Steinberg Discontinuing VST2 Support in its products

[Teksonik]

This was done and failed.

Oh really? I don't remember such an association and I've been around the plugin world since day one. Perhaps you could refresh my memory with some facts about an organization of all the major music software developers coming together to develop standards for the benefit of artists and musicians around the world".

https://en.wikipedia.org/wiki/Generaliz ... _Interface

It was big news in 2003, and kind of died in 2005.

Well my memory has failed me then but....

"Generalized Music Plug-in Interface (GMPI) is a working group of the MIDI Manufacturers Association set up to consider the feasibility of a standard interface for audio and MIDI software plug-ins".

It was the MMA not a group of plugin developers? Were you involved back then? Any other developers beside Jeff of Synthedit fame?

So it failed in 2005. A lot has changed in 17 years. I still think a coalition of developers working together for the common good not only for themselves but for the customer would be beneficial.

Whether it will happen or not is another matter.

[SparkySpark]

This was done and failed.

Oh really? I don't remember such an association and I've been around the plugin world since day one. Perhaps you could refresh my memory with some facts about an organization of all the major music software developers coming together to develop standards for the benefit of artists and musicians around the world".

https://en.wikipedia.org/wiki/Generaliz ... _Interface

It was big news in 2003, and kind of died in 2005.

Well my memory has failed me then but....

"Generalized Music Plug-in Interface (GMPI) is a working group of the MIDI Manufacturers Association set up to consider the feasibility of a standard interface for audio and MIDI software plug-ins".

It was the MMA not a group of plugin developers? Were you involved back then? Any other developers beside Jeff of Synthedit fame?

So it failed in 2005. A lot has changed in 17 years. I still think a coalition of developers working together for the common good not only for themselves but for the customer would be beneficial.

Whether it will happen or not is another matter.

I was a rep in the MMA in 2000-2002. At the time, we mostly dealt with standards for MIDI playback (etc) on cell phones. I don't know, but perhaps they had more related things to work on as cell phones evolved into what we today call smart phones. So maybe they were just too busy for getting on with GMPI.

As Teksonik wrote, a lot has happened in the past seventeen years. It's also worth remembering that the relations between developers and customers have changed completely, not least with the advent of social media and Youtube. ...so hopefully, the end user can have more of say this time around.

[Urs]

Well, GMPI was merely a rather large discussion among plug-in devs (and probably some host guys as well, I remember Ron of Cakewalk to play a major role). The first two weeks were so intense, I had to pull out to keep my sanity. It felt like, 50 people sending messages to a mailing list simultaneously, tens if not hundreds an hour.

Some people managed to hold out and ended up defining a list of claims. Like, what does the ideal plug-in interface for audio look like. Mind you, back then MIDI was thought to be obsolete sooner rather than later. A stance that should not hold true.

I guess it was during a NAMM show that this project was presented to a larger audience, and some time later somehow handed over to the MMA to be the body to govern it. As the story goes, the MMA however rejected the idea that a standard was proposed to them: Apparently they wanted to define it themselves, including defining the API. Which, a the story goes, they never did. IDK.

[Tj Shredder]

So CLAP might become the Linux of plugin standards. GNU agreed finally that Linux was what they always wanted to do but never did. It needs a small core to start such a project. If its convincing enough and the direction is clear, it will evolve…

[SparkySpark]

...
The first two weeks were so intense, I had to pull out to keep my sanity. It felt like, 50 people sending messages to a mailing list simultaneously, tens if not hundreds an hour.
...
I guess it was during a NAMM show that this project was presented to a larger audience, and some time later somehow handed over to the MMA to be the body to govern it. As the story goes, the MMA however rejected the idea that a standard was proposed to them: Apparently they wanted to define it themselves, including defining the API. Which, a the story goes, they never did. IDK.

Interesting, thanks for sharing!

[dionenoid]

Good move by Steinberg. Let's hope that this pushes devs in making full use of the vst3 possibilities, in stead of simply porting vst2 to make vst3 versions with the same feature set.

Fabfilter is already offering vst3 versions with more functionality, NI is slowly implementing new functionality in their latest vst3 updates. My guess is that more will follow soon. A lot can happen in 2 years.

