SoundForgePro Official 15-day trial

Mastering

Audio Export QA Checklist: Catch Problems Before Delivery

Guide navigation Table of contents Sections
    Stereo waveform passing through level, phase and listening checks before final audio delivery

    An export is not finished when the progress bar disappears. The delivered file can still have the wrong channel layout, a clipped encode, missing tail, stale metadata or a perfectly valid format that the recipient did not ask for. This checklist separates the creative approval from the technical handoff.

    1. Freeze the delivery specification

    Write the destination requirements before opening the render dialog. “High quality WAV” is not a specification. Record the container, codec, sample rate, bit depth or bitrate, channel layout, loudness target, peak ceiling, filename pattern and metadata requirements.

    Audio export moving through edit, level, channel, listening and delivery checks
    Approve the rendered file itself: verify the edit, meters, channel behavior, reopened export and final delivery copy.
    FieldExample onlyVerify against
    Container and codecWAV PCM or AAC in MP4Client, platform or broadcaster specification
    Sample rate48 kHz for many video workflowsThe destination and source project
    Bit depth / bitrate24-bit PCM master; delivery bitrate set by platformRequired master and distribution copies
    ChannelsMono, stereo or named multichannel layoutPlayback route and delivery contract
    Loudness and peakProgram-specific LUFS/LKFS and true-peak ceilingCurrent destination standard
    MetadataTitle, artist, episode, ISRC or BWF fieldsRelease sheet or asset manifest

    Targets vary. EBU R 128 uses −23 LUFS as its programme loudness reference for European broadcast workflows, while web platforms publish their own encoding recommendations. YouTube's current upload guidance, for example, recommends 48 kHz audio and AAC-LC or Opus in supported video containers. Neither example should be copied into an unrelated delivery.

    2. Preserve a clean master

    • Save a new project or file version before final processing.
    • Keep a lossless master at the native production sample rate and an appropriate bit depth.
    • Create distribution encodes from that master instead of repeatedly transcoding a lossy file.
    • Apply dither only when reducing bit depth, and avoid applying it again to an already dithered master.

    3. Check the edit before rendering

    • Confirm the start is clean and intentional. Look for a clipped first transient or too much pre-roll.
    • Listen through the final fade, reverb and ambience tail. Make sure the selection does not cut it short.
    • Check edits, crossfades, mute regions and processing bypass states.
    • Confirm markers, regions and channel routing if the export depends on them.
    • Bypass temporary monitoring correction that should not print into the file.

    4. Check level, loudness and headroom

    Peak normalization and loudness normalization solve different problems. Peak normalization moves the highest sample to a chosen ceiling. Loudness normalization changes gain according to an average programme measurement. Sound Forge's documentation also warns that compressed formats can create peaks near or above the original ceiling during encoding, which is a reason to leave margin rather than aiming blindly at 0 dBFS.

    • Measure the full programme, not a highlighted excerpt.
    • Check sample peaks and true peaks when the destination specifies them.
    • Confirm the loudness measurement mode and gating match the requested standard.
    • Listen after any limiter, normalization or codec stage for pumping, distortion and softened transients.
    • Do not raise a quiet master simply to make the waveform look larger.

    5. Verify channel layout and polarity

    Open the rendered file and inspect it as a new asset. Confirm that mono has not become dual mono by accident, left and right are not swapped, and a multichannel file follows the expected order. For stereo material, check both channels separately and perform a mono compatibility listen if the destination may sum them.

    The left/right speaker test helps verify the playback path. It does not replace checking the channel metadata and waveform inside the delivered file.

    6. Reopen the actual export

    Do not approve the timeline and assume the render matches it. Close the render dialog, import or open the output file, then listen to the beginning, several transitions, the loudest section and the end. Check duration, silence, fades, sample rate, bit depth or codec, channels and metadata from the exported asset itself.

    7. Test the delivery copy

    • Play it in a second application or on the destination device when practical.
    • For a lossy encode, compare it with the lossless master at matched loudness.
    • Confirm the filename, extension and version number are unambiguous.
    • Check that album, episode or reel assets use the same naming pattern.
    • Create a checksum when files move through a formal handoff or archive.
    • Keep the master, delivery file and notes together so the export can be reproduced.

    A compact sign-off block

    CheckPass condition
    SpecificationFormat, sample rate, depth/bitrate, channels and targets match the written requirement
    ContentCorrect version, complete beginning and tail, no accidental mute or bypass
    LevelRequested loudness and peak limits verified on the full export
    ChannelsLayout, order, polarity and mono behavior are intentional
    MetadataRequired fields present and stale fields removed
    PlaybackRendered file reopens and plays in a second path
    HandoffFilename, checksum and accompanying notes identify the approved version

    Sources and scope

    The numeric examples above are destination-specific references, not universal mastering targets. Confirm current requirements in the EBU R 128 documents, YouTube encoding guidance, the receiving platform's specification, and Sound Forge's normalization documentation.