Plugins will NEVER get better…. This is why... 

What is getting better is the realistic 3D shading on my screws in my plugins

If there’s one thing the industry can get behind together it’s definitely rereleasing every plugin in a new format that we need to purchase again.

CLAP support is on the way for FL, it's currently being tested in the 2024 beta

Your studio is looking very nice these days.

I have to say this is a pretty nice idea. The problem with most developers are they're obsessed over re-creating 60 year old designs and workflows in a computer. "The eagles didn't have 1176s talking to each other - why would we need to have plugins talk to each other?" We need more developers like DMG, Fabfilter, Goodhertz etc who do not start with the ancient workflow and start fresh with a perspective with current needs and processing capabilities in mind.

the genius of Dave Smith’s midi. RIP

I think the open CLAP standard is seeking to solve what you're talking about.

Reaper can handle upsampling a whole chain of plugins like that already, and then downsampling the chain, to my knowledge. That's not solving any problems mentioned otherwise, but quite a fine thing.

I am nobody in this sector, only i play piano (amateur), but i appreciate you so much for your integrity, i like to see your videos for it, you are so authentic!!! My great compliments to you

Many have already said this but CLAP is basically the plugin standard equivalent of a standard like Vulkan for graphics. Open Source, developed and supported by multiple DAW vendors (Bitwig, Presonus) and actually makes sense from a API sense. Something to also add is the DAWProject format that Bitwig and Presonus have introduced. Another open-source standard that actually manages to do cross-daw project sharing (no more stem/file exporting!).

Audio processing tech has always been behind. I've been saying this for years. I'm so glad you're saying this!

FL has Clap support in beta I hope to see it in Studio One sometime

A lot of the things you described are already possible, but not in plugins, but DAW features. For example in Cubase Pro you can apply an offline processing effect chain to any audio event you want. It can include any realtime plugins you have and if you tweak a parameter in the middle of the chain it will automatically re-render the interleaved audio from that effect on. The only problem is that they didn't add this feature for a whole audio track, but I'm pretty sure they could do that if they wanted to. They already have implemented the pretty similiar track freeze feature. Another thing you mentioned was the oversampling thing. Why can't a plugin tell the next plugin that they could share the same oversampler to save some resources and sound quality? Well, Reaper has a feature, where you can wrap a whole effect chain in an oversampler. And Bitwig has even more useful containers, like a mid/side-container or several multiband ones, that make it easier to use small plugins, that don't bring a lot of features themselves. So all in all I think you should just focus on DAWs instead of plugins with the things you want. DAW features can solve all these problems without a new plugin format. The job of an audio plugin is just to synthesize or process audio or MIDI and it has all the available tools and information available for doing that already.

Fl Studio is about to support CLAP. In 2 years since its release 2 major DAWs adopted CLAP and there's others thinking about adopting it. I truly believe CLAP will be the future but it may take a while.

TLDW: Wytse argues for a modern global universal standard for plugin development, DAW implementation and hosting. This means that plugins, DAWs and instruments would be able to communicate with one another and know what kind of processing is going on across a whole project.

Thank you for not using clickbait titles and sharing priceless knowledge instead.

Another standard we need to be widely adopted is some kind of royalty-free audio networking. I wish AVB - a 12 year old standard- was more widely adopted. ADAT I/O is not enough

Very good Wytse! Thanks for always pushing the audio world to get better and better!

The thing with CLAP is that most of the advantages it has over VST and AU are developer focused, not user focused. There are some advantages over VST for users as well, like better performance for modern CPUs, better plugin metadata and organization, built-in per-note automation and modulation (MIDI 2.0 and "MPE on steroids"), and ofc great cross-platform support. We will probably see more user centric improvements in the future, but until CLAP has become the de-facto plugin standard, most new improvements will probably be on the developer side just to make sure that most developers follow suit.

Just in case you don't know: Digital Performer will PRE-RENDER plugins--which is pretty much what you're talking about--if you want it to.

Great insight! Thx!

