Is VST3 worth it?

[PeterP_swe]

I'd prioritise the use of VST3 plugins if there was a tangible benefit to doing so - e.g. lower latency, less CPU strain, etc. - but if there is nothing like that on the table, why care or bother with it?

Quoting from Steinbergs page: "VST3 helps to improve overall performance by applying processing to plug-ins only when audio signals are present on their respective inputs. Instead of always processing input signals, VST3 plug-ins can apply their processing economically and only when it is needed."

Other hosts may do that optimization for VST2 plugins as well, but Cubase apparently does not.

[EvilDragon]

In VST2 that's not necessarily a host optimization, the plugin can stop their processing if there's no input signal at all.

[BDeep]

Ableton 10.1 actually got VST3 support.

Also ARA support is not directly tied to VST3, it can be implemented in VST2.4 plugins as well.


VST3 suffers from over-engineering for no good reason whereas VST2 is a much simpler SDK overall. I wouldn't call it redundant and outdated, many of the things VST3 touts were possible in VST2 as well, and in fact nearly all the things VST3 touts could've been a part of VST2.5 without completely changing how the plugins are built and how the SDK is laid out. But no, why should things be simple when they can be complicated?

You and your facts, you're taking away all the fun in mindless bashing

[Functional]

and in fact nearly all the things VST3 touts could've been a part of VST2.5 without completely changing how the plugins are built and how the SDK is laid out

It is what it is now. Although I agree, I'm not sure what was the rationale behind this particular thing but there's no going back now from that. And that considered, my main point still stands as far as I'm concerned

[EvilDragon]

It was marketing and trying to get a stranglehold on the market, I assume. Typical Steinberg thing. Well, it didn't quite work out as they'd hoped.

[andypryce]

I think also with vst3, automation shows real values instead of generic 0-100

[EvilDragon]

That's possible with VST2 too.

[samsam]

and in fact nearly all the things VST3 touts could've been a part of VST2.5 without completely changing how the plugins are built and how the SDK is laid out

It is what it is now. Although I agree, I'm not sure what was the rationale behind this particular thing but there's no going back now from that. And that considered, my main point still stands as far as I'm concerned

Apart from the bit about Ableton.

[Functional]

Apart from the bit about Ableton.

I said main point, which wasn't really the bit about ableton. I don't hold much against Ableton, aside from missing customizable shortcuts. Or maybe newer version does have that too in.

[Googly Smythe]

Maybe the major devs could get together with Steinberg and come up with version 4, all the benefits of 2 and 3 and none of the drawbacks.
Wow, a flying pig!

[Functional]

And then version 5 with the tagline "I can't believe it's not VST4!"

[BONES]

I could write this in a long story but I decided to do it in a short story: if every major DAW would support VST3, then plugin developers would only have to develop for VST3 and it wouldn't matter whenever they would benefit from VST3 format itself or not.

So no reason at all, then. As I suspected. Is it harder to support VST 2.4 than it is to support AU? Everyone seems to bother with AU versions when most Mac hosts have VST support.

Quoting from Steinbergs page: "VST3 helps to improve overall performance by applying processing to plug-ins only when audio signals are present on their respective inputs. Instead of always processing input signals, VST3 plug-ins can apply their processing economically and only when it is needed."

That sounds like it would only apply to effects, which aren't the things that use up all your CPU in the first place. So, again, no reason to care.

[Mushy Mushy]

Effects don’t use much CPU? Is that what was just said?

Molekular, Softube Modular, B2, Kaleidescope to name but a few.

[BlueprintInc]

Native sidechain support reduces the need for different binaries. This was the case in VST2 when e.g. have a 3 input channels - what is it, stereo in + sidechain or LCR or mono with 2 sidechains? Also VST3 has managed to improve the whole channels and bus management.

Audio input for instruments as well as sample accurate automation is quite useful too. The latter one is possible in VST2 too if the host slices the buffer. But as seen in case of FL Studio which does this perfectly, many plugins can't handle this.

VST3 parameters are not bound to their order anymore, but to IDs which allows it to add or remap them without destroying projects.

Plugin shell handling has become more easy.

Is it harder to support VST 2.4 than it is to support AU? Everyone seems to bother with AU versions when most Mac hosts have VST support.

AU is pita. Not only do you have to learn a new language (Objective C), it is also much more cumberstone. And the fact that Apple changes things frequently with every OS update does its thing too.

[jancivil]

Apple has sought to somewhat catch up with VST3 with AU3. Unfortunately the release may as well be a public beta. as pertains to uses I would have anyway.

AU: no midi to the outside world; restricted to one port (so Logic had an awkward workaround to get more than the one port in an instance of VE Pro), there are others which wouldn't occur to me