Automating Multitrack Post-Production with Auphonic: A Complete Broadcast Guide

Podcasters and broadcast audio engineers face a consistent dilemma: multi-mic studio recordings capture exceptional vocal nuance, but they introduce compounding technical headaches. When three or four microphones share the same physical room, acoustic bleed, cumulative background room hiss, volume disparities between dynamic speakers, and differing vocal timbres demand hours of manual editing. Post-production often turns into an exhaustive cycle of spectral gating, fader automation, EQ balancing, and loudness measurement against stringent delivery specifications.

Auphonic tackles this operational bottleneck through its Multitrack Post-Production engine. Unlike traditional single-track mastering tools that apply broad dynamic range compression to an already collapsed mixdown, Auphonic ingests isolated stems—or splits multichannel field recordings—and applies contextual, inter-track intelligence. It identifies who is speaking at every millisecond, strips correlated crosstalk across open microphones, suppresses inactive mic room tone, levels disparate dialogue tracks, and renders a broadcast-compliant master configured precisely to international loudness targets.

This guide provides a step-by-step technical blueprint for configuring Auphonic Multitrack productions, creating reusable presets, dialing in noise reduction and leveling algorithms, and automating professional broadcast workflows.

1. How Auphonic Multitrack Processing Works

Singletrack processors operate blind: they must guess whether a quiet section is an intended conversational pause, room ambience, or mic spill from a secondary speaker. In contrast, Auphonic Multitrack evaluates all incoming stems simultaneously, creating a composite analysis before committing to any signal transformation:

• Microphone Bleed Removal (Crossgate): In co-located panel recordings, Speaker A’s voice inevitably spills into Speaker B’s microphone with a slight acoustic delay. Gating alone rarely eliminates this bleed without unnatural cutting or phasing. Auphonic compares the direct waveform on Speaker A's track with the correlated leakage on Speaker B's track, phase-cancelling or attenuating the spill so each speaker sounds completely isolated.
• Adaptive Noise Gating: Mixing four open microphones multiplies the ambient noise floor fourfold. Auphonic dynamically expands and attenuates idle tracks while preserving the active speaker's channel, dropping ambient studio rumble and room tone during conversational pauses without cutting off word endings.
• Multitrack Adaptive Leveler: Speech dynamics are equalized across two dimensions: mid-term level corrections balance uneven speakers (e.g., quiet guests versus loud hosts), while micro-dynamics compression tames transient spikes and plosives. Crucially, vocal leveling occurs independently of music beds or ambient soundscapes.
• Stem-Specific Denoising & Filtering: A host on an Shure SM7B in a treated room requires different spectral treatment than a remote guest calling over VoIP with low-bitrate compression. Auphonic allows engineers to assign distinct denoising profiles and voice enhancement models to each individual track.
• Global Loudness Normalization & True Peak Limiting: Once individual stems are cleaned, leveled, and mixed according to user-defined panorama and ducking rules, the combined master passes through a multi-pass ITU-R BS.1770 / EBU R128 loudness normalizer and a 4x oversampled True Peak limiter.

2. Audio Preparation and Ingestion Rules

To produce clean, artifact-free mixdowns, input tracks must satisfy strict structural criteria:

The Golden Rules of Multitrack Stems

1. Strict Zero-Time Alignment: Every audio track must start at the exact same timestamp (00:00:00.000). If a guest only speaks thirty minutes into a recording, do not trim the preceding silence in your digital audio workstation (DAW) without rendering leading silence. Auphonic aligns tracks in parallel from frame zero; shifted stems will desynchronize dialogue and break bleed detection.
2. Dedicated Stems for Speech and Music: Never combine intro/outro music and dialogue on the same track prior to uploading. Music must sit on its own stem so the classifier does not mistake melodic instruments for speech or apply aggressive vocal leveling.
3. One Speaker Per Track: Assign each participant their own physical input or track file. Grouping multiple speakers onto a single stem disables inter-track crosstalk cancellation.
4. Multichannel Recording Support: If you record on hardware field recorders (such as Zoom F-series, Sound Devices MixPre, or RØDECaster Pro) that output polyphonic WAV files with multiple embedded channels, Auphonic can split these directly in the browser via Import Multiple Tracks > Multichannel File. This extracts up to 8 distinct stems (handling iXML metadata, mono tracks, and paired stereo tracks) without forcing you to split files manually in a DAW.