Your thoughts on offline processing at 5:15 remind me of the experimental DAW Blockhead where you see the waveform of your track audio get altered in real-time as you apply FX and tweak parameters.

And what about ARA? You know, the thing Melodyne uses - seems to have offline access to multiple tracks :)

Some brilliant ideas you have there! Hope to see them one day become a reality

Interesting video. Thanks Witse.

I worked in standardization for years, be in a completely different field. I recognize everything you say, and you are right about everything. I often felt like an alien, because you hardly ever can explain why everything can't just be connected to each other and just exchange data. Standardization is in a way always a compromise and comes with (experienced) limitations. Getting all parties agreeing takes a lot of time and is more a matter of culture, egos and diplomacy. Als everyone has to make explicit what they want and do and that turns out to very difficult. There also have to be shared goals, and you need to trust each other. By now it may be clear we are detached from the technical level (and lost the developers and programmers). The key may be the concept of plugin that should work with every DAW. The more parties will hae to work with it, the less you can do. Interesting may be de facto standards like USB. Or how MIDI came about. In a way Germans seem to be good in standardization. Let's figure out why that is. From the start I thought CLAP was a stupid name which wouldn't facilitate it's acceptance.

When you said, that VST3 was finalized in 2008 and considering, that most plugins from 5 years ago where only supporting VST2, it tells you much about the development and adoption rate issues for any standard. Specifically, an open standard always begs the questions: who is going to invest their time in the development and support? What happens, when that development and support just stops? What I wished to be part of the standards next to inter-plugin and extended plugin-DAW communication was multicore CPU usage and (with hopefully enough love of GPU audio) GPU usage. Can you do the following: Have an interview with a developer on how VST3 vs. CLAP comparison looks like and why there are so little implementations of CLAP so far (like none of the major brands in the instrument and FX business aside from u-he). Because just asking for a new, better standard doesn't say much about if the current standards already have the features and are just not implementing it as they could and should. Also, a new standard means it needs to be adopted, specifically for business economic reasons. If it takes too much time or effort, it will not be adopted. If there's already a solution, why not use that and pay a few bucks? As of now, CLAP is mostly supported by rather small, unknown devs after two years. That doesn't sound good if you want a standard to take off.

The 'proxy render' option would be cool. It is something that's already very common in the Video Editing softwares. Something that would eliminate the need to do freezing or rendering the tracks etc. Basically it would 'pre-render' the track in real time, and it would keep the effects on the tracks, turning off the plugins (until they are tweaked again). It would make working on big projects easier and faster, even on worse computers.

There are open protocols developed by the Linux community called LADSPA and LV2. Any standard that is not open is doomed to fail to be consumer-friendly as we have seen it so many times across the industry. I just wish LV2 had native support for DAWs running in MacOS

Dont worry, I'm sure we'll get CLAP support in Ableton by 2052 maybe

I have been thinking this for a while too, so many possibilities open up when vsts have "offline access". When I started I was very surprised you couldn't just check the LUFS whenever. Fortunately music software is an area where passionate people occasionally create things that don't necessarily make sense from a business standpoint (Reaper's Winrar-like selling strategy). So I do have some hope for the future. Awesome video, totally agree.

I work in a mix mode doing both in DAW and offline. Would be nice to have a nice interface for this. Take a lots of work when adjusting automation

I may be wrong but I’m pretty sure FL Studio added CLAP recently.

I think you are right and things will close up. But it will be a time to look forward to. Because it can also be offline and can REALLY have the best out of both worlds, Analog&digital: Look at UDO Super Gemini. FPGAs are the future of audio devices. Hardware feel because of like 8x real time oversampling with close to no latency- and it will sound great. But it will need a hardware device to load „logic“ onto. FPGAs can load software structures as Hardware (programmed with HDL Hardware Description Languages) and run the logic at samplerates like 25 Mhz instead of 192 Khz. And they run processes in parallel and can even be used in parallel to normal MCU based environments

