Steinberg SpectraLayers Pro 13 [WiN]

• Product: SpectraLayers Pro 13
• Developer: Steinberg
• Version: 13.0.0
• Requirements: Windows 10 or later
• Source: steinberg.net/spectralayers

SpectraLayers Pro 13 is a spectral audio editor and AI-assisted source-separation application operating on a spectrogram — a visual frequency-versus-time display where every region of a sound’s spectrum is individually selectable and editable — rather than on a conventional audio waveform. It functions as a standalone application and, through ARA2 integration, as a bridge from inside Cubase, Nuendo, Studio One, Logic, and other ARA2-compatible hosts. Version 13 adds nine new modules, including Reconstruct for spectral resynthesis, Unmix Sound Effects for isolating discrete sounds from continuous background, and three voice-processing tools, alongside full keyboard customization and saveable interface layouts. Its differentiator is editing resolution: problems that exist only within a specific frequency range at a specific point in time — a whistle embedded in dialogue, a resonance inside a musical note, a secondary voice overlapping a primary one — can be addressed precisely at that frequency and time without affecting the surrounding audio. For engineers asking whether SpectraLayers Pro 13 handles their specific restoration or separation problem better than iZotope RX, this is the technical basis for that comparison.

Key Takeaway

Activates when a recording problem is frequency-and-time-specific rather than broadband — a lip smack that occupies a specific frequency region at a specific moment, an S-tone that pokes above surrounding content at a consistent spectral position, or a second voice that needs separation from a primary one without eliminating it entirely. Displaces conventional noise reduction and spectral repair plugins when the problem can be seen and addressed directly on the spectrogram rather than treated with a broadband process applied across the full signal. Isn’t an automatic repair utility: independent review coverage is explicit on this — SpectraLayers requires an engineer to work inside its interface, understand its layering system, and make deliberate editing decisions, rather than running an automatic pipeline that produces a cleaned result with no manual intervention.

Reconstruct Module: Spectral Resynthesis as a Surgical Repair Tool

Reconstruct is the standout new module in version 13, replacing and significantly expanding on the former Heal function: it uses synthesis to rebuild spectral data within a selection from the surrounding material, analyzing the before/after region to reconstruct tonal content and the above/below region to reconstruct transients and noise. The three reconstruction targets — tones, transients, noise — have individually configurable amounts, so a repair focused purely on tonal content (a sustained tone disrupted by a click) can weight toward tone reconstruction without pulling in transient or noise character from the surrounding material.

Reconstruct’s performance is tied to the interface’s FFT Size setting, which functions like pulling focus on a camera: smaller values sharpen time-domain events like transients and short sounds, while larger values increase frequency precision for sustained tones and complex textures. An independent post-production specialist who worked through the alpha and beta stages describes Reconstruct as the most likely “Swiss Army knife” tool in the new module set, using it to remove whistles from crowd recordings, stray barks, thumps, and stray dialogue pieces — content that was previously patched over with room tone rather than spectral-repaired in place. The Attenuate module, added alongside Reconstruct, handles content that needs to be turned down rather than replaced — a sharp sibilant that stands out from its surroundings gets brought to the level of surrounding content rather than removed entirely.

Unmix Sound Effects and the Post-Production Separation Workflow

Unmix Sound Effects separates short, discrete sound effects from continuous environmental or ambient background — traffic, crowd, wind, nature — placing each on its own layer for independent editing, export, or further processing. The module analyzes a user-specified duration of material to distinguish continuous ambience from discrete events, with Threshold, Tolerance, and horizontal/vertical Smoothing controls governing how cleanly those two categories separate. Lower Smoothing values preserve attack precision and frequency detail in the separated layers; higher values smooth unmixing results at the cost of potentially softening transients or losing frequency edges that distinguish the sound effect from the background.

Ambience Heal directly addresses the gap left behind after any Unmix operation — dialogue unmixing, sound effect separation, or any other operation that creates a hole in the spectral content of one layer. Rather than leaving silence or an obvious transition where content was removed, Ambience Heal fills the gap using room tone or environmental sound already present elsewhere in the recording, using the source layer’s remaining spectral content to generate material that bridges the gap transparently. A three-voice dialogue cleanup that previously required sourcing replacement room tone from elsewhere in the recording can now use Ambience Heal to fill the gaps after each Unmix operation without leaving audible transitions.

Three New Voice Tools and What Each One Actually Does

Unmix Two Voices separates two overlapping speakers into independent layers without requiring voice profile registration, using AI analysis to distinguish the two voices from their spectral interaction. This differs structurally from Voice DeCrosstalk, which removes a secondary voice from a primary recording rather than placing both in their own separate, independently editable layers — DeCrosstalk is a cleanup tool for minimizing bleed, while Unmix Two Voices is a separation tool for extracting both parties as usable audio. Voice DeClick handles the third category: mouth noise removal, specifically lip smacks and tongue clicks that occur within a voice performance rather than as environmental interference. All three perform differently depending on FFT Size: Voice DeClick benefits from smaller FFT values that sharpen time-domain precision around the click’s position, while Unmix operations aimed at complex spectral separation between two speakers benefit from larger values that improve frequency-domain resolution.

These three modules address a cluster of post-production problems that previously required either a separate restoration application, manual spectral painting, or accepting a less precise result from SpectraLayers 12’s Voice Denoise module. Voice DeCrosstalk existed as a function inside that module in version 12 and has been promoted to its own dedicated module in 13, reflecting how frequently it’s used in dialogue editing as a standalone operation rather than as part of a broader denoising pass.

Workflow Customization and the Pro Tools Integration Path

Every module and the majority of functions in SpectraLayers 13 are now assignable to user-defined keyboard shortcuts, with Interface Layouts — saveable screen configurations of open modules and panel arrangements — recallable via the same shortcut system. This means a Pro Tools editor can map SpectraLayers functions to the same keystrokes used for equivalent Pro Tools actions — fade keys, zoom keys, panel toggle keys — so muscle memory built in Pro Tools transfers to SpectraLayers without a separate keybinding layer to learn. Interface Layouts function analogously to Pro Tools window configurations: set up the screen with the modules relevant to a specific workflow (dialogue cleanup, music stem separation, loudness checking), assign a keystroke, and recall that exact layout mid-session.

The Pro Tools AudioSuite Bridge plugin gains multi-clip round-trip editing in version 13: multiple clips can now be sent from Pro Tools to SpectraLayers simultaneously, each in its own tab, processed individually or via batch processing across all open tabs, then returned to Pro Tools in a single pass. This replaces the previous clip-by-clip workflow that required separate round-trips per clip, which accumulated into meaningful additional time on sessions with many discrete elements requiring spectral treatment. Batch processing has also been extended to run across multiple open project tabs simultaneously rather than one project at a time — a direct response to how SpectraLayers is used in post-production workflows where multiple elements from the same recording session need consistent spectral treatment.

Nine New Modules, Still Requiring an Engineer Behind Each One

Reconstruct, Ambience Heal, Unmix Sound Effects, and the three voice tools expand what can be addressed at a specific frequency and time — but every one of them depends on an engineer understanding the spectrogram, calibrating FFT Size for the task, and evaluating the result before committing it.

FAQs