3. Configuring Track-Level Algorithms

When opening a new Multitrack Production in the Auphonic web console (https://auphonic.com/engine/multitrack/), configure each stem according to its acoustic characteristics:

[Host 1 (Local Mic)]    --> [AutoEQ + Static Denoiser]      ----+
[Guest 2 (Local Mic)]   --> [AutoEQ + Static Denoiser]      ----+--> [Bleed Removal + Crossgate]
[Guest 3 (Remote Call)] --> [Studio Voice / BWE + Dynamic]  ----+--> [Adaptive Leveler / Ducking]
[Music Bed / Sound FX]  --> [No Denoise + Ducking (-18dB)]  ----+--> [True Peak Limiting / EBU R128]
                                                                        |
                                                                        v
                                                               [Broadcast Master]

Track Identification & Stereo Panorama

• Assign meaningful Track Identifiers (e.g., "Sarah - Host", "Marcus - Guest", "Music Bed"). Auphonic's Whisper Automatic Speech Recognition (ASR) uses these identifiers directly as speaker labels when generating transcripts and shownotes.
• Set the Stereo Panorama (Balance) for each vocal track. For natural conversational separation, pan co-hosts slightly left or right (e.g., L15 / R15 or L25 / R25). Music beds should remain centered stereo. If all inputs are mono and left centered, the final master output will render as mono.

Adaptive Filtering and Voice Enhancement

For each vocal stem, select the appropriate filtering method:

• Adaptive High-Pass Filtering: Applies zero-phase, linear-phase high-pass filters tailored dynamically to the fundamental pitch of the speaker's voice (male or female). It eliminates mechanical desk bumps and air conditioning rumble without introducing phase smearing or asymmetric waveform skew.
• Voice AutoEQ: Analyzes the speaker’s frequency curve and optimizes spectral balance dynamically. It controls harsh sibilance (de-essing), removes boxiness, and suppresses plosive pop sounds (de-plosive).
• Voice AutoEQ + Bandwidth Extension (BWE): Ideal for remote guests recording through telephone lines or web conferencing tools where codecs cut off frequencies above 4 kHz or 8 kHz. BWE synthesizes high-frequency harmonics to restore brightness and intelligibility.
• Studio Voice (Beta): An advanced neural voice reconstruction model. Rather than simple filtering, it reconstructs vocal timbre damaged by aggressive bit-rate compression, severe distortion, clipping, or digital processing artifacts.

Targeted Noise & Reverb Reduction

Auphonic provides four distinct denoising algorithms:

For treated studio recordings with multiple open mics, set vocal tracks to Static Denoiser (or Dynamic Denoiser with moderate settings), leaving Remove Reverb at default or lowering it slightly if natural room acoustic warmth is desired. For remote guests calling from untreated bedrooms, enable Dynamic Denoiser with Remove Breaths enabled to strip gasping inhalation sounds.

Fore/Background Ducking and Loop Settings

On music beds or ambient audio tracks:

• Set Fore/Background to Background or Duck this track.
• Set Level of Background Segments (default is -18 dB). Whenever speech is active on any dialogue track, Auphonic automatically ducks the music by 18 dB, fading smoothly based on the Ducking Fade Time (default: 500 ms).
• If your background music bed is shorter than your 45-minute discussion, enable Loop / Trim. Auphonic automatically loops the track seamlessly until the final speech segment finishes.

4. Configuring Master Mixdown Algorithms

Once track-level stems are conditioned, global algorithms process the summed mixdown:

1. Crossgate (Remove Mic Bleed)

Ensure the Remove Mic Bleed checkbox is checked on the Master Track. Auphonic correlates acoustic arrivals across tracks. When Host A speaks, their voice is subtracted from Host B's open microphone, eliminating phasing artifacts, comb filtering, and muddy hollow reverberation.

2. Adaptive Noise Gate

Enable the Adaptive Noise Gate on the Master Track. This instructs Auphonic’s classifier to dynamically expand the noise floor on silent channels, attenuating ambient bleed and room tone when speakers pause.

3. Master Adaptive Leveler & Broadcast Mode

Under Master Track settings, switch the Leveler Mode from standard leveling to Broadcast Mode to enforce broadcast compliance:

• Maximum Loudness Range (MaxLRA): Restricts the overall dynamic range of the master. For news broadcasts and technical interviews, set MaxLRA between 3 LU and 5 LU. For dramatic narrative audio or talk shows, set MaxLRA between 5 LU and 8 LU.
• Maximum Short-term Loudness (MaxS): Constrains transient volume bursts (measured over a 3-second window) relative to the integrated loudness target. Setting MaxS to +5 LU relative ensures compliance with strict European television advertising directives (such as EBU R128 S1).
• Maximum Momentary Loudness (MaxM): Regulates instantaneous peaks (measured over a 400 ms window) to prevent abrupt dynamic spikes.

