.png)
.jpeg)
Audio post production is the set of tasks that turn rough location or production audio into a polished, deliverable soundtrack: clear dialogue, believable sound effects and ambience, music balance and a final mix. This short playbook gives you a fast, task-oriented workflow for Premiere Pro and DaVinci Resolve users, with practical checks, time budgets and handoff notes so you can stop hunting libraries and start shipping client-ready sound.
Audio post-production is the process of refining and finalising all the sound that sits under moving images. It takes the raw production tracks and shapes them into dialogue that reads clearly, sound effects that sell movement and location, ambient beds that create space, and music that supports emotion without fighting the voice.
Immediate benefits for editors
• Dialogue clarity, so the viewer understands every line without cranking volume.
• Emotional storytelling, where ambience and music underline mood.
• Fewer client rounds, because technical problems are resolved before review.
• Deliverables that integrate cleanly with picture edits, keeping deadlines intact.
Typical final deliverables
• Dialogue stems or edited dialogue tracks, useful for localisation and mixers.
• SFX beds and grouped categories like footsteps, crowds and transitions.
• Ambience tracks for scene continuity and room tone healing.
• Final mix or reference mix, plus stems (dialogue, SFX, music).
• Game-ready exports where required: one-shots and adaptive music stems for middleware.
Clients see and hear cleaner picture sync, intelligible VO in noisy locations, believable footsteps and prop hits, and ambience that feels like the scene’s space. Music sits in the right emotional pocket, not louder than the conversation. Those are the moments clients call back for.
Export dialogue stems with handles and timecode, an SFX bed split into categories, an ambience bed, and a stereo reference mix. Include a clicktrack or guide track if the editor or composer needs it. Label everything with scene and take numbers plus frame rate so your files slot straight into other timelines.
This numbered workflow prioritises speed and repeatability. Treat each phase as a time-boxed task where the objective is a usable, unambiguous result for the next step.
Play through the rough cut and flag scenes by priority: high for close up dialogue or FX-heavy beats, low for background shots and cutaways. Use markers in Premiere or Resolve to note problems: noisy room tone, mouth clicks, missing props. If a scene only needs level tweaks or an ambience fill, mark it for quick fixes rather than full post.
Time budget tip, for a typical short: spend 10 to 20 percent of total post time on spotting, less for short-form content. Prioritise scenes that will make or break comprehension or emotion.
Order your dialogue work: sync the audio to picture, cut breathy gaps and extraneous noise, de-clip any peaks, run denoise if needed, then apply subtractive EQ to remove rumble or boxiness. In Premiere use Merge Clips or Manual Sync and the Essential Sound panel for quick denoise and EQ. In Resolve, use Fairlight’s clip gain, Noise Reduction and the EQ presets, riding levels with automation for performance nuance.
Fast wins include clip gain for consistent levels before any plugin processing, handheld low-cut filters for all tracks under 150 Hz, and a gentle de-esser for sibilant lines. Keep a copy of raw audio and export intermediate stems so fixes are reversible.
Choose ADR when the original performance is unusable, the line is critical to story or emotional detail is lost, and editing or noise reduction cannot recover it. Quick decision checklist: can you fix intelligibility with EQ and reduction within 30 minutes? If not, schedule ADR.
For quick ADR, prepare a reference take sheet, provide picture in-sync, and record several editorial-quality takes. Label takes with scene, slate and timecode. If ADR is marginally required, consider clever editing or library replacement for non-vital lines to save time.
This checklist assumes a single-day finish on a short sequence or a polished scene. Time allocations are estimates; adjust to project scale.
Organise your session, load a template with tracks named Dialogue, Foley, Ambience, SFX, Music. Import the EDL or XML from the video edit and confirm frame rate and timecode. Drop markers for problem lines and important picture hits. Templates and naming save minutes that add up fast.
1. Sync and rough trim dialogue, 30 to 60 minutes.
2. Clip gain and basic de-click or de-clip, 30 minutes.
3. Denoise and EQ on dialogue stems, 30 to 60 minutes.
4. Critical SFX and Foley hits (footsteps, prop sync), 60 to 120 minutes. Use a small, organised library and layer hits for realism.
5. Ambience bed and continuity patches, 30 minutes.
6. Quick balance, panning and automation for intelligibility, 30 to 45 minutes.
7. Loudness check and final export for review, 15 to 30 minutes.
Must-dos are dialogue intelligibility, sync accuracy, and a steady ambience bed. Nice-to-have items are fancy micro Foley layers or complex sound design that can wait until the next pass.
