my suspicion is that steinberg introduced vst 3 merely to clean up the mess that vst 2.x was - from a software design perspective. they simply did not want to deal with that anymore. which i can appreciate. i mean AudioEffectX as subclass of AudioEffect with an AEffect-struct member ...WTF? from a brief skimming over the new VST, the design seems to be much cleaner. had they marketed VST3 like this to plugin developers, the acceptance would probably have been far greater.
as for switching off plugin processing when no signal goes in: VST 2.x had the getTailSize() function for exactly that purpose (@AQ: so it was already possible - regardless whether desirable or not. i actually think that it is desirable. YMMV).
to be clear: i'm not opposed to VST3 - cleaning up a bad design is a good thing. i just don't like the dishonest marketing with all those - supposed to be - new features
How important is VST3 support for you as a customer
-
Music Engineer Music Engineer https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=15959 - KVRAF
- 4390 posts since 8 Mar, 2004 from Berlin, Germany
-
- KVRAF
- 4735 posts since 18 Jul, 2002 from London, UK
Yeah, unfortunately by cleaning up the old (and well known) mess they created a whole new (and somewhat unknown) one...
AudioEffectX and AudioEffect are immaterial at the level of the actual VST standard API (which is all defined in AEffect plus the extra AEffectX enumeration constants). It would have been trivial to create a new 2.3/2.4 100% compatible SDK using the existing enums and classes but without most of that mess (I've done so myself anyhow, but can't redistribute).
True there are some messes & ambiguities in the VST 2.x spec that would require rather more than an SDK reorganization, but they're pretty well recognized and understood by now, and I'll go so far as to speculate that part of the reason 2.x was so widely adopted was because of those ambiguities - a design that is overly strictly specced can seem like so much work to implement at all that developers are put off from doing so, or create fragmentory or competitive standards as some aspect of the spec is incompatible with their design.
To give an example - the VST 2.x spec has never mandated which thread setParameter() is called from. Trivial plug-ins don't need to know or care, so for them the spec is fine. Host vendors don't want to be told which thread it is, as it can be a pain in the a** depending on your host's design if a specific thread is mandated. But for complex plug-ins, especially those with multiple automation code paths & asynchronous messaging, this ambiguity is an annoying mess. However, unless you've got a Comp Sci Ph.D., by the time you ever realise it's a problem, you've already adopted & implenented the SDK and just have to figure out a way to hack around it.
AudioEffectX and AudioEffect are immaterial at the level of the actual VST standard API (which is all defined in AEffect plus the extra AEffectX enumeration constants). It would have been trivial to create a new 2.3/2.4 100% compatible SDK using the existing enums and classes but without most of that mess (I've done so myself anyhow, but can't redistribute).
True there are some messes & ambiguities in the VST 2.x spec that would require rather more than an SDK reorganization, but they're pretty well recognized and understood by now, and I'll go so far as to speculate that part of the reason 2.x was so widely adopted was because of those ambiguities - a design that is overly strictly specced can seem like so much work to implement at all that developers are put off from doing so, or create fragmentory or competitive standards as some aspect of the spec is incompatible with their design.
To give an example - the VST 2.x spec has never mandated which thread setParameter() is called from. Trivial plug-ins don't need to know or care, so for them the spec is fine. Host vendors don't want to be told which thread it is, as it can be a pain in the a** depending on your host's design if a specific thread is mandated. But for complex plug-ins, especially those with multiple automation code paths & asynchronous messaging, this ambiguity is an annoying mess. However, unless you've got a Comp Sci Ph.D., by the time you ever realise it's a problem, you've already adopted & implenented the SDK and just have to figure out a way to hack around it.
This account is dormant, I am no longer employed by FXpansion / ROLI.
Find me on LinkedIn or elsewhere if you need to get in touch.
Find me on LinkedIn or elsewhere if you need to get in touch.
-
- Skunk Mod
- 21249 posts since 10 Jun, 2004 from Pony Pasture
Dr. Seward informs me that Mr. Renfield has been asking for... a cat.Angus_FX wrote:Yeah, unfortunately by cleaning up the old (and well known) mess they created a whole new (and somewhat unknown) one...
-
SuitcaseOfLizards SuitcaseOfLizards https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=2363 - KVRAF
- 10879 posts since 3 Apr, 2002 from Austin, TX USA
Thanks for the fish option, which normally I would choose, but this time I got serious and voted "Don't care".
As long as it works, does what I need, plays well in Reaper and Acid Pro 7.. I don't give a caudal peduncle what version of the VST spec it's using,
As long as it works, does what I need, plays well in Reaper and Acid Pro 7.. I don't give a caudal peduncle what version of the VST spec it's using,
Bandcamp: https://suitcaseoflizards.bandcamp.com/
Linux Mint, Waveform 13 Pro, U-He synths, Audio Damage effects,.
Linux Mint, Waveform 13 Pro, U-He synths, Audio Damage effects,.
-
- KVRAF
- 13140 posts since 7 May, 2006 from Southern California
I rarely use Cubase or Studio 1 so it's not an issue for me at all.
-
- KVRAF
- 14498 posts since 16 Feb, 2005 from Planet Earth, Somewhere
I am curious form an end user/beta tester only perspective....does all the ambiguities in the vst 2.4 spec allow so many incompatibilities with different hosts? I am always amazed how important it is for developers to have to test their plugins in so many hosts and different versions of those hosts...would adaptation of vst 3.X spec decrease the need for that in any respect?
rsp
rsp
-
- KVRist
- 194 posts since 16 Sep, 2004
VST3 has never been a dealbreaker for me, but it can be the feature that mackes me choose one plug over another. I think it may become even more so with 3.5 though, even if it's a little early to know for sure yet.
