Steinberg SpectraLayers Pro 11

SpectraLayers continues to develop at a spanking pace. We check out what’s new.

Given how fully AI is now integrated into the processing algorithms of SpectraLayers, it’s perhaps not wholly surprising that Steinberg have adopted a fairly rapid update cycle for their flagship spectral editing software. There is no sign of this slowing down; just 12 months after reviewing v10, we now have SpectraLayers Pro 11 (SLP11). So, what’s new in this release?

Workflow

Well, quite a lot and — while many of the highlights are AI related — they are not exclusively so. These include UI refinements such as the Display Panel now allowing the control selection to be fully customised, a new Home screen on launch for quicker access to new or existing projects, and the Colour Dimming option available within the Selected Layers view of the Layers panel. The latter is very useful, allowing you to see selected layers in their full colour, while showing non‑selected layers ‘dimmed’ by a user‑defined amount in the background.

A new Modules panel provides easy access to all of the processing modules, with full customisation of the selection shown. For those with limited screen real estate, SLP11 also offers a ‘compact’ viewing option (accessed via the small double arrow icon at the very top of the Panels area of the UI) where panel options are compressed and represented by a thin vertical strip of icons.

However, from a workflow perspective, a particular highlight is the combination of the new Modules Chain option and the ability to batch process multiple files using these chains. Incidentally, this includes batch processing files through an unmixing process, and the generated layers are handled elegantly alongside their original source file. This whole workflow is impressive, allowing you to set SLP11 off on what might be quite a lengthy series of routine processing jobs while you multitask (for example, by consuming a hot beverage).

New Editing & Processing Options

SLP11 also brings new and improved editing and processing options. For example, there are new options to create fades around regions as you make your selection and a new loop preview option for auditioning as you edit. SLP11 also adds volume automation at either the file or layer level, accessed via new envelope buttons that sit alongside the existing mute, solo and polarity buttons within the Layers panel.

For those working on dialogue, the Voice DeNoise module has been improved with the addition of a ‘strong’ noise reduction ratio option and, if clipping is a problem, the brand new DeClip module can do a remarkable job of rescuing dialogue that might otherwise be unusable. Sound designers will enjoy the new Reverse module that, when applied to layers unmixed from an individual sound, can generate all sorts of weird (and sometimes wonderful) sonic outcomes. There is also a new Transient Pencil tool, allowing you to draw transients into the spectrogram either for creative sound design or for repair.

From an editing perspective, perhaps most impressive is the new Transfer Brush. The concept is simple; it lets you draw over spectral data contained within one layer of a file and have it automatically transferred into a different (the target) layer of your choice. The process is totally non‑destructive and, as a means of manually cleaning up layers created by one of SLP11’s unmixing processes (yes, I’ll get to those in a minute), it’s absolutely brilliant. The style of the brush can be configured to suit the nature of the spectral data you want to select for transfer.

The new Transfer Brush makes it much easier to tidy up small details from an unmixing process.

Unmixing

The area of spectral editing where AI has made the most remarkable impact — and made the most jaws hit the floor — is in unmixing. Even with just 12 months between major releases, SLP11 doesn’t disappoint on this front and Steinberg have weaved the latest AI magic into pretty much every unmixing process under the SLP11 hood.

This includes the Unmix Song module which now offers a new Sax & Brass layer option. However, the most remarkable difference is in the quality of the separation. The drum stems seem to be crisper and cleaner, while the vocal stems are achieved with much finer detail and fewer of the ‘phasey’ artefacts unmixing algorithms can sometimes generate. Reverb and delay on the vocals also seem to be ‘caught’ within the vocal layer in a more coherent fashion. In the vast majority of examples I tried, v11 produced a clear and noticeable improvement over v10 (and that was already very good). This improvement was particularly obvious in somewhat busier mixes.

iZotope’s RX11 (which I reviewed in the July 2024 issue) also offers impressive AI‑powered song unmixing but, like all vocal extraction processes (including SLP10), it works well where just a lead vocal is present within a mix but can be challenged when backing/harmony vocal lines are also present. In that review, I commented that whoever is first to crack the nut of separating lead vocals from backing vocals will take a leap ahead of the competition. Well, Steinberg seem to be first off the blocks on that front with SLP11’s new Unmix Chorus module. If required, this can be applied to your vocal layer from the Unmix Song process and will then attempt to separate out lead and backing vocals into two further layers. Yes, it depends upon just how successful that original vocal layer extraction was and, no, the results are not (yet) perfect, but it’s still impressive. It’s also a prime candidate for further clean‑up using the Transfer Brush mentioned earlier. I’m sure other developers will soon bring their own take on this process but hats off to Steinberg for taking the leap.

Two further new unmixing modules have also been added. Unmix Crowd Noise does pretty much what it says on the tin and would be useful for processing stereo recordings of live music performances if you need to adjust the balance between the music and the audience participation. The new Unmix Mid/Side module lets to create separate layers for the Mid and Sides content of a stereo or multi‑channel audio file and it offers an ‘expanded’ mode where the Sides content can be placed into a stereo layer. If you want to apply SLP11’s spectral editing magic to adjust or enhance the stereo image of your audio, these new layer options give you a means to do so.

There have also been improvements to other existing unmixing modules. For example, the Unmix Drum module, which in SLP10 allows you to take a drum layer from the Unmix Song process (or a drum loop) and create separate kick, snare and cymbal layers, now adds a separate hi‑hat layer to that process. Whether it’s to rebalance a drum recording, or to provide isolated triggers for a drum replacement process, the separations are now cleaner and more usable. And, alongside the other dialogue‑related improvements noted earlier, the Unmix Multiple Voices module has been improved, allowing you to register individual voices (by selecting sections where a specific voice is heard in isolation) prior to running the unmixing process and thereby improve the way the AI can extract each voice. Whether for rebalancing on‑set dialogue or isolating a single voice for dialogue replacement, film, TV and podcast audio editors will feel some love.

Finally, it’s worth noting that SLP11 improves the options for moving multiple layers from SLP11 and into your ARA‑friendly DAW (such a Cubase or Nuendo) or on to your desktop file system. The latter is particularly useful; you can just drag and drop from the Layers panel and SLP11 handles all the file creation duties required.

If SpectraLayers is part of your regular workflow, then the unmixing advances and the combination of Modules Chain and batch processing provide compelling reasons to update.

Take My Money?

These are, of course, just the obvious highlights of the new release but there are other additions, improvements and fine‑tuning that will undoubtedly appeal in specific use‑case scenarios. SpectraLayers Pro 11 has come a remarkable distance since Steinberg acquired it from Magix in 2019 and I think the improvements and additions made for this release — AI and otherwise — do warrant a whole point update. Yes, more casual users could be forgiven for rolling their eyes at the prospect of another paid‑for upgrade only 12 months after the last one but, if SpectraLayers is part of your regular workflow, then the unmixing advances and the combination of Modules Chain and batch processing provide compelling reasons to update. This is, however, something of an AI arms race. It’s exciting to see it develop but it is not going to stop any time soon. I therefore fully expect to see you here again in a further 12 months. In the meantime, SpectraLayers Pro 11 is undeniably impressive.