Steinberg releases SpectraLayers 12 to boost unmixing and speech processing capabilities

Steinberg has announced the availability of SpectraLayers 12, which includes a range of new and enhanced tools that take the unmixing and speech processing capabilities of SpectraLayers to a new level.

SpectraLayers 12 includes two brand-new unmixing modules, which bring more precision and flexibility to the unmixing process. 

The new Unmix Soundtrack module separates audio into dialogue, effects, and music layers — especially useful for post-production workflows involving mixed content such as film soundtracks, trailers, and radio broadcasts. By isolating the core audio components, it enables more precise dialogue cleanup, music replacement, and effects enhancement.

The new Unmix Instrument module isolates and extracts user-identified instruments by analysing the target instrument before unmixing. With the flexibility to extract custom elements from mixes, it’s now easier to rework arrangements, target processing, and create stems for remixing and restoration. 

Enhanced unmixing features, alongside the new unmixing modules, the Unmix Song, Unmix Drums, and Unmix Noisy Speech processes have been significantly improved to deliver even better results.

Unmix Song now delivers dramatically better vocal separation, major enhancements to drum, bass, and sax/brass unmixing and subtle improvements to guitar and piano unmixing. 

Updated algorithms of the Unmix Drums process mean it can extract six discrete kit elements — kick, snare, toms, hi-hats, ride, and crash cymbals, while the Unmix Noisy Speech module has been significantly improved to deliver much clearer speech/noise separation, along with a substantial reduction in artefacts.

SpectraLayers 12 takes voice processing to new levels of accuracy, with new and enhanced modules focused on vocal analysis and improvement. The new Voice Enhance module uses generative AI to intelligently process and then reconstruct poor-quality voice recordings, noticeably improving quality while preserving the speaker’s unique identity. 

The new DePlosive module automatically eliminates distracting plosives in speech recordings, with adjustable controls for threshold, frequency, and reduction ratio for fine-tuning. Improvements to the Voice DeNoise and Unmix Transcription modules, respectively, deliver much cleaner results when the background to speech is noise or music, alongside transcription quality that supports twice as many languages and is twice as accurate as before.

There are many workflow improvements which make the powerful processing of SpectraLayers 12 faster and more intuitive to use. A headline feature is the ability for all SpectraLayers tools and processes to operate on multiple layers simultaneously. 

In standalone mode, this operates vertically across layers, allowing events spanning multiple stems to be edited at the same time. The feature extends horizontally in Cubase or Nuendo in ARA mode — for example, when using the speech processing tools in post, on multiple clips located anywhere in a project.

The selection of frequencies, harmonics, and transients have all been significantly upgraded in SpectraLayers 12. Narrow frequencies can be tracked more precisely, with elastic frequency selection allowing shifting content to be more accurately processed. 

Harmonics selection is more intuitive — no more parameter tweaking necessary — and transient selection is more detailed and flexible. These improvements all result in cleaner, more focused selections, reducing the time spent on manual fine-tuning.

“Since we first introduced unmixing algorithms in SpectraLayers, there hasn’t been any other application that could match its quality. Now, with version 12, the clarity of the separated stems is exceptional… You could almost believe that they were never part of a mix,” said marketing manager Pro Audio, Luis Dongo. 

“This version also has a strong focus on the needs of the post-production industry, with innovative features like soundtrack unmixing and speech enhancement. Existing voice processing algorithms are constantly being refined and updated, while workflow and productivity improvements requested by our loyal users haven’t been forgotten.”

Buy SpectraLayers 12 here.