I think, to be able to do what you want, is to rethink how tracks in DAWs work. It is not a matter of a plugin-format. But a new way of thinking on tracks. I would also like to see the use of DATA streams integrated to the tracks, for controller functions, as well a ways of encoding time-codes, to do "look ahead", even with analog outboard gear, as it could buffer time-coded streams and sync them, to allow for time aligned parallel processing. And also the ability to stream tracks through the network, to use other computers to offload, with the data streams also enabling control of remote plugins, in an integrated way. I was thinking about writing Behringer about this when they were asking about what people would like in their new DAW. Not that I think they are necessarily the ones to pull it off well. But at least they were starting off fresh, and has some money behind them. The other thing I really would like to see is something like dawproject, being adopted by the whole industry. And DAW developers to the extent possible, releasing their custom plugins as standard format plugins, that I would be fine with requiring the ownership of their DAW to get access to. I mean, if it is possible to run Ableton Devices, as plugins in any other DAW, one might be more inclined to use them, if one is on a a multi-dawset-up. Whereas one might otherwise find 3rd party plugins, just to know that one is able to run them in any daw, at any point. That would of course also be required for a true multi daw format. I think the only ones really to potentially lose out with a proper multi-daw format is pro-tools, unless they are the ones inventing it, and getting some license fees for it. For everyone else, if they lose a customer for a couple of generations of the DAW, at least when they add the features people were looking for, the reason they switched, at least then it would be easy to switch back. And many more would be willing to be multi-daw users, expanding the over all market. And if I could dream, imagine being able to send a project, that comes with project specific licenses to plugins. So anyone opening it, could download the plugins, and if they don't have a license, they can use the project specific one. Yes, there would have to be some rules to how much one can change the project, to not be able to use it as a way of sharing plugins. But it would be fantastic for so many use cases. For collaboration. Or for the audio engineer that is tasked with mixing or mastering, being able to actually open up the plugins used and correct things as the plugin level, instead for having to ask the producer to fix and re-send. I would also see other smart ways of dealing with licenses, so that one is not locked out, when forgetting to either bring the USB key, or to release a license... There could be ways of solving it, by having to later connect to the same network as the computer that had the license in the firt place, not to be locked out. Or basically, getting rid of most offline licenses, but not requiring authorization ever time it is used, but at some interval, with limited possibilities of pausing autohorization.

I think image-line hasa clap support for fl studio in the current beta.

Oversampling one is spot on...

Reason tried to do a closed ecosystem, look at them now. UAD recently opened their plugins for use outside their hardware.

I think one way companies could accomplish this is by all agreeing to compile their programs into a universal file format and get daws to accept this new format and then program it’s language to be in communication with neighboring plugins of the same type, It’s a long shot because considering the business model of the corporations hiring the programmers, they would never fully give up their own formatsbecause they wouldn’t trust globalizing their plugins unless it provides them a benefit.

All we.need is more cowbell

I suspect the 'look ahead' wouldn't work in the way you imagine. Although computers are very powerful now, rendering audio offline actually takes a lot of resources. You can see that with the new mastering feature in Logic: when you activate it, the DAW has to pause for quite a long time while it chews through the audio. And the problem is that every time you change something in your DAW, move a single fader, turn a single dial or or add in a single new plugin, your 'look ahead' algorithm would need to be updated. Could end up being a huge extra load on the CPU. So I suspect just doing this in the background all the time may not be that feasible?

NEVER say never!

I really love that you referenced the xkcd standards comic :D

What you suggest is also marred with problems. Plugins looking ahead to a render ends up in a vortex in that the render needs to bounce down other plugins other than itself to get a picture, then others do the same, but also plugins are material sensitive to what they get in a moment in time and in a sequence, it basically can’t happen, everything needs to be sequentially interpreted in the moment like it is currently. Oversampling, to some degree the solution is to use a mega high sample rate in the project instead, but not everything needs to oversample, so there’s wastage. What CLAP brings is mainly freedom from Steinberg and a slightly better interface, but the DSP approach has to be the same sequentially. The biggest problem with plugins is CPU and what they have to do in short blocks without hogging too much. Thats why emulations try to capture the best perceived characteristics of what they are based on but can never manage it all, they skim because that’s all they can do.

