VST3: Is it worth supporting?

[sqigls]

I as a Cubase user strongly vote for VST3 support. The easy sidechain setup alone is worth it. Whereever supported I use VST3 plugins, like Voxengo, Melda, Waves, etc...

As opposed to the 15 seconds it took to set it up the old way? Side-chaining is not a new feature.

The 'old way' involved creating a quadra bus and sending to the back 2 channels of it, after making visible the back 2 and re-routing bla bla... In use, it was a bit convoluted compared to just activating and sending straight to the insert of a channel. Unnecessary.

[AdmiralQuality]

I as a Cubase user strongly vote for VST3 support. The easy sidechain setup alone is worth it. Whereever supported I use VST3 plugins, like Voxengo, Melda, Waves, etc...

As opposed to the 15 seconds it took to set it up the old way? Side-chaining is not a new feature.

The 'old way' involved creating a quadra bus and sending to the back 2 channels of it, after making visible the back 2 and re-routing bla bla... In use, it was a bit convoluted compared to just activating and sending straight to the insert of a channel. Unnecessary.

And how many music software products have YOU developed, sqigls?

It was absolutely necessary, as that was the way to do it.

Sure, it's a place that could have been improved. But NONE of the things that are awful about VST 3.x were required to implement that.

I'd rather take a minute to set up a sidechain than lose MIDI and VST 2.4 compatibility (which, again, Steinberg CLAIM can be produced from a VST 3 project, but they're f**king LYING).

[helium]

I my opinion MIDI sucks. I use it because I have to use it, not because I want to use it. In general I think decoupling plugins from MIDI is a good thing. Now you can support different more powerful protocols which the host translate to one common format that the plugin has to support.

[Juanjo]

From my mixed state of ignorance and knowledge... isn't the situation getting fast to a point where it will be either swallow vst3 or start working towards an open (as opposed to proprietary) and feasible standard for audio plugins? I don't see Yamaha backpedaling ever, they seem to be way down that road...

I'm getting into audio software development this year and vst3 looks like a no-go to me, and most devs I've read/heard on the issue seem to more or less agree (I struggle to find any indie devs who actually like the damn thing overall).

[AdmiralQuality]

I my opinion MIDI sucks. I use it because I have to use it, not because I want to use it. In general I think decoupling plugins from MIDI is a good thing. Now you can support different more powerful protocols which the host translate to one common format that the plugin has to support.

No, it's a bad thing. Because wrapping one standard in a non-standard means you're only as strong as the implementation of the non-standard. And Steinberg's MIDI support is a VERY weak link.

So it's not "more powerful" if you can't do things you could do before.

Again helium, you claim to be a developer, how did your VST3 project work out? Because THIS developer couldn't get even a sample project to compile. (I haven't tried in a few versions, but I still closely monitor all the suckers going through hell with it and having to hack-in their own fixes. When they stop reporting immense problems, I might take another look at it. Maybe. Again, nothing was "broken" about my plug-ins before that needed "fixing" by VST3, so I resent the perception that somehow our customers are missing out. They're not. We're sticking to 2.4 FOR our customers.)

[AdmiralQuality]

From my mixed state of ignorance and knowledge... isn't the situation getting fast to a point where it will be either swallow vst3 or start working towards an open (as opposed to proprietary) and feasible standard for audio plugins? I don't see Yamaha backpedaling ever, they seem to be way down that road...

I'm getting into audio software development this year and vst3 looks like a no-go to me, and most devs I've read/heard on the issue seem to more or less agree (I struggle to find any indie devs who actually like the damn thing overall).

VST 2.4 is that "open" standard. (And in fact, it's particularly secure as it's no longer being developed, so nobody can f**k it up. It's frozen in time at a period of maximum stability and functionality. If we tried to make a competing open standard by committee we'd just make an even worse abomination than VST 3. Not to mention making everyone's jobs that much harder by introducing yet another standard that they'd presumably feel compelled to support.)

Yama-berg pissed away their industry lead. Just ignore them, we don't need them for anything.

[fmr]

VST 2.4 is that "open" standard. (And in fact, it's particularly secure as it's no longer being developed, so nobody can f**k it up. It's frozen in time at a period of maximum stability and functionality. If we tried to make a competing open standard by committee we'd just make an even worse abomination than VST 3. Not to mention making everyone's jobs that much harder by introducing yet another standard that they'd presumably feel compelled to support.)

