More AI features in Adobe Audition for sound editing

1. AI-Powered Features

Artificial intelligence represents the biggest near-term opportunity for Audition to differentiate itself. The following features would automate time-consuming manual tasks and deliver meaningful results for podcast producers, voiceover artists, broadcast engineers, and music editors alike.

1.1  Semantic noise removal

Instead of manually dialing in noise reduction thresholds, AI should identify and classify noise types — HVAC, traffic, room tone, electrical hum — and apply context-aware profiles automatically. The interface should expose a single adjustable 'aggressiveness' slider rather than a bank of technical parameters. This makes the tool accessible to non-engineers without sacrificing precision for advanced users.

  ● High priority  

1.2  AI-powered dialogue cleanup

Automatically detect and remove mouth clicks, breath sounds, mic pops, and plosives without affecting the surrounding audio. The current workflow requires manual identification and repair, which is extremely time-consuming. The ideal implementation is one-click processing with per-detection override — allowing editors to review and restore anything the AI incorrectly flagged.

  ● High priority  

1.3  Smart silence detection and pacing

Existing silence detection is based purely on audio level. An AI-based replacement should understand natural speech rhythm — detecting pauses that feel long relative to the speaker's cadence — and allow trimming or shortening them without making the result sound unnatural. Ideally, the system should learn from an editor's decisions over time to improve per-project.

  ● High priority  

1.4  Automatic room acoustics matching

When audio from different recording environments is combined — remote interviews, multi-location podcasts, overdubbed narration — the acoustic mismatch is immediately noticeable. AI should analyze the room tone and impulse response of each track and normalize them to a shared acoustic profile. This alone would save hours on professional podcast production.

  ● High priority  

1.5  Speaker diarization and track separation

Automatically detect and separate multiple speakers in a single recording, assign each speaker to an individual track, and label them by identity. This is already available in tools like Descript and is increasingly expected in professional audio software. It is particularly valuable in podcast, interview, and conference recording workflows.

  ● High priority  

1.6  Stem separation

Integrate machine-learning-based stem separation — vocals, instruments, drums, bass — directly into Audition. This would allow editors to isolate or remove individual components of mixed audio without requiring third-party plugins like iZotope RX or Spleeter. Use cases include music podcast production, archival restoration, and sound design.

  ● Medium priority  

1.7  AI mastering assistant

Analyze a finished mix and suggest or automatically apply mastering corrections: dynamic range adjustments, loudness normalization to LUFS targets, and EQ balance corrections. The system should offer platform presets (podcast, broadcast, Spotify, Apple Music, Netflix) and present its suggestions transparently before committing changes.

  ● Medium priority  

1.8  Voice cloning for gap-fill

Using an existing recording of a speaker as a reference, generate short corrective phrases or fill missing words. This is invaluable for podcast editing and voiceover work where re-recording a session is impractical. The feature must include ethical safeguards: audible watermarking options, disclosure metadata, and explicit user consent workflows.

  ● Medium priority  

1.9  Natural language editing commands

Allow editors to type or speak instructions — 'remove all background noise from 00:30 to 01:15' or 'bring up the second speaker 2 dB in the chorus' — and have AI execute them on the timeline. This lowers the barrier for less experienced users and accelerates repetitive operations for professionals.

  ● Medium priority  

2. Core Editing Capabilities

Several fundamental editing capabilities are either missing or underdeveloped in Audition compared to competing tools. Addressing these would significantly improve the experience for experienced editors working on complex projects.

2.1  Fully non-destructive effects chain

All effects should be applied as a live, non-destructive chain that can be reordered, bypassed, and removed at any point in the project lifecycle — similar to a plugin rack in a DAW. Currently, many operations in Audition's waveform editor are baked into the audio file. This should be resolved by moving to a fully non-destructive architecture throughout the application.

  ● High priority  

2.2  Waveform-linked transcript editing

An automatically generated, editable transcript that is directly linked to the waveform. Deleting or rearranging words in the transcript removes or moves the corresponding audio. This feature — available in tools like Descript and Hindenburg — has become a standard expectation for podcast and spoken-word production workflows.