The Mooger Fooger Plugins also communicate with each other + the Moog Marina Plugin

The CLAP format looks interesting also from a licensing standpoint. It’s got a license that is friendly to both opensource as well as commercial development, unlike VST which has this awkward dual license model. Steinberg where once innovators that brought a widely adopted plugin format but they have been resting on their laurels. When they created VST, they also created ASIO to go along with it. It was all fine and dendy in the 2000’s but users needs have changed since then. Development on ASIO is completely stagnant. Not much has changed to it since 1999! Meanwhile the landscape of audio hardware has changed, USB interfaces came along and since the last couple of years also synths, grooveboxes, drum machines etc with USB audio have appeared. But there is one thing holding them back, and that’s the rather old fashioned single-device architecture of ASIO. I feel that it is stuck in the 90’s and that a new driver standard is also required that allows multiple devices to be used under one driver in a reliable way, just like how you can plug multiple things into an analogue mixer and it just works. I’d like to see along side the CLAP plugin format a new, and open audio driver model. With open standards innovation can happen that can result in new solutions that can solve today’s needs as well as anticipate future needs. Wouldn’t it be cool if there could also be a specialised, audio oriented OS that could run any DAW and Plugin that doesn’t require a locked down computer?

Could you make a video about the difference between a clipper and a limiter and how to use them in conjuction with eachother? Thank you in advance

2:05 it is caller propellerhead reason (before it allowed VST user remained decades without vst...never made so many track that at that time)

very interesting, so is this time in music and the world. Thanks!

I too have often thought about look ahead processing. I never really understood the limitations of VST or other formats, I always thought that the reason it was mot available was the stubbornness of old school industry types Your video really provoked some thought for me. Thank you

I've been doing so much mixing and mastering in last year mostly EDM. I have learned so much, and one of things I've learned VST's cannot replace Hardware, and you will always have limitations so you have to adapt your knowledge to VST. Many engineers and "big" producers have sold their equipment, yet I hear phasing in their tracks, and things that would never appear if they used proper hardware. I hope one day there will be good enough analog emulation that can provide same results as it did 20 years ago, even tho tracks sounded "muddier" they were way cleaner and you could hear each element of track, something that digital VST's cannot provide unfortunate.

Reaper is the DAW for rebels

Ideas are possible. Here's one: a plugin wrapper that lets you tab transients and make break points so you can turn regions' volumes up and down. Then the plugin wrapper lets you use another plugin like a limiter or compressor. After the compressor, the plugin automatically reverses all your volume changes, restoring the macro dynamics. The point would be to let you prep your signal to feed consistently into a compressor or limiter so the processing works more like a micogroove shaper rather than a leveler, but then gives you your dynamics back on the back end. It could do busy work for you like analyze the new waveform for the nearest transient break points (they will shift around with processing). Or even automate the frontend of the process. Its kind of like conjugation in abstract mathematics g^-1*h*g. Curious about Sound Radix Pi. It uses info from all the tracks. Is the distinction that it can't have a global view of the session data?

I’ve been playing around with Suno lately and while the bitrate sounds pretty degraded, the overall mix quality is better than what I can produce with tons of paid plugins and a few years of learning what I can. It’s crazy to me that it can create a rounded professional style mix in seconds but side chaining is the most we can do to have tracks communicate with each other.

Thank you for advertising your ideas. And by the way, before watching this video, I had no idea about the problem of sampling filters of several plugins. I think that this can be solved inside DAW, with a special built-in container plugin that internally upsamples and has a downsampling filter on the output. And the idea with proxy rendering is also great. In fact, everything for it is already inside the software. Since, for example, I do something similar manually. I'm just doing a bounce at the bottom of the original track. Then I turn it off. This allows me to always go back a step and change something. It also saves a huge amount of resources. All you need is to simply automate this process.