[Markus Krause]

VST3 does not offer significant advantages for synthesizers compared to VST2. The opposite is even true.

Q: Should I migrate my VST2 plugins to VST3?
A: For you as a producer there is no need at all to migrate to VST3. All hosts that currently support VST2 will continue to do this also in the future.
VST3 does offer no significant advantages for synthesizers compared to VST2. The opposite is even true frequently. VST3 does not offer proper MIDI support.

Q: Which plugin-format is the most reliable one?
A: The most popular plugin format is VST2. Nearly all hosts support it natively. As a result VST2 plugins are more well-tested and stable than AU or VST3. Many people say that VST3 is less stable than VST2.

Q: Which plugin-format is loading faster?
A: The VST3- and AU-interface is more complex than VST2. These formats have more overhead and the binaries are bigger. That's why they usually load slightly slower than VST2 plugins.

Q: Steinberg's marketing claims that VST3 plugins offer 'improved CPU performance' compared to other plugin-formats
A: Steinberg's marketing is misleading. They conceal that this is not true for synthesizers (a synth usually does not process incoming audio). In practice their claim is only true for very few effect-plugins that support this feature. Apart from this 'bypassing the processing when no input signal is present' is possible with any plugin-format. We supported this already a decade ago in various VST2 plugins. In practise the CPU performance of a plugin depends on the developer's implementation and not the plugin-format.

Q: After more than a decade VST3 still has not established properly. What's the reason for this?
A: VST3 does not offer significant benefits compared to VST2 for most plugins. Most developers don't like Steinberg's VST3 SDK (software-developer-kit). They say it's badly designed, complicated, over-engineered, didn't come with a proper documentation and taylor-made just for Steinberg's own demands. Even experienced developers have problems to compile it and make it work. Adding VST3 support to existing plugins/hosts using the VST2 SDK is extremely difficult and laborious, since VST3 is not downward compatible at all. It is a completely different plugin-format with serious weaknesses.

Q: Does VST3 support MIDI?
A: VST3 is a plugin-format that is suitable for simple effect-units, but it's architecture does not work well for complex synthesizers. It does not offer proper Midi support. Steinberg continuously ignores the most important industry-standard that exists since decades and that was created by the major players in the music-industry. As a result many VST3 plugins do not offer MIDI-learn or the developers have to do ugly workarounds like exposing hundreds of 'Midi CCs' as fake-automation-parameters. Steinberg is criticized heavily by developers and customers for this critical design-flaw.

Q: Are re-sizeable GUIs only possible with VST3?
A: No. GUIs with different sizes are possible with any plugin format. This is not a VST3 feature.

Q: VST3 offers 'sample accurate automation'. What's the benefit?
A: In practice: None. Extremely accurate automation without glitches is also possible with AU and VST2. The accuracy of automation depends mainly on the implementation of the plugin and not the plugin-format.

Q: Is 'Audio In' for synthesizers a new VST3 feature?
A: No. This is also possible with other plugin-formats.

Q: Is sidechaining only possible with VST3?
A: No. This is also possible with other plugin-formats.

Q: Is multicore-CPU support a VST3 feature?
A: No. This is possible with any plugin-format. We already supported this 20 years ago.

Q: VST3 offers hierarchical plug-in and automation parameter categorization. What's does it mean?
A: It is a useful feature that allows grouping automation-parameters like 'oscillator', 'filter' etc. Sadly there are still many hosts that do not support it.

Q: VST3 offers the possibility to change the number of out- and inputs dynamically. Why do not many plugins use this?
A: This feature is problematic. It can cause serious crashes if the host as well as the plugin do not use a completely correct implementation.

[zvenx]

So basically Steinberg is lying and you are speaking the truth..

I have tried to keep out of threads you are in, but you just go on and on and on..
smh.
rsp

[Markus Krause]

Please note that our products are available in several plugin formats: VST2, VST3 and AU. All licenses are cross platform. So I don't care which format is used.
I spent the last 20 years in coding for the interfaces and know them very well

[MirkoVanHauten]

So basically Steinberg is lying and you are speaking the truth..