Yama-berg pissed away their industry lead. Just ignore them, we don't need them for anything.

I am no developer too, but I have to agre with Admiral about this. There is a rule in the industry that, unfortunately, has not been followed very much lately: If it aint broke, don't fix it.

Also, a developer friend of mine once told, and I agree: A "standard" is a code that is running flawlessly for over ten years.

We have that in VST 2.4. Maybe we need a couple of things, but I sincerely hope that hosts do not abandon supporting it. I'm told there is a problem with the SDK (Steinberg is no longer licensing it) and with some legal stuff. I don't know if these can be solved, but if they can (I mean, if we can still get a working SDK and the legal problems get sorted), then VST 2.4 should absolutlely be kept, ate least for the time being, until Steinberg perfects their new "standard" (it is far from ten yars yet).

And yes, this is the abominated Apple business model being transposed

[helium]

Again helium, you claim to be a developer, how did your VST3 project work out? Because THIS developer couldn't get even a sample project to compile.

Just for you I tried compiling the sample again projects. Here is the result not even 5 minutes later:



Both Vst2 and Vst3 versions running.

[AdmiralQuality]

There is a rule in the industry that, unfortunately, has not been followed very much lately: If it aint broke, don't fix it.

Or Yamaha/Steinberg/Apple/ProTools style... if sales are down, BREAK IT AND MAKE 'EM BUY IT AGAIN!

Also, a developer friend of mine once told, and I agree: A "standard" is a code that is running flawlessly for over ten years.

We have that in VST 2.4. Maybe we need a couple of things, but I sincerely hope that hosts do not abandon supporting it. I'm told there is a problem with the SDK (Steinberg is no longer licensing it) and with some legal stuff. I don't know if these can be solved, but if they can (I mean, if we can still get a working SDK and the legal problems get sorted), then VST 2.4 should absolutlely be kept, ate least for the time being, until Steinberg perfects their new "standard" (it is far from ten yars yet).

And yes, this is the abominated Apple business model being transposed

Thanks for saying so. Nice to see some non-developers get this as well. If Steinberg were telling the truth it would be great. But they're not. They're serving Jonestown Kool-Aid to whoever will drink it.

And no worries yet about the 2.4 SDK. It's currently still included in the VST 3.5 SDK (if buried, and with the sample projects removed). And I don't think any new players will ever have that much trouble locating a copy, even if Steinberg stop publishing it. Steinberg have no way of knowing where anyone got their SDK copy.

Frankly, I think Steinberg is going to be gone in a few more years, if not sooner. They've f**ked themselves. (Or their Japanese masters have. Whatever.)

Oh and VST 3 will be 10 years old in two more years. Yes, it's THAT old and has been uselessly bug ridden the whole time. It's been out for 8 years and still virtually zero adoption.

[AdmiralQuality]

Again helium, you claim to be a developer, how did your VST3 project work out? Because THIS developer couldn't get even a sample project to compile.

Just for you I tried compiling the sample again projects. Here is the result not even 5 minutes later:



Both Vst2 and Vst3 versions running.

As I already mentioned, my experience was with VST 3.0, 6 or 7 years ago. I rejected it then, and haven't found reason since to be confident enough in to try it again. My time is valuable.

Your screen cap doesn't let me test them. And wow, one parameter and no MIDI (or whatever) implementation. That definitely proves it all works.

[fmr]

Oh and VST 3 will be 10 years old in two more years. Yes, it's THAT old and has been uselessly bug ridden the whole time. It's been out for 8 years and still virtually zero adoption.

Well, I have seen developers reporting problems, but I would not say "still virtually zero adoption"

There are quite a few already adopting, and some have been severe critics of the norm, like Urs (U-He). Waves is supporting it, FabFilter also, and I know a few others, whose names I don't remember right now (I'm away from my computer). For what I was told, there are ways to support it without implementing all the features, like the mentioned "note expression", right from the beginning.

[Juanjo]

VST 2.4 is that "open" standard...

