I try to stay on top of what’s new because being on top of the tools is one way for me to separate myself from every other bedroom engineer out there. But I agree most of the big moves in DSP in recent years have been more about workflow and less about the actual DSP getting better. For instance @Artificer brought up Valhalla Vintageverb. I agree, the number of reverbs that I think sound straight up better than Vintageverb I can count on one hand, and they’re all expensive - eventide SP2016, a good Lexicon emulation (there are few, including some by Lexicon), and Valhalla Room when you need something more neutral. So two emulations of vintage units and another Valhalla product… And the other reverb I’m interested in is VSS4 HD native, which is an emulation of another 20+ year old hardware unit.
But one that sounds almost as good and gets me where I want to be faster is Izotope Neoverb. I picked that up for like 10 bucks during black friday last year and it quickly took over most of my FX tracks. I can use 3 instances of Vintageverb on multiple busses running into each other to get the sound I’m after, or I can slap Neoverb on one bus and tweak a few sliders after I hit the “reverb for me” button.
There are a few tools pushing at the edges of what can be done with classic processing ideas. If you look at Tone Projects Kelvin, it’s a nice saturator/distortion. More or less gives you the flexibility of Saturn with more broad strokes controls. But the magic slider inside of there is a “width” control which takes the harmonics you’re adding and puts them in progressively more and more of the sides signal, which means that during mastering I can make a track wider without touching the original audio by just blending in a little saturation on the side channels. That’s pretty cool, but is that really what 10 years have gotten us from Decapitator?
I think the actual biggest jump in DSP in recent years was zero-delay feedback filters being implemented in DSP for the first time… which happened 10 years ago in Diva for the first time. And it’s still not something I can run on more than one instrument at a time. Finding a way to accelerate that workload algorithmically in a way that’s easy for a CPU to run natively (without going through an abstraction layer to do calculus) is probably the next big jump (outside of AI powered stuff).
Also just a shoutout for the Output FX, I got Portal last year and that made granular FX fun and easy to automate in my tracks. A bit expensive for mostly a preset machine, but I have multiple songs built around ideas that came out of playing with it so yeah… Thermal looks fun too. I’d like to see something with a similar interface for reverb, kind of like a souped up Neoverb…
Oh, and if you don’t have one, get an EQ that can oversample. I didn’t think that would make a difference until I tried to EQ the top end with versus without oversampling. Every EQ plugin that has a rep for a “smooth” or “natural” or “effortless” top end is just the sound of oversampling keeping the reconstruction filters away from whatever EQ moves you’re trying to make.