4. Global Loudness Normalization & True Peak Limiter

Select your target delivery standard:

Enable the Dual Mono flag when rendering mono podcast feeds. Auphonic automatically applies a -3 LU offset (-19 LUFS) so that mono files played back on stereo consumer devices sound perceptually equivalent to -16 LUFS stereo productions.

5. Automatic Cutting & Editor Integration

Auphonic’s AI can eliminate structural speech imperfections directly on the timeline:

• Cut Silence: Trims dead air caused by conversational hesitations or mic checks without affecting deliberate natural pauses between spoken sentences.
• Cut Fillers: Detects and cuts speech fillers ("um", "uh", "ah", "euh", "ähm") across multiple languages. In multitrack setups, if Speaker A utters "um" while Speaker B is mid-sentence, Auphonic fades the filler to silence on Speaker A’s track instead of slicing the timeline, preserving Speaker B’s dialogue.
• Cut Coughs: Removes sneezes, throat-clearing, and harsh respiratory sounds.
• Non-Destructive DAW Workflow: If you prefer manual review over destructive rendering, set Cut Mode to Export Uncut Audio or Set Cuts To Silent, and add a Cut List file to your output formats. Auphonic exports EDL and CSV cut markers for Reaper, Adobe Audition, DaVinci Resolve, Final Cut Pro, and Audacity, allowing engineers to inspect and refine edit boundaries on their own DAW timelines.

6. Delivery Formats, Presets, and Automated Publishing

Once algorithms are chosen, streamline distribution:

1. Multiple Output Encodings: Add parallel export targets:
   • Master MP3 (192 kbps or 256 kbps CBR/VBR) with embedded ID3v2 tags and chapter markers.
   • Uncompressed 24-bit 48 kHz WAV master for archival storage.
   • Individual Tracks (ZIP archive): Select Individual Tracks under Output Files to download a ZIP file containing every processed, leveled stem in clean WAV format. This allows sound designers to perform further surgical automation in a DAW while benefiting from Auphonic’s bleeding removal and denoising.
2. Metadata & Chapters: Auphonic automatically embeds chapter markers (including titles, URLs, and chapter art) into MP3, AAC, and Vorbis Comments. It can also generate dynamic audiogram videos with embedded cover art for YouTube, complete with burned-in subtitles.
3. Automated Preset Deployment: Save your complete multitrack configuration as a Preset (https://auphonic.com/engine/preset/). Connect external cloud services (such as Amazon S3, Google Drive, SFTP, Dropbox, Libsyn, or Podbean). You can configure a Watch Folder so that dropping raw multichannel stems into an S3 bucket or Google Drive folder triggers automated processing and immediate distribution to podcast hosting platforms without opening a web browser.

7. Operational Trade-Offs and Best Practices

While Auphonic dramatically shortens post-production turnaround, engineers should evaluate key operational trade-offs:

• Acoustic Ambience versus Sterile Dialogue: Combining aggressive dynamic denoising, speech isolation, and crossgating can strip desirable acoustic air, leaving voices sounding clinical or "hollow." If an ensemble discussion begins to sound unnatural, reduce Remove Noise and Remove Reverb by 10–20 dB on dialogue stems, or deactivate the Adaptive Noise Gate to retain subtle room ambience.
• Credit Economics: Auphonic bills Multitrack productions based strictly on the duration of the finished output audio file, not the number of input tracks. A 60-minute four-track episode consumes 1 hour of processing credit (the free tier provides 2 hours monthly; multitrack productions under 20 minutes can be processed for free). Unused recurring monthly credits reset, whereas one-time credits never expire.
• When to Keep Manual Control: Complex dramatic audio fiction, scripted film dialogue with intricate Foley beds, or multi-instrumental musical performances require dedicated DAW automation. Auphonic excels at dialogue-dominant content: interview podcasts, panel broadcasts, educational webinars, and radio roundtables.

By pairing stem-level cleanup with cross-track bleed suppression and broadcast-standard loudness limiting, Auphonic turns hours of tedious multi-mic post-production into a reliable, consistent, and automated broadcast pipeline.

References

• Auphonic Documentation — Auphonic Help 2026 documentation
• Audio Algorithms — Auphonic Help 2026 documentation
• Multitrack Post Production Algorithms — Auphonic Help 2026 documentation
• Multitrack Best Practice — Auphonic Help 2026 documentation
• Multitrack — Auphonic Help 2026 documentation
• Pricing FAQ — Auphonic Help 2026 documentation