sequenced.ai
Articles/Voice & video/Blueprint//7 min read

Auphonic gives podcast editors an adjustable first pass at a finished mix

Auphonic can align separate microphones, balance speakers and deliver to a loudness target. Credit rules and editable gain curves shape the production decision.

By Sequenced deskAI-assisted, source-led · how we work
Visit Auphonic website ↗
Sync TracksAlignmentCombines clues about sound, conversation and recording times.
Gain pointsLevel editingRefine the local volume change applied to a segment.
True peakDeliverySet a peak limit alongside a loudness target.
Time creditsBilling unitProcessing duration determines the charge.
Auphonic mark
Auphonicauphonic.com · independent research

Represent this company? Verify your work email to access its workspace, or send the desk a factual correction.

The guest recorded in a quiet study. The host recorded in a kitchen, leaned away from the microphone twice and started their recorder before the guest. Both files sound usable by themselves. Played together, the question lands on top of the answer, the room tone jumps between speakers, and the guest disappears whenever the listener turns the volume down. Auphonic is an audio post-production service for exactly this stretch between recording an interview and delivering a mix people can follow.

In brief
  1. 01Starting material Bring an existing recording or individual microphone files from a session.
  2. 02Editorial control Inspect the alignment and change the leveler’s decisions where emphasis is part of the story.
  3. 03Cost driver Estimate the duration submitted for processing, including intros, instead of multiplying the show length by its number of microphones.

01 / ProductFrom separate voices to a deliverable mix

Auphonic’s production overview covers podcasts, videos, educational recordings and audiobooks. Its tools address problems at different points in the production process. An alignment correction helps put two speakers back into the same conversation. A leveler adjusts relative loudness when a voice grows or fades. Delivery normalization aims the resulting program at the loudness specification requested by a publisher. Each acts on a different property of the recording.

A file can hit a loudness target while one speaker remains hard to understand. For example, the music under a quiet guest may contribute enough energy to the mix that the overall program measures as expected. Turning everything up would also increase the music. The relevant intervention is to rebalance those elements, or to revisit the original recording if the speech is already masked. Keeping separate stems gives an editor better evidence to make that decision.

The algorithm descriptions distinguish speech and music handling, noise and reverb reduction, multitrack adjustments and true peak control. For a weekly talk show, choose settings around the sound the audience expects. A convincing documentary may need some of the crowd or room noise that a studio interview would lose. Before processing a long series, compare a busy field segment with a clean studio exchange through the same proposed preset.

02 / AudienceWho benefits from a repeatable finishing pass?

An editor servicing a network of short weekly shows may spend more time correcting basic level changes than making creative decisions. A repeatable first pass is valuable if it leaves that editor with a small, identifiable set of moments to review. Choose sections with laughter, overlapping speakers, a whisper and a theme tune. A preset that treats every pause as empty space can change the meaning of a difficult answer even when the overall sound is polished.

A single lecturer recording from a fixed microphone has fewer alignment problems. For that person, consistency between lessons and the ability to produce different delivery formats from the same session may matter more. A concert producer seeking a deliberately wide dynamic range faces a different question. Listen to whether a crescendo or the decay after an important note still has its intended shape. A speech-friendly output can be technically tidy and musically inappropriate.

03 / WorkflowRebuild a remote interview from its two original tracks

In a proposed evaluation, ask a host and guest to keep the original microphone recordings from a short, permitted conversation. Retain a reference mix from the call if one exists. Choose a stretch containing interruptions and one long answer, then mark the same exchanges near the beginning and end. Those points let the editor distinguish an incorrect start from a recording that gradually drifts out of time.

Auphonic explains in its Sync Tracks engineering note that the model looks for shared audio, conversational timing and clues from the files. Where there is shared sound, it can estimate clock drift. Two isolated headphone recordings may instead supply a pattern of speech and silence, which gives a weaker cue about how the exchange fits. Inspect the suggested offsets and listen to ambiguous moments. A deliberate overlap between speakers should still sound like an interruption.

Start from unedited stems. The tool describes a single shift per track, so a cut inserted into the middle of only one recording can leave one side of the cut correctly aligned and the other side wrong. If a producer already edited out a pause on the host track, retrieve the original or account for that cut before trusting the alignment. Verify the last exchange as well as the first because a clean opening alone says little about drift.