  ● High priority  

2.3  Spectral repair with content-aware fill

The spectral editor should be expanded with multi-selection, lasso tools, and frequency-locked brushes. More significantly, it should include AI-powered content-aware fill — analogous to Photoshop's — that seamlessly removes tonal artifacts such as sirens, ringing phones, or microphone interference by reconstructing the surrounding spectral content.

  ● High priority  

2.4  Clip gain automation

Clip-level gain should be independently automatable with its own lane on the timeline, separate from track volume automation. This is a standard feature in Logic Pro and Pro Tools that is conspicuously absent in Audition, and its absence forces inefficient workarounds for fine-grained level control.

  ● Medium priority  

2.5  Improved punch-in and comping

Punch-in recording should include a configurable pre-roll and post-roll buffer by default. Audition should also offer automatic take comparison — using pitch and timing analysis to identify the best take — and allow layered comping directly in the waveform view, rather than requiring editors to work around the current limited implementation.

  ● Medium priority  

2.6  Improved loudness metering

LUFS, LRA, and true peak meters should be permanently docked, always-visible panels — not buried in the Diagnostics or Amplitude Statistics panels. They should include selectable platform presets for Spotify, Apple Podcasts, Netflix, EBU R128, and ATSC A/85, and update in real time during playback.

  ● High priority  

3. UX & Workflow Improvements

Several workflow and usability issues slow down everyday production work. The following improvements would reduce friction for recurring tasks and bring Audition closer to the efficiency expected of a professional-grade tool.

3.1  Customizable macro system

Users should be able to record, name, and assign keyboard shortcuts to sequences of actions. This is critical for repetitive podcast production tasks — for example: apply loudness normalization, add fade in/out, export to MP3 at 192 kbps. A macro system would make Audition significantly more competitive with tools like Adobe Premiere Pro's own action recording features.

  ● High priority  

3.2  Batch processing overhaul

The batch processor requires a full redesign. It should support conditional logic, allow drag-and-drop reordering of processing steps, show a preview before committing changes, and present a modern navigable interface. The current dialog is functionally limited and visually outdated compared to even basic audio converters.

  ● Medium priority  

3.3  Project templates with full configuration

Users should be able to save complete project configurations — track layouts, routing, effects chains, loudness targets, export presets — as named templates. Starting a new podcast episode, voiceover session, or radio spot should take seconds, not minutes of setup.

  ● Medium priority  

3.4  Plugin management improvements

A centralized plugin manager with category tagging, favorites, A/B comparison while auditioning, and the ability to save and name effect chains as reusable presets available across all projects. Currently, managing third-party VST plugins in Audition is cumbersome and lacks the organizational tools that professional users require.

  ● Medium priority  

3.5  Cloud-based collaboration and version history

Cloud-synced projects with version history, the ability to add time-coded comments on clips or regions, and a shareable review link — similar to Frame.io's model, but built specifically for audio. This is particularly valuable for remote podcast production teams, where client approval workflows are currently handled outside the application entirely.

  ● Nice to have  

Summary

The table below provides a quick-reference overview of all proposed improvements by section and priority.

Feature             Category         Priority

Semantic noise removal   AI features     High

AI dialogue cleanup              AI features     High

Smart silence detection    AI features     High

Room acoustics matching                               AI features     High

Speaker diarization                AI features     High

Stem separation                       AI features     Medium

AI mastering assistant        AI features     Medium

Voice cloning / gap-fill         AI features     Medium

Natural language editing   AI features     Medium

Non-destructive effects chain                      Core editing  High

Transcript-linked editing   Core editing  High

Spectral repair + AI fill         Core editing  High

Improved loudness metering                         Core editing  High

Clip gain automation            Core editing  Medium

Punch-in and comping       Core editing  Medium

Macro system                             UX & workflow                            High

Batch processing overhaul                              UX & workflow                            Medium

Project templates                    UX & workflow                            Medium

Plugin management              UX & workflow                            Medium

Cloud collaboration              UX & workflow                            Nice to have

End of document