Csound within Mixbus

[cauldron]

Linux can not compete with Windows/OSX commercial plugins solutions. The software companies generally do not support linux (apart Harrison and a few others).
There are two ways to use the Windows/OSX plugins with linux:
1) hosting Windows plugins within Reaper with wineasio
2) use Windows/OSX as a network audio black-box with Jack Audio network infrastructure
.

Both ways have their advantages and disadvantages. In any case the complexity of the system increases significantly (especially in the case 1)). Minimizing the complexity is the best choice (...just use Windows or OSX...).

Harrison mixbus offers many integrated cross-platform plugins tools. Also Xrange and Xtools greatly expand the DSP scene with a hardware only Harrison implementation.

It would be interesting to know some development guidelines within mixbus(/ardour) to develop custom plugins using csound audio engine.

Also Cabbage is a software for prototyping and developing audio plugins with the csound audio synthesis language. It provides users with a powerful toolkit for the development of cross-platform audio software. Cabbage is a framework for audio software development. Using Csound as its audio engine, Cabbage users can target Windows, OSX, Linux and Android with a single piece of source code. As well as developing standalone software, users can also export their instruments as fully functional audio plugins, thus paving the way for full integration of Csound into any number of digital audio workstations...

Csound: http://csound.github.io/
Cabbage: http://cabbageaudio.com/

[superb]

I would love to run MB in Linux. Both options sound viable, but I'm curios to know if I would be ahead in terms of performance by running Linux - reaper with waves plugins wineasiio into MB vs only running MB with waves plugins in Windows 10...


Sent from my iPhone using Tapatalk

[cauldron]

I would love to run MB in Linux. Both options sound viable, but I'm curios to know if I would be ahead in terms of performance by running Linux - reaper with waves plugins wineasiio into MB vs only running MB with waves plugins in Windows 10...


Sent from my iPhone using Tapatalk

If you use linux then you can reduce, customize and optimize everything as you like until you get to an embedded low latency system. The big advantage is the reliability of minimalism.

OSX is more optimized than Windows. If your Windows system is not well optimized then it could happen this on linux/reaper platform:
- The plugin will not run faster (especially if there are many graphical interface effects)
- There is not a guarantee of compatibility (especially if there are hardware protections as Steinberg eLicenser)
- There is no support
.

If the plugin should work fine then the only advantage is the optimized linux environment and of course the ability to use the plugin with linux. For example, I use Voxengo 32-bit plugins with AV Linux 2016 64-bit and Reaper 64-bit.

The fact that it works does not mean that it is a thing well done!

[superb]

Thanks for the reply! I suppose I could test it out sometime to see if there are any gains to be had...


Sent from my iPhone using Tapatalk

[BigstevE]

Linux can not compete with Windows/OSX commercial plugins solutions.

Can't say I agree with that assertion. Though I would say, with respect to the multitudes of hardware/software configurations of Linux users, most Linux plugin providers are generally at a financial and compatibility disadvantage when compared to the bigger fish in the pond. Thing is, it's a new pond out there and most of those fish are emperors with no clothes...

Much of the sound processing software, be it in plugin form, or an audio synthesis language available for Linux is no less effective in creative use, or sound quality that what's available for Windows or OSX.

It would be interesting to know some development guidelines within mixbus(/ardour) to develop custom plugins using csound audio engine.

You might want to look here. It seems this is where Ardour(/Mixbus) is headed. If Csound can be called with Lua functions then who knows what's possible.. Though Csound is a wondrous powerful language, there are also a number of other languages that you can use to build and prototype plugins with these days. Faust immediately comes to mind. Csound, I believe is also as much a MIDI and scoring-compositional language (using orchestra and score files) whereas Faust is more of a purely DSP based language for sound processing in the 'acoustical' sense, if that makes any. There is also SuperCollider, ChucK and PureData but I'm not so sure how relevant they are for creating stand-alone plugins in DAW compatible formats. I guess it also depends on kind of plugins you want to create with it, and how you want it to interact with Mixbus.

Best,
Steven

[x42]

Linux can not compete with Windows/OSX commercial plugins solutions. The software companies generally do not support linux (apart Harrison and a few others).

Well, ask your plugin vendor to provide a Linux version...

It would be interesting to know some development guidelines within mixbus(/ardour) to develop custom plugins using csound audio engine.

There's still csladspa, it's a bit dated and audio only but available on most linux-distros out of the box.

If you're thinking of tight integration (not just generating VSTs), it'll definitely have to happen on the Ardour side first (Mixbus codebase is 99% identical to https://github.com/ardour/ardour) and realistically it'd be Csound only. But even pulling that off and keeping it in sync with upstream would be a lot of maintenance work. Personally I think it would be very cool (I'm a big Csound fan), yet overkill.

IMHO the way forward is to just use exiting AU and VST export features of Cabbage.

Can you be a bit more specific what you mean with development guidelines?

[cauldron]

Thanks BigstevE and x42 for the advice.

Ardour does not come with any built-in signal processors of its own (other than volume faders) then Mixbus complements the GUI and "basic" DSP engine very effectively (and also it offers several "Mixbus Plugins").
Your advice (csladspa & AU/VST) seems to me the best thing.

I am thinking of the guidelines to scale with the computing resources (a quad-cores Intel processor has about 100 GigaFLOPS). Think of 300 tracks of a pipe organ that sounds! Many of these tracks may be the result of an additive synthesis, physical modelling synthesis, ..., and many DSP post-processing (many TeraFLOPS power)! CUDA, OpenCL, Xeon Phi, NUMA architecture, Distributed Computing, ...

But Ardour is DSP-agnostic so they are useless talk.
But at least in this case Linux has great potential compared to Windows/OSX. The Linux kernel supports NUMA vSMP architecture for transparent power scalability: ScaleMP http://www.scalemp.com/

[Frank]

Hi

while it is always useful to have expertise on guidelines to scale with the computing resources, if one considers supercomputers, i am afraid we will end up with a Ferrari engine for a remote controlled model car.

And remember we are using real time systems. And SMP is not an ideal model for RTS.

But let me not hold you back !

regards