Plugin Development Uni Project ! Please Help

[jkjkjk123]

Hi guys,

I am currently at uni studying a degree in audio production and would like to create a VST as my final project. I also host a worldwide weekly radio show. As i am constantly on the go playing shows (DJ) I do not have to time to mix my weekly radio show live hence I use Logic and a few plugins to mix my show. It usually takes me about 3 hours or so to finish the whole mix as i have to adjust each parameter on the plugins I use for each track. I would love to make this a more simpler process. I would like to design and create a VST that mixes 2 tracks seamlessly while having the control over how long it takes and from what points of the tracks it mixes. I currently do this on my DAW using the Volume Parameter and all 3 Bands on a 3 Band EQ plugin including a limiter to prevent any clippings that may occur. I have to manually adjust each parameter for each track which takes a long time when making a 1 hour mix. I would love to get some advice on how I can create my VST. Im a beginner in coding but I am willing to learn everything and anything !

Best Regards,
Jay

[Mayae]

Have you considered using something like Ableton and Max4Live? It seems well suited for this purpose - you can automate the DAW and other plugins through algorithms / midi / events in a visual programming environment.

Since you're using Logic, I doubt you'll be utilizing a VST (but AU instead). If you insist on programming it your self, take a look in the top of this forum where there is a lot of information. Also, audio plugins don't have any knowledge of 'tracks' or 'samples', they are only fed a stream of samples from N channels, and generates some output. An audio plugin solution will not have an integrated workflow in your DAW, unless you use something like Ableton as previously mentioned.

[paterpeter]

Im a beginner in coding but I am willing to learn everything and anything !

What time frame are we talking about? If you've never written a non-trivial program before, I'd recommend not to do this as your final project, definitely not if we are talking about a native plugin written in C++. You will likely have a very hard time coming up with anything useful and stable in the typical 6 months time frame. Don't forget you'll need to write a thesis, too, which takes quite some time.

If you want to get into plugin programming, try it in your spare time and see how it goes. But doing a final project that relies on a skillset you don't have yet will put unnecessary pressure on you.

[paterpeter]

(contd.)

If you want to use your basic idea in a thesis, you should probably do it like this: define a problem ("mixing tracks is a laborious, repetitive task"), propose a solution ("automatic track mixing") and do the research to prove how to do it (or that it isn't feasible, which is an equally valid outcome). You'd probably define the criteria and parameters for a "good" mix, determine what the problems are when mixing etc. You could then use audio examples and a "manual automatic" example to prove your findings. You'd probably leave the implementation of your approach as a usable product as future work.

I'm that's plenty of stuff to do in your typical final project time frame and it sounds like a good contribution to the scientific community.

BTW, this sounds more like an offline tool, not like something that should be a VST. As Mayae mentions, plugins are for audio streams.

[Zaphod (giancarlo)]

A plugin maybe not, but you have time for creating an external application which simply automate some kind of off line processing. In java or python it is quite straightforward, in c++ less but feasible.

[Ap0C552]

I am currently at uni studying a degree in audio production and would like to create a VST as my final project.

Unless you have solid prior programming knowledge, this sounds like a bad idea for a final project. So much work, I would not want to risk a grade on it.

[jkjkjk123]

Hey guys, thanks for the advice ! I don't think it will be likely after the feedback but I will definitely look at it in my free time. Thank you as well paterpeter for the idea on the thesis as well !! I may follow that idea !

[jkjkjk123]

(contd.)

If you want to use your basic idea in a thesis, you should probably do it like this: define a problem ("mixing tracks is a laborious, repetitive task"), propose a solution ("automatic track mixing") and do the research to prove how to do it (or that it isn't feasible, which is an equally valid outcome). You'd probably define the criteria and parameters for a "good" mix, determine what the problems are when mixing etc. You could then use audio examples and a "manual automatic" example to prove your findings. You'd probably leave the implementation of your approach as a usable product as future work.

I'm that's plenty of stuff to do in your typical final project time frame and it sounds like a good contribution to the scientific community.

BTW, this sounds more like an offline tool, not like something that should be a VST. As Mayae mentions, plugins are for audio streams.

Thank you !

[PurpleSunray]

(contd.)

If you want to use your basic idea in a thesis, you should probably do it like this: define a problem ("mixing tracks is a laborious, repetitive task"), propose a solution ("automatic track mixing") and do the research to prove how to do it (or that it isn't feasible, which is an equally valid outcome). You'd probably define the criteria and parameters for a "good" mix, determine what the problems are when mixing etc. You could then use audio examples and a "manual automatic" example to prove your findings. You'd probably leave the implementation of your approach as a usable product as future work.

