Bye bye VST2

[Urs]

Show the data.

I feel tempted to harvest that data just for the fun of it. We should place bets!

If we had a script that scans VST3/AU plug-ins for markers of VST2... say, something like entry points to VST2 called from within portions that entry points of AU or VST3 (does a com layer have that actually?) point to... how has the percentage shifted over the years, particularly since JUCE became popular, and how many plug-ins from the time prior to that are still installed on user's computers? If and when were AU plug-ins wrapped with Symbiosis the majority, and how has that changed since the move to Apple Silicon? Have they moved to a different wrapper?

Ah, here's some discussion about different wrappers, actually using JUCE as a wrapper:

Using JUCE to wrap older VST 2.4 plugins

What are the bets that tens, hundreds or maybe even a thousand developers chose such an approach once Symbiosis seemed abandoned?

But then you, know, I personally am happy with the feedback from the developer community. I don't really feel like asking any of my employees to write scripts just because someone on the internet doesn't like my insights.

[Urs]

Argumentum ad verecundiam isn't evidence.

I would like to see the data from this survey. That would be evidence.

So I guess I should have taken notes then?

[Music Engineer]

What I am actually concerned with is future-proofing and finally seeing some stability.

These concerns are actually precisely what CLAP addresses (and they are my concerns, too). For a future-proof plugin format, you need two ingredients: (1) permissive licensing, (2) simple and extensible API. CLAP ticks both of these boxes whereas VST3 ticks none.

I stopped buying plugins that weren’t VST3 and swapped out VST2 for VST3 in every song project I have a long time ago. So I’m not lamenting now, because I prepared a long time ago. And that’s what I’ve advocated here for others for a long time now.

Aha! So that's the reason why you seem to be so emotionally attached to VST3. You invested a lot of work and effort into getting your setup up and running with VST3 exclusively and now it looks like all this effort may have been wasted if CLAP catches on? ...but I wouldn't worry too much. VST3 is not going to go away anytime soon even with CLAP on the rise. Plugins manufacturers will still target VST3 (and even if not, there will be a wrapper) - but as a replacement for VST2 inside the internal development framework, it's just not suitable. VST3 is too - let's say - opinionated. ...and complicated. It's kinda like a whole framework that may interfere with one's own. And then there are the licensing risks.

[muzicxs]

Presented without evidence.

Before you demand evidence from others, where's your evidence? You were already called out, but blatantly ignored the request to supply evidence.

The reality is that over 90% of developers had no hesitation in adopting VST3 15 years ago.

I guess if you're going to lie you might as well lie big.

that's just lolworthy.

"more than half" wouldn't look half as stupid. did you just decide to yolo it?

[seafire]

What I am actually concerned with is future-proofing and finally seeing some stability.

These concerns are actually precisely what CLAP addresses (and they are my concerns, too). For a future-proof plugin format, you need two ingredients: (1) permissive licensing, (2) simple and extensible API. CLAP ticks both of these boxes whereas VST3 ticks none.

You also need everybody on board. Every dev of every host and plugin . That is a long way off.

[seafire]

Argumentum ad verecundiam isn't evidence.

I would like to see the data from this survey. That would be evidence.

So I guess I should have taken notes then?

Always. He will want to see your accounts too after you threw in the 100k plus figure

[Urs]

Argumentum ad verecundiam isn't evidence.

I would like to see the data from this survey. That would be evidence.

So I guess I should have taken notes then?

Always. He will want to see your accounts too after you threw in the 100k plus figure

We're in the EU. There's public record for that.

[gaggle of hermits]

My actual experience with automating physically modeled bowed instruments is that instruments built with VST2 are not capable of consistently reproducing bowed articulations correctly, while the VST3 versions are.

So in real-world application, VST2 is deficient for my needs.

aw, gee, and you used to be such a fan of arché: viewtopic.php?p=7975040#p7975040

maybe it has sod all to do with vst2 given that swam strings runs as both vst2 and vst3? do you really think vst3 is giving the incoming midi a good polish but the vst2 just barfs it out?

what bizarre factoid will jamcat come up with next? find out later today.

[Music Engineer]

What I am actually concerned with is future-proofing and finally seeing some stability.

These concerns are actually precisely what CLAP addresses (and they are my concerns, too). For a future-proof plugin format, you need two ingredients: (1) permissive licensing, (2) simple and extensible API. CLAP ticks both of these boxes whereas VST3 ticks none.

You also need everybody on board. Every dev of every host and plugin . That is a long way off.

Yes - I should probably have said "at least" two ingredients. I already had the hunch that there might be more. VST3 currently has the lead in the adoption rate over CLAP - that is true. But it was adopted very reluctantly and it's not really the case that all the adopters are "on board" in the sense that they really like the format for its merits. It was mostly adopted out of apparent necessity and Steinberg had to exert some force to get there - in 15 years. And there are good (technical and legal) reasons for that. CLAP is completely new. This has to be factored in when comparing the adoption rates. In the long run, I'm pretty confident that CLAP will win the race. And there will be wrappers anyway. As said: VST3 will not go away. It will still be targeted. It just can't replace VST2.