Im okay with my fab filter suite 😎

i've had the same thought about wishing plugins could do lookahead/offline processing

When I hit play on my daw, the audio engine starts calculating all the tracks outputs and bus outputs and mix outputs. I press stop, and then play again, and the audio engine does exactly the same calculations all over again, and gets the same result again. Why not cache each tracks render as a hi quality wav file in RAM (we all have huge amounts of RAM these days), and then simply play the sample back next time. If the track changes or a plug in adjusted, then next playback calculate it again. The load on the processor would plummet. Saving the project could give the option to save all the renders out as well if needed. Then a plugin could easily 'look head' as the wave would be sitting there ready to inspect.

Coming from a web dev background and being a hobbyist musician, would be great to see something cross-platform so no matter which platform you're using whether PC or Mac it should work. I'm curious whether the DSP for Mac and PC are somehow different which is why people develop specific to that platform or just lazy to code it. Would love to learn more about how all this works since I'm curious what are the limitations

How about the ability to embed a frozen state of plugins into the host save? Meaning that even if you no longer had the plugins installed / authorized, you would be able at any time in the future, open a project on any computer with any newer OS/ Hardware , edit the audio and MIDI, remove frozen or add new plugins, render the audio and so on. In effect being able to save a project archive, complete with plugin state and without having to think about rendering all the audio tracks..?

For the first point there is ARA (Audio Random Access) which was developed by the company of Melodyne. Else a plugin like Melodyne would not work.

Apples and oranges. And always will be. The direction McDsp took with their APB line is the first break-through since ProTools TDM in 1995 promised a true in-the-box studio when combined with great analog summing boxes.

regarding the track fx communication like oversampling in reaper you can set a global oversampling for the channel, if you disable oversampling within the plugins it should do what you're saying no?

Everyone has a different opinion but I personally enjoy the open architecture of the PC instead of the closed Apple system however each succeeded due to this fact. The original Pro Tools system required 30 thousand pounds or dollars of hardware before you could use a single plugin eliminating most musicians from the ProTools recording business. There's definitely a rightful debate for both sides and always has been. Good video!

Hasn't Reaper solved the up/downsampling thing by making it possible to upsample the signal within the whole plugin chain?

What about Reaper's integrated oversampling functionalities? No communication between plugins, but we can set manually chain oversampling if I remember correctly.

I like the idea, I think some plugins don't upsample if you feed it 192 or something like that, I don't know because I only work in 48. Anyway I hope something cool comes out of this!

you are my favorite plugin guy

I wonder if an upsampling/downsampling container vst is possible. Where you would load all the vst's into a container and let the container do the upsampling.

Amazing content again. I'm not wiser now but I'm way more informed. This is important!

Korg gadget instruments are now useble outside of gadget as well

I know that CLAP brings polyphonic modulations and things like that but I’m not sure it brings any improvements regarding how audio streams are processed.

I've often wondered why oversampling isn't managed at the DAW level. It seems like a logical step towards achieving consistency in audio processing across different plugins. Having oversampling centralized within the DAW could streamline the workflow and potentially enhance audio quality by solving the issue of multiple points of sample rate conversion within a processing chain.

Final Cut Pro already does automatic rendering whenever you stop manipulating stuff for a second. I can imagine a DAW making « freezing » tracks more automated. However, I think it would struggle with sidechaining tracks with each other.

I come from the IT environment and am a big supporter of the OpenSource community, so I also contribute to it. I mean, especially in IT, the whole world is based on OpenSource projects. But I've also noticed that this open source and standardization community in the audio sector has fallen asleep a bit. I really like your idea and this approach to change something about it

I’ve noticed my fabfilter plug ins are the only clap plug ins I have It came up after updating them I wondered what that was 😅

once computer speed and space increase maybe plugins can be packed with more algorithms to further sound like hardware. ?

It would be awesome if white sea studio reviewed some 3rd Party Libraries for Acustica Nebula!

What is it that the VST standard can't accommodate? Oh you answered that. I should have been more patient 😂

I constantly think of how Abobe has Dynamic Link between Premiere and After Effects. Alter something in Premiere, immediately gets altered in After Effects. Certainly it’s possible, but like you said, not with current plugin standards.

