVinci Resolve, Premiere Pro, Final Cut Pro or Avid, through the same panel and export workflow that sends a cut to the NLE in the first place. That XML handoff has no AAF step in it at all; the picture edit gets built and locked in the NLE first, and only then does that locked sequence get its own, separate AAF export to the mix stage, built and sent entirely from inside the NLE. Same general idea as exporting a paper edit into Final Cut Pro or Avid, two different jobs, two different formats, one chain.

In that XML handoff, every sound channel of a recording keeps its own track, so a stereo mix or a pair of lav mics arrives in the NLE already separated, which matters later when that same locked sequence heads off to a mixer as an AAF.

AAF And NLE Compatibility

Does every editing tool open an AAF the same way?

Mostly, but not perfectly. Avid Media Composer and Pro Tools, built by the same company and part of the original consortium, read AAF most completely, including track structure and some effects metadata. Premiere Pro and DaVinci Resolve both import and export AAF for audio handoff, but effects and transitions sometimes need rebuilding, which is a different problem than the one Resolve's text-based transcript tools solve. The safest habit, on any system, is a short test export before the real handoff: three minutes of the actual sequence, opened on the receiving end, checked for sync and track order, before sending the full cut.

Common Mistakes

What should you double-check before an editor hits send on a sound mix handoff?

• Confirm picture is actually locked, not "locked unless notes come back," before exporting.
• Decide embed versus reference audio before export, and tell the mixer which one they're getting.
• Match frame rate and sample rate between the picture session and the mix session.
• Send a short test AAF first on any new pairing of NLE and DAW.
• Keep the original locked timeline untouched after export, in case a reconform is ever needed.

For tightening captions and burned-in text on a finished cut before it ever reaches the mix stage, the free SRT to VTT converter handles the subtitle side without opening an NLE.

If this handoff keeps eating a day of every schedule, here is how we'd start: build and lock the story first inside ScriptCut, send the approved cut to the NLE as XML, finish picture there, and only then export the AAF the mixer actually needs. Fixing the chain upstream of the AAF saves more time than fixing the AAF itself ever will.

Sources

• Microsoft: Advanced Authoring Format specification release, 1998
• Advanced Media Workflow Association (AMWA)
• Advanced Authoring Format overview, Wikipedia
• Avid Community: OMF file size limit
• Microsoft 1998 release announcing AAF