[Haskell-cafe] wondering on Haskell kickstarter project: VST plugin / mobile synthesizer

[Evan Laforge]

On Mon, Dec 21, 2015 at 2:03 AM, Anton Kholomiov
<anton.kholomiov at gmail.com> wrote:
> @Evan
>
> Thanks for detailed and sincere reply. I can see your points.
> That VST/mobiles market is very crowded business and it's really
> hard to compete with super giants especially when you
> don't provide any significant change or creative opportunity
> in the app.
>
> So I think this project is going to be more a prove of the
> concept and example that it's really possible to create
> real-time audio with Haskell and wrap it in non-Haskell UI.
> Hope I would manage to do the mobile version of it too some days.
> I would really like to add non equal temeprament scales, I feel
> that it can make a difference to the mood of the music.

This is a basic requirement for me, because most of the music I'm
interested in doesn't use equal temperament.  MIDI makes it really
inconvenient though, which is one reason why if I did my own
synthesizer it wouldn't use MIDI.  Supporting multiple MIDI channels
is about as convenient as MIDI can get, but very few VSTs do even
that.  But computers are fast and loading say 8 copies of the same VST
actually works, though it's aesthetically unappealing.  Actually, it
only works for cheap synthesis like FM... so maybe it doesn't work
that great.  As far as I can tell, if you're using MIDI there is no
satisfactory answer.

Some synthesizers can load scala files too, which more complicated and
not as flexible, but it can sidestep the multiple VST hassle.  Of
course it's static so it only works for instruments like percussion
that are not pitch continuous.

Even though there is a crowded market of giants, from my perspective
they all have fundamental holes in their functionality, but it works
for them because they aim at a market not interested in those things.

> Maybe it's better to look in the direction of media installations.
> The Python has many cool libraries for non-trivial user interaction
> like computer vision. It would be great to stream this data
> in Haskell generated synthesizer. Maybe something can be created in this
> area.

Indeed, it does sound interesting.  Ultimately it's whatever is
interesting enough to you that doing it would be its own reward.  I
did a little experimentation with interactive musical toys in python
back in college, and it was kind of neat... though ultimately didn't
pursue it any further.  With a powerful language like haskell and
access down to the low level synthesis you could express more
complicated kinds of control.