-
- u-he
- 30250 posts since 8 Aug, 2002 from Berlin
For VST 3.x you only need to test in about 3 hosts...zvenx wrote:I am curious form an end user/beta tester only perspective....does all the ambiguities in the vst 2.4 spec allow so many incompatibilities with different hosts? I am always amazed how important it is for developers to have to test their plugins in so many hosts and different versions of those hosts...would adaptation of vst 3.X spec decrease the need for that in any respect?
rsp
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Nice one Urs!Urs wrote:For VST 3.x you only need to test in about 3 hosts...zvenx wrote:I am curious form an end user/beta tester only perspective....does all the ambiguities in the vst 2.4 spec allow so many incompatibilities with different hosts? I am always amazed how important it is for developers to have to test their plugins in so many hosts and different versions of those hosts...would adaptation of vst 3.X spec decrease the need for that in any respect?
rsp
-
- KVRAF
- 14498 posts since 16 Feb, 2005 from Planet Earth, Somewhere
which I assume is a vast improvement? well in the end only some of us end users will be shafted. Like I said I use Cubendo and audio in and sidechain etc doesn't work unless it is a vst3 spec... there will come a time as more plugins become vst3 that that will be a criteria in my purchasing...glad to know you seem to be on board Urs...
rsp
rsp
sound sculptist
-
- KVRAF
- 1729 posts since 26 Feb, 2008
So, to rehash -
VST3.X is great for devs
VST3.X really doesn't make a difference to consumers, except maybe certain complex plugins experience subtle benefits - and even then only if you use a host that supports VST3 standard anyway
How close am I?
VST3.X is great for devs
VST3.X really doesn't make a difference to consumers, except maybe certain complex plugins experience subtle benefits - and even then only if you use a host that supports VST3 standard anyway
How close am I?
Snare drums samples: the new and improved "dither algo"
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Urs was joking, there only ARE about 3 VST3 hosts TO test in.zvenx wrote:which I assume is a vast improvement? well in the end only some of us end users will be shafted. Like I said I use Cubendo and audio in and sidechain etc doesn't work unless it is a vst3 spec... there will come a time as more plugins become vst3 that that will be a criteria in my purchasing...glad to know you seem to be on board Urs...
rsp
Poly-AnaFX (VST 2.4) is an instrument that accepts audio input which you can process and/or use as modulation sources. There are many, many others as well. (Technically they're actually an effect that accepts MIDI input, not an instrument that accepts audio input, but whatever... same difference.)
Sidechains have been trivial for a long time too, though it did require an extra step or two. (Make a 4 channel track to host your effect. Pan signal to chans 1 + 2 and pan the modulator signal to chans 3 + 4). If you can't do that in newer versions of Cubendo, that's because they broke it, not because it wasn't possible before.
SX 3.1 forever!
-
- KVRAF
- 14498 posts since 16 Feb, 2005 from Planet Earth, Somewhere
That Urs 
, are they really only three VST hosts ok...thought they would be more...
I know for instance, to sample from your host using NI's Maschine, you have to use the Maschine FX version, as an insert in the track you want to sample, save the sound and pull it up in the vst instrument version... I know Angus created a clever software to do all that in one go for Geist, but for the end user that is at the very least inconvenient, for other users like me a PITA....I guess we shall see where all of this settles, but knowing SB I know they aren't going back to VST 2.4 and as more plugins come to vst 3 I have a feeling at some point some will be left behind.
rsp
edit: ah I get it now. There are more than three host that utilise vst, but so far only three have jumped on the vst3 bandwagon!
, are they really only three VST hosts ok...thought they would be more...I know for instance, to sample from your host using NI's Maschine, you have to use the Maschine FX version, as an insert in the track you want to sample, save the sound and pull it up in the vst instrument version... I know Angus created a clever software to do all that in one go for Geist, but for the end user that is at the very least inconvenient, for other users like me a PITA....I guess we shall see where all of this settles, but knowing SB I know they aren't going back to VST 2.4 and as more plugins come to vst 3 I have a feeling at some point some will be left behind.
rsp
edit: ah I get it now. There are more than three host that utilise vst, but so far only three have jumped on the vst3 bandwagon!
Last edited by zvenx on Thu Jan 20, 2011 12:14 am, edited 1 time in total.
-
- KVRAF
- 14498 posts since 16 Feb, 2005 from Planet Earth, Somewhere
There is a reason why some of us moved from hardware to software. for me it is the convenience.... the harder it is to use, the more likely I never use it. case in point rewire... I stopped using reason because it is just so much easier to use a vst.......AdmiralQuality wrote:Urs was joking, there only ARE about 3 VST3 hosts TO test in.zvenx wrote:which I assume is a vast improvement? well in the end only some of us end users will be shafted. Like I said I use Cubendo and audio in and sidechain etc doesn't work unless it is a vst3 spec... there will come a time as more plugins become vst3 that that will be a criteria in my purchasing...glad to know you seem to be on board Urs...
rsp
Poly-AnaFX (VST 2.4) is an instrument that accepts audio input which you can process and/or use as modulation sources. There are many, many others as well. (Technically they're actually an effect that accepts MIDI input, not an instrument that accepts audio input, but whatever... same difference.)
Sidechains have been trivial for a long time too, though it did require an extra step or two. (Make a 4 channel track to host your effect. Pan signal to chans 1 + 2 and pan the modulator signal to chans 3 + 4). If you can't do that in newer versions of Cubendo, that's because they broke it, not because it wasn't possible before.
SX 3.1 forever!
I am not going to do all of that to use sidechain, so much easier for me to find a compressor that works with sidechain in the vst3 spec....
took ppl awhile to move from tape to digital....some still haven't
, but all the tape companies eventually went out of business....nuff said..rsp