[ampetrosillo]

It has a lot of deficiencies. And it's not nearly as widely used as Microsoft software.

VST3 has sample-accurate automation, VST2 does not.
Automation with VST2 is sloppy, to say the least, and it was incurable without a rewrite. That alone is reason enough for VST3.

Let's even assume that this is true. How often is sample-accuracy essential? Most automation essentially consists in "riding the faders". Whenever does this need to be sample accurate? It's like presenting BEVs as a superior vehicle on account of their 0-60 times. Yes, you can easily get sub-4s times with EVs (while essentially wasting a lot of energy), but who actually needs this? VST3 may even be superior in some respects, but if (for instance) the licensing is a show-stopper then it counts f**k all.

[jamcat]

Aha! So that's the reason why you seem to be so emotionally attached to VST3. You invested a lot of work and effort into getting your setup up and running with VST3 and now it looks like all this effort may have been wasted if CLAP catches on?

No, if CLAP truly caught on then I would just be covered by both VST3 and CLAP.

While I try to explain to people why that will likely never happen, and just how long and slow of a process it would be if it ever actually were to happen, that doesn't mean I'm anti-CLAP. Because I'm not anti-CLAP. I'm just pro-reality. And someone needs to bat down the ridiculous anti-VST agitprop from the CLAP zealots. Believe it or not, VST3 plugins really can be built. They actually are real, and work quite well, despite the nonsense some people on KVR with an agenda would have you believe. Just maybe not in Ableton Live.

[Blue Cat Audio]

Just my two cents, from our customers' standpoint (as a developer I could unfortunately write pages about the topic):

• VST3 is still missing features that were available in VST2, even after 16 years (and they will probably never be implemented). So for several of our products I just cannot recommend it over VST2 or any other format. The whole MIDI support thing is incomplete and such a mess that it is a shame and will never work flawlessly.
• VST3 plug-ins and hosts are still not as stable as their VST2 counterparts in most cases (we do have both hosts and plug-ins so we have a pretty long experience and a large number of reports on this topic). I won't argue about the reasons here, but it is not likely to change anytime soon, that's a pity.
• Many VST2 plug-ins will never get translated into VST3 or any other format (either developers have stopped working on them, or companies have gone out of business, or these are simply older versions of plug-ins that still exist). Keeping backward compatibility for these seems like the minimum thing you can do for your customers. If VST3 had been compatible with VST2, it would have been a different story of course.
• The fact that the only way to kill VST2 after so many years is the legal way stands for itself. Back in the day, DirectX plug-ins just faded away without the need for anyone to deprecate this format (you can actually still build DirectX plug-ins, but who does?).

All in all this is a sad situation for everyone, and whatever we developers think about it, the problem is that the impact will be painful for customers, and that's what I am worried about.

BTW, as far as I understand, if you do not upgrade to the latest VST3 SDK and do not receive any notice from Germany about the termination, you can still release VST2 plug-ins and hosts.

[AdvancedFollower]

It has a lot of deficiencies. And it's not nearly as widely used as Microsoft software.

VST3 has sample-accurate automation, VST2 does not.
Automation with VST2 is sloppy, to say the least, and it was incurable without a rewrite. That alone is reason enough for VST3.

Let's even assume that this is true. How often is sample-accuracy essential? Most automation essentially consists in "riding the faders". Whenever does this need to be sample accurate? It's like presenting BEVs as a superior vehicle on account of their 0-60 times. Yes, you can easily get sub-4s times with EVs (while essentially wasting a lot of energy), but who actually needs this? VST3 may even be superior in some respects, but if (for instance) the licensing is a show-stopper then it counts f**k all.

It's certainly not unheard of to draw in automation manually instead of riding faders, including instant vertical changes on the automation lane to coincide with transients which can be just a few milliseconds long. Usually there are more efficient ways to deal with such automation, such as sidechaining, MSEG's etc. but like I said, manually drawn automation is not unheard of.

[jamcat]

Argumentum ad verecundiam isn't evidence.

I would like to see the data from this survey. That would be evidence.

So I guess I should have taken notes then?

Always. He will want to see your accounts too after you threw in the 100k plus figure

Why wouldn't I believe that? It's not really impressive. At ~20 employees, that's only 5k each per month. Arturia brings in over $2 million per month, perhaps $4 million. IK Multimedia brings in $1.5 million per month. And you can be sure neither of them are hacking VST2 plugins into a VST3 wrapper.

[jamcat]

• The fact that the only way to kill VST2 after so many years is the legal way stands for itself. Back in the day, DirectX plug-ins just faded away without the need for anyone to deprecate this format (you can actually still build DirectX plug-ins, but who does?).

Well, it did take Steinberg abruptly dropping DX support from Cubase 4 (the one that came after SX3) to finally shift the industry away from it on Windows.