Right now sure, but that's my point... I'm not saying we're "there", I'm just saying we are rapidly approaching that crossroads because VST2.4 is open as in "everybody has access now", but it's not "open" as in future-proof because it's proprietary and Yammy are making quite clear that they are moving away from it... sure you still can kinda get it with the vst3 sdk, but for how long?

Imho a "standard" which no new devs can adopt is not really a standard, isn't it? And afaics, vst2.4 IS going that way...

[AdmiralQuality]

Oh and VST 3 will be 10 years old in two more years. Yes, it's THAT old and has been uselessly bug ridden the whole time. It's been out for 8 years and still virtually zero adoption.

Well, I have seen developers reporting problems, but I would not say "still virtually zero adoption"

There are quite a few already adopting, and some have been severe critics of the norm, like Urs (U-He). Waves is supporting it, FabFilter also, and I know a few others, whose names I don't remember right now (I'm away from my computer). For what I was told, there are ways to support it without implementing all the features, like the mentioned "note expression", right from the beginning.

Don't take my word for it. Read this entire thread. Like I said, there's a very rare consensus among developers on this one.

And I qualified my statement with "virtually" for a reason. Proportionally, VST3 doesn't chart. (Just look at all the popular HOSTS that don't support it. And that's after it's been out for 7 or 8 years!)

And sorry, if you want to interface with controllers, "note expression" or whatever they call it is it. There's no choice other than abandon instrument control entirely, or implement "note events" and abandon MOST OF the control you used to have. Doing nothing is not an option. Some of us have existing products that would LOSE functionality if implemented as VST 3.

[AdmiralQuality]

VST 2.4 is that "open" standard...

Right now sure, but that's my point... I'm not saying we're "there", I'm just saying we are rapidly approaching that crossroads because VST2.4 is open as in "everybody has access now", but it's not "open" as in future-proof because it's proprietary and Yammy are making quite clear that they are moving away from it... sure you still can kinda get it with the vst3 sdk, but for how long?

Imho a "standard" which no new devs can adopt is not really a standard, isn't it? And afaics, vst2.4 IS going that way...

I don't think you know what proprietary means. They can't stop us from using the VST 2.4 SDK for perpetuity. And like I said, even if they stop distributing it, it's out there. My copy was obtained legitimately, and will be forever valid for me to use forever. As for those who will eventually get their copy through illegitimate channels, there's no way for Steinaha to know who those developers are. So even if it was enforceable (which it's not) they wouldn't know who to attack.

[hokkaido]

I my opinion MIDI sucks. I use it because I have to use it, not because I want to use it. In general I think decoupling plugins from MIDI is a good thing. Now you can support different more powerful protocols which the host translate to one common format that the plugin has to support.

No, it's a bad thing. Because wrapping one standard in a non-standard means you're only as strong as the implementation of the non-standard. And Steinberg's MIDI support is a VERY weak link.

So it's not "more powerful" if you can't do things you could do before.

Again helium, you claim to be a developer, how did your VST3 project work out? Because THIS developer couldn't get even a sample project to compile. (I haven't tried in a few versions, but I still closely monitor all the suckers going through hell with it and having to hack-in their own fixes. When they stop reporting immense problems, I might take another look at it. Maybe. Again, nothing was "broken" about my plug-ins before that needed "fixing" by VST3, so I resent the perception that somehow our customers are missing out. They're not. We're sticking to 2.4 FOR our customers.)

I don't really like VST and wished there was a truly open standard. Wouldn't it be cool if there was such a standard, with a working group, a bug tracker, github repo, good documentation and a FOSS compatible license? Unfortunately this will probably never happen.

Now, you seem to be on a crusade with VST3. Being less prickish about your dislike about VST3 might bring you more converts imho.

That said, as a programmer I'd also be hesitant to just rewrite my entire code base just because there is a new specification.

I'm a software developer by day but I haven't touched dsp programming and c++ in a long time, but even I was able to build their VST3 sample projects and adapt them without a problem. Of course I'm also experienced enough that this doesn't mean shit. Dabbling in personal projects isn't the same as truly working with a component / system / library for a prolonged time and learning all its quirks and problems. From what I've read one of the main problems with VST3 seems to stem from the asynchronous communication between processor and controller.

Of course, coming back to my main point, the problem aren't so much the bugs but the closed-source nature of the specification and SDK in the first place. And this is as much a problem with VST 2.4 as with VST 3+