SpectraLayers Performance on Mac: Transparency and Long-Term Direction

[nabw]

I think one of the recurring issues around SpectraLayers is that discussions about rendering performance on Mac keep resurfacing without enough technical clarity from Steinberg.

If rendering is genuinely slower in certain environments, then users need clearer information about why it is slower and where the actual limitations are coming from.
Right now, people are forced to dig through forum posts and piece together assumptions themselves, which only leads to the same arguments repeating over and over again.

Of course, Steinberg has a broad user base, and I understand the need to support less technical users as well.
But at the same time, for professionals using these tools in real production environments, there also needs to be more transparency regarding technical constraints and architectural limitations.

I honestly think Steinberg is at a major crossroads right now.

For example, if GPU acceleration is becoming increasingly important, then Windows may currently be the more suitable platform in some cases.
And if Mac support comes with structural limitations that affect performance, that’s fine — as long as those limitations are clearly communicated.

At the moment, however, users only see the result (“it’s slow”), while the underlying reasons remain unclear.

The same applies to ARA integration.
ARA has been heavily promoted for years as part of a modern workflow, but in professional post-production environments it still often feels unreliable or difficult to trust in larger sessions.
At some point, the discussion has to move beyond marketing language and into realistic conversations about stability, scalability, and whether these features are actually production-ready for high-pressure workflows.

The bigger issue for me is that the forum currently feels less like a place for constructive software development discussion, and more like an endless cycle of repeated frustration and noise.

What I’d really like to discuss is not just complaints, but how SpectraLayers and Nuendo can actually continue improving in meaningful ways for professional workflows.

[Sophus]

@nabw Here are the most relevant comments from @Robin_Lobel

As I understand it, it essentially boils down to the fact that you would need a separate version of SpectraLayers tailored exclusively to macOS. And this is likely not feasible, as it would double the development effort. And they don’t want to use Python, because it’s not part of SpectraLayers but runs as a separate application.

SpectraLayers 12.0.30 Maintenance Update

Simply because I don’t have any reliable answer to that question, it would be unreasonable to make predictions when there’s no simple technological solution to it, without relying on an entire external Python/Pytorch environment that would need to be installed somewhere on your system, and would not be integrated directly within SL. There’s a reason why the closest competitors to SpectraLayers, RX and Acoustica, don’t provide any GPU acceleration at all. These hardware technologies are highly …

SpectraLayers 12.0.30 Maintenance Update

@al808 Logic is a product developed by Apple, which only work on macOS. So they have access to all internal Apple resources, and they can afford training a model specifically for the mac platform and hardware, using purely Apple technologies, which won’t be compatible on other platforms and hardware. SpectraLayers being cross platform and cross hardware, it has to rely on cross platform technologies, which can’t go as deep in the Apple tech stack as pure Apple technologies. @Cubase-Erklaerbaer …

[al808]

I still don’t understand why SL can’t do better on M series macs.

Here’s a screen shot showing UVR5 giving my cpu & gpu a good workout doing stem splitting (mac mini m4 pro). It’s a freeware app that also has a windows version.

[Sophus]

@al808 Ultimate Vocal Remover is a Python application. As @Robin_Lobel wrote, they don’t want to use Python, for whatever reason.