Don't think that advertising cosmetics overhaul or updates to working stuff is technically lying, rather presenting it as innovative is an issue. I mean they even say it: VST3 factory replaces the VST2 shell plugin format (so it already was possible) or VST3 instroduces a new approach to GUI stuff (and yes resizing VST2 exist); is sample accurate automation impossbile with VST2? No, FL does it for example. These MIDI points are new? No, VST2 could do this because IT WAS ABLE TO HANDLE PURE MIDI EVENTS, there was no workaround needed, no need to introduce this stuff.

I can confirm most of the stuff Markus wrote, or Urs in other threads about VST3 limitations and what already was possible with VST2.

[zvenx]

"I can confirm most of the stuff Markus wrote, or Urs in other threads about VST3 limitations and what already was possible with VST2." yes, as workarounds/hacks.

I just said this elsewhere and have said it on kvr at other times, maybe not in this exact way.

I beta test for SB. I have for over 10 years I believe. Unlike most here I get to hear the other side of the argument, most at kvr only get to hear one side of that argument since there are many no fans of vst3 developers here, and SB for whatever reason, only participate in the developer forum here.

What I can say from these years of interacting with SB developers is that: 1) I am not aware there is any financial gain (beyond extra time to make sure Cubendo is both vst2 and vst3 compatible) of SB pushing vst3.. They really do believe it is much better tech than vst2. More structured, more efficient, with more possibilities of adding new functionality.
2) it isn't that they pushed vst3 and are just to shame to admit now that vst2 was better, as I have seen inferred not on this thread but other threads here at kvr. 3) (I had other points but can't remember them now )

rsp

[zvenx]

There is a post either on this thread or another thread by a Reaper's developer which I think is the most balanced of the developer's post I have seen, looking for it now..


edit:

Found it:

Screen Shot 2022-02-03 at 9.12.28 AM.jpg

viewtopic.php?p=8331849#p8331849


rsp

[Erisian]

1) I am not aware there is any financial gain

Do devs not have to pay for a new license to build vst3? Commercial companies usually create new things to make more money, not out of the goodness of their hearts.

[zvenx]

No they don't..
there is afaik no fee to use vst3.. you have to sign an agreement of use*, but no fee afaik.

https://developer.steinberg.help/displa ... +Licensing

(I think they are two licenses, one you have to sign an agreement with SB for use, and one if you are making a plugin that is open sources, where you don't have to sign that agreement with SB but the condition is you have to make your source code open sourced... the faq spells it out I believe.
rsp

[Funkybot's Evil Twin]

zvenx, they potentially gain in marketshare when Cubase and Neundo are the DAWs that best support VST3 [and by that I mean, they're the only ones who know how it's supposed to work due to the lack of documentation]. There's also nothing preventing them from killing off VST2, waiting for VST3 to become "the new standard", then changing the licensing requirements in the future. They could start charging devs for the ability to compile plugins, or ask for a cut of sales...who knows what the future holds? It sounds like hysteria and is a worst-case scenario, but corporations have done worse things to abuse monopoly positions than that. And that's what Steinberg is trying to do: make VST3 ubiquitous like VST2 has been. If it's the only game in town after killing VST2, what's to stop Steiny from abusing their marketshare by changing their licensing terms? Nothing.

You're also a beta tester for Steinberg, which is great, but keep in mind, you may be getting the corporate line. Steinberg employees certainly 1) may indeed have a more rosy view than developers and 2) may be inclined not to share their real thoughts with a beta tester, if their real thoughts are not positive. No ones wants to lose their job for sharing their true opinions about the company they work for with people outside the company. Certainly not in an electronic communication. They also simply may know more about the format they created than everyone else who has to suffer through the poorly documented SDK.

The fact of the matter is that the people who do development, like Markus, Urs, and dozens of others are all pretty much saying the same thing. I've never heard a commercial developer come out and praise VST3 for its ease of use and terrific feature set. I'll bet there's one out there somewhere, but there's a few nutjob scientists who also ignore the consensus in the scientific community that climate change is man-made. The consensus in the plugin development community is basically that VST3 is not a good replacement for VST2 and took many steps backwards. Despite Steinberg's strong assertions otherwise. If it truly was "the better format" it would've become the dominant format on its own. I trust the majority of people who make living developing plugins, aren't lying about it just to make Steinberg look bad.