After alignment, listen to the balance on ordinary earphones. Mark a quiet sentence that should carry weight and a loud laugh the audience should still experience as a laugh. Auphonic’s editable gain curve lets the editor use gain points to change particular passages after the adaptive leveler has done its work. Try a local correction and listen across the boundary into the following phrase. Avoid a sudden jump that calls attention to the adjustment.

Use the publisher’s loudness and true peak requirements for the deliverable. Save the final output and a reference identifying where the editor intervened. If one guest always needs a large correction, inspect that recording setup before adding another stage of automated treatment. A better microphone position for the next episode may save more editing time than further work on the finishing preset.

04 / PricingWork out the cost from the running time

OptionPriceDuration and conditions
Free$0Two hours per month, jingle on output and multitrack length restrictions.
Recurring S monthly$13 per monthNine hours each month, unused recurring credits reset.
Recurring S yearly$132 billed yearlyNine base hours each month, equivalent to $11 per month.
One-time package$12 onceFive processing hours that do not expire.

Selected US-dollar options before any applicable tax, consulted 19 November 2026. Auphonic credit pricing.

The credit rules count the duration submitted, with a three-minute minimum per production. An intro adds to the billed duration. Additional input tracks or multiple output formats in the same production do not individually multiply that time. For a proposed weekly interview lasting 45 minutes with a one-minute intro, four productions cover 184 minutes. An editor can compare that monthly total with a credit package and allow room for other shows and additional work.

The choice changes when production is irregular. A show that releases a six-episode season and then takes a break may use one-time hours more efficiently than a recurring allowance that resets. A network publishing throughout the year can consider the annual commitment and loyalty increases offered after continuous subscription milestones. Check feature eligibility too. The free tier limits multitrack productions to under 20 minutes and omits multilingual speech recognition and automatic shownotes.

05 / DistinctionsWhat to compare with a recording platform or an edit suite

Riverside’s production workflow is a relevant starting point when a remote recording has yet to happen. Decisions about how to capture and retrieve each guest’s track affect the material a finishing service can use later. If a program repeatedly receives only a flattened call recording, solving the capture problem can be more important than optimizing the same mix after every episode.

Descript’s editing workflow addresses another stage, where the editor is deciding which words and sections belong in the published piece. For an interview containing a sensitive correction, agree on the final spoken sequence before generating a delivery file. Processing may make a splice less audible, but it cannot establish whether joining two statements preserves the speaker’s intended meaning. Keep an approved edit and a separate finishing master.

06 / QuestionsWhat evidence should an editor keep?

Retain the unprocessed recording, the approved timeline and short excerpts of the opening and closing sync references. Those small pieces make it possible to explain a late-discovered echo or to reproduce a change when the show moves between editors. If the original is clipped or two people share a single microphone, document that constraint. A polished preview does not demonstrate how much recoverable information was present in the source.

Consider the destination as well as the processing step. A podcast feed, an internal lesson and a broadcast slot may require different files or metadata. Listen to the encoded output the audience will actually receive, including the transition from intro to speech, and check the delivered specification against the brief. An acceptable studio master can still leave the listener with an unexpectedly loud theme tune on a small speaker.

07 / DecisionChoose the job you want Auphonic to take over

01

An independently produced interview

Use a short, representative pair of microphone files to evaluate alignment and compare local gain changes. Preserve the original tracks and listen to both ends.

Move to a paid configuration for longer multitrack episodes.
02

A network publishing every week

Count real submitted minutes for a month, compare recurring and one-time rates, and agree which exceptions each editor will audition.

Choose a repeatable preset and a named listening pass.
03

A documentary built around location sound

Select a scene where ambience matters to the narrative. Evaluate speech clarity alongside the sound that tells the listener where the scene happened.

Keep only the processing that serves the scene.
What should we explore next?

A business worth understanding.

Suggest your business or one you find interesting. Tell us what you want to understand about its product, positioning, design or workflows.

Suggestions are free. Selection and publication stay with the desk.

Continue reading

All in this category