....CLAP?

Oh wow, very interesting thoughts

7:11 a DAW should be able to do upsampling/downsampling in a chain though. Since it already knows which plugin is doing what. Much easier solution.

Creating a new standard will just just be that, another standard that developers will need to write to in addition to the existing standards. VST2.4 is still widely used and it is 30 years old now. Currently the cost of development is so high that this doesn't make economic sense to try and create another new, competing standard. It is simply money. Something as big as a new effects SDK requires a lot of money and time not only in creating the standard, but in getting companies to support that standard.

If you had access to offline processing and inter-plugin communication, what kind of plugin would you want to make?

The new plugin format they are developing right now is called CLAP

Plugins Are Great, And it’s More about learning when and When not to use them once you actually Learn How they Work and Not To Buy Duplicates which Just have A Different U i.

Especially compressor plugins could be much closer to hardware and have more realistic attack behaviour, if they didn't have to process in real time. I've been thinking about this quite some time and it pains me that there isn't really a solution out there (except ARA... eh). It really seems like offline processing would be the missing link/next step for realistic hardware emulations.

yeah, if plugins to talk to each other i feel like it would be really good especially for atmos mixing

This up sampling down sampling things is interesting. I always felt my mix downs where I’ve just used the native DAW plugs ins sounded bigger and more powerful, I wonder if this is because of all the nonsense happening with a multitude of 3rd party plugs.

Yes. I have been thinking about this. There could be a more integrated approach. Maybe DAWs could be split up into 2 areas. 1. GUI and 2. processing. For example, I like using Reaper because of the UI. But, what if a company like SSL could build the back end processing for Reaper? So that you could have the customizable workflow UI and the signal processing of an SSL console. So DAW manufacturers could sell a version of their DAW that is just the GUI. So that software processing manufacturers could sell the back end processor. Maybe companies like SSL, API, NEVE, etc. I like the CLA mixhub. However, It doesn't have sends in the buckets. I understand why. But it is clunky. Or, I like the way Mixbus 32c has all of the EQ, Dynamics, Bussing built in. Or Studio One has “Mix FX.” If reaper had that feature I would probably use that. I find that just a touch of the crosstalk and drive does some really cool things. I like how it deals with things on a lower level - like crosstalk. I don’t know how this could be accomplished. But, I do know that the mixing in the box experience could be much better than it is.

As much as I love that I can make music with DAWs. Some major parts of how they work have not developed in the least. The stuff that you're talking about definitely, but also channel summing wich could have a global plugin controlling how the summing is being calculated. Imitating an analog console sound would be a breeze with that. Plus it would open up other global level sonic processing. Theres been suggestions for a different sampling format. One wich could eliminate aliasing. Its still only an idea but I think it holds a lot of promise.

May I please correct one of the statement. Clap format is created and developed by Bitwig and U-He. Not only the latter one. Thank you. PS: Clap. Clever Audio Plugin API. Also Gonorrhea. Not sure I dig this name much… 🤭

when you move into live sound some close systems are already there, like soundvision. you should get a programmer ally and start making a daw from scratch the way it should be with proper features as stock 😁

Hardware is more expensive . Or is it ? Maybe time to reconsider hardware ?

clap clap clap, well said

I feel the same. There's so much possible and I feel constantly limited like we live in some stone age era, wasting computer resources and amazing workflow possibilities. Actually some solution would be DAWs implementing such internal plugins with offline capabilities and inter-talking themselves though, if the plugin companies won't be able to create some cool common standards

I don't understand the argument. A new, open standard would not guarantee that it supports all the nice features you're asking for. Portability of source code is essentially already there for quite some time through cross-platform frameworks such as JUCE. It's not optimal, but apparently a working solution.

I've also wondered about why there has not been any offline processing style plugins. Now I know a bit more about why it hasn't been done.

Coffee ad cookies ... absolutely brilliant analysis of our current state of bit candy ... you rock