I'm that's plenty of stuff to do in your typical final project time frame and it sounds like a good contribution to the scientific community.

BTW, this sounds more like an offline tool, not like something that should be a VST. As Mayae mentions, plugins are for audio streams.

Thank you !

Also often seen is matlab code.
You don't really want to deliver a fully functional, super-stable, high-performance, nice UI, high quality .... VST plugin as a result of this project, but the project is about to show how to get there.
So as already described, describe the problem, your research on it and the solution you propose.
If you have any fancy algorithms you want to prove or show, simply do it in matlab.
Like.. problem is that I need to touch the knobs on the mixer while mixing... my research shows that is such a dump task that it could be actually automated... this is my smart-mixing-EQ idea.. and here is the implementation on mathlab to prove that it works.
This is actually pretty close to how 'real' VST plugins are made. At that point you hand over your matlab code to a VST developer and tell him "do this is C++, with a nice UI on top" .. all the thinking work has already been done by you and is verified - he only needs to translate from mathlab to C++ (and this is outside of you project .. )

[Obsolete236871]

On the JUCE homepage there is a step-by-step tutorial that looks quite interesting:

http://www.juce.com/doc/tutorial_create ... sic_plugin

[jkjkjk123]

(contd.)

If you want to use your basic idea in a thesis, you should probably do it like this: define a problem ("mixing tracks is a laborious, repetitive task"), propose a solution ("automatic track mixing") and do the research to prove how to do it (or that it isn't feasible, which is an equally valid outcome). You'd probably define the criteria and parameters for a "good" mix, determine what the problems are when mixing etc. You could then use audio examples and a "manual automatic" example to prove your findings. You'd probably leave the implementation of your approach as a usable product as future work.

I'm that's plenty of stuff to do in your typical final project time frame and it sounds like a good contribution to the scientific community.

BTW, this sounds more like an offline tool, not like something that should be a VST. As Mayae mentions, plugins are for audio streams.

Thank you !

Also often seen is matlab code.
You don't really want to deliver a fully functional, super-stable, high-performance, nice UI, high quality .... VST plugin as a result of this project, but the project is about to show how to get there.
So as already described, describe the problem, your research on it and the solution you propose.
If you have any fancy algorithms you want to prove or show, simply do it in matlab.
Like.. problem is that I need to touch the knobs on the mixer while mixing... my research shows that is such a dump task that it could be actually automated... this is my smart-mixing-EQ idea.. and here is the implementation on mathlab to prove that it works.
This is actually pretty close to how 'real' VST plugins are made. At that point you hand over your matlab code to a VST developer and tell him "do this is C++, with a nice UI on top" .. all the thinking work has already been done by you and is verified - he only needs to translate from mathlab to C++ (and this is outside of you project .. )

GENIUS !!!! I have never used mathlab, ill guess ill start working on it !!! Thanks !!!!

[jkjkjk123]

(contd.)

If you want to use your basic idea in a thesis, you should probably do it like this: define a problem ("mixing tracks is a laborious, repetitive task"), propose a solution ("automatic track mixing") and do the research to prove how to do it (or that it isn't feasible, which is an equally valid outcome). You'd probably define the criteria and parameters for a "good" mix, determine what the problems are when mixing etc. You could then use audio examples and a "manual automatic" example to prove your findings. You'd probably leave the implementation of your approach as a usable product as future work.

I'm that's plenty of stuff to do in your typical final project time frame and it sounds like a good contribution to the scientific community.

BTW, this sounds more like an offline tool, not like something that should be a VST. As Mayae mentions, plugins are for audio streams.

Thank you !

Also often seen is matlab code.
You don't really want to deliver a fully functional, super-stable, high-performance, nice UI, high quality .... VST plugin as a result of this project, but the project is about to show how to get there.
So as already described, describe the problem, your research on it and the solution you propose.
If you have any fancy algorithms you want to prove or show, simply do it in matlab.
Like.. problem is that I need to touch the knobs on the mixer while mixing... my research shows that is such a dump task that it could be actually automated... this is my smart-mixing-EQ idea.. and here is the implementation on mathlab to prove that it works.
This is actually pretty close to how 'real' VST plugins are made. At that point you hand over your matlab code to a VST developer and tell him "do this is C++, with a nice UI on top" .. all the thinking work has already been done by you and is verified - he only needs to translate from mathlab to C++ (and this is outside of you project .. )

GENIUS !!!! I have never used mathlab, ill guess ill start working on it !!! Thanks !!!!