Export quick review mixes as stereo MP3s or WAVs with timecode burned in for notes. For handoff to a mixer, export stems at 48 kHz 24-bit with 1 to 2 seconds handles and include a reference mix.
Workflows should reduce friction between edit and mix. Use standard formats and consistent labelling to avoid wasted time.
Decide whether to export stems or full tracks. Export stems when multiple audio sources are mixed together or when the mixer needs consolidated groups, such as Dialogue stem, SFX stem, Music stem. Export full tracks when you want the mixer to have each on separate channels. Use AAF or OMF for multitrack exchange where your DAW supports it, or export WAV stems named with scene_shot_track_take_timecode. Include handles for flexibility and keep frame rate metadata intact.
For Premiere, use Export AAF via Adobe Media Encoder or round-trip using XML into Resolve or Pro Tools. For Resolve, use the Deliver page to export stems or use AAF/OMF exports from Fairlight.
Keep category libraries sorted: footsteps, clothes, foley props, whooshes, impacts and ambiences. Use procedural tools or batch processing to generate pitch/key variations of whooshes or crowd ambiences quickly. Layer three to four samples instead of hunting for a single perfect clip. Tools that let you perform sounds in real time can speed creative iterations and reduce library browsing time.
Krotos-style tools that allow real-time performance and quick variation can be a shortcut when you need bespoke SFX without a long edit session. Keep each SFX as an editable stem so you can tweak later.
For Unreal or Unity, export one-shots and multi-variant WAVs with clear naming: event_context_variant_take. Provide both long and trimmed versions if the engine needs loop points. For FMOD and Wwise, export organised banks or folder structures that map to events and parameters. Use consistent bit depth and sample rate across assets, commonly 48 kHz and 24-bit, and include short documentation for trigger conditions and randomness parameters.
AI-assisted tools can accelerate tasks like generating atmospheres, expanding libraries and creating fast variations. Use them where speed matters, but keep clear records and controls.
When to use AI and when to record
• Use AI for variations, non-critical ambience and to jump-start creative ideas.
• Prefer human recording for character performances, unique Foley moments and anything requiring legal or ethical certainty.
Verification and licensing checks
• Keep a changelog for assets created or modified by AI, including prompts and parameters.
• Verify licences, whether assets are from libraries, procedural tools or AI generators. Keep written confirmation if assets will be redistributed or sold.
Approvals workflow
• Disclose use of generative tools to stakeholders when it affects creative ownership or licensing.
• Deliver editable stems and source files so clients can approve or request changes. This keeps creative control and avoids legal exposure.
Document sources and processing steps, keep editable stems, note any AI prompts used, and provide clients with clear options: recorded replacement or retained AI-generated asset. Use versioning and keep a clean archive of raw audio.
Do not pass AI-only assets off as recorded performance without disclosure. Avoid assuming unlimited licensing for generative content. Don’t skip a human-quality check, especially on dialogue, lead performances and anything that represents an identifiable person.
Before you deliver, confirm provenance, have a signed delivery spec if in doubt, and keep a simple client approval step that lists any generative tools used.
Try this in your next session: use templates, versioned stems and a short release note that says what was generated and what was recorded. It keeps everyone aligned and reduces rework.
Ready to try a hands-on workflow? Download a fast-start session template, grab a free trial of Krotos tools to speed SFX performance and join the Krotos community forum to swap presets and workflow tips with other creators. A short trial and a tidy template often saves more time than a weekend of library hunting.
Roles vary by project size. On small projects one person may cover dialogue editing, Foley, SFX, and mixing. Larger projects have specialists: dialogue editors, Foley recordists and editors, sound designers, re-recording mixers and supervising sound editors. Producers or post supervisors coordinate deliverables and client notes. Each role focuses on a specific part of the chain so work is faster and errors are easier to reverse.
ADR, or automated dialogue replacement, is recording dialogue in a studio to replace or improve production audio. It is about capturing a voice performance that matches picture and emotional intent. Foley is the recording and syncing of everyday sounds such as footsteps, clothes rustle and prop handling, performed to picture for realism. ADR replaces spoken lines, Foley creates or enhances physical sound effects.
The canonical five are dialogue, Foley, sound effects, ambience and music. Dialogue carries the story and must be intelligible. Foley covers bespoke physical sounds tied to on-screen action. Sound effects include designed impacts, whooshes and crowd cues. Ambience provides location and continuity. Music supports mood and pacing. Together these elements form a balanced soundtrack that reads with the picture.