RS-MET
RS-MET copied to clipboard
RobinSchmidt
I'm looking for customers for your library
Robin, I am actively looking to develop plugins for people with your library. It is very important that you take my feature requests. I could perhaps have them pay me hourly while at the same time giving you money for your library.
So, be ready. It's going to happen.
As a matter of fact, it would help to have some idea of pricing for your library right now... like... now now.
as said, i'm aiming at a pricing corridor between 500 and 2000 euros per product, depending on size and scope of the product/company, the role of my library within the product, etc.
What about supporting developers who are just starting out? That pricing scheme makes your library uncompetitive. That pricing makes sense for developers/companies who know what they are doing, but think about my situation. I can't work for a big company. What do I do? My first handful of clients won't have that kind of money.
500 to 2000 euros per product, your framework needs to be ready to make an amazing product from scratch and quickly. It's not there yet due to lack of easy custom GUI and preset browser.
i actually never planned to license the framework (i don't really know if it's any good, to be honest) - i'm talking more about the dsp backend and i see the competition more among the likes of zplane - and they charge more and don't even provide source code (just headers and compiled libraries). but as said, it really depends on the company and what kind of money they have or make with the product. these prices are more for established companies. maybe one can also think of some kind of revenue sharing model
although zplane's libraries are different in that they are focused on a single dsp task and do that really well whereas my dsp library is more broad and less deep, i guess
wait... im confused.
The gui stuff was not meant to be part of your library? or part of the deal? that's just for your personal use? and I'm riding off of that? And... so........ I shouldn't be bugging you about custom UI?
but I thought you wanted a one-stop-shop from scratch to finished product? no?
hmm...well, i see my role more in dsp algorithm department - not really so much in the making-a-polished-product with multi-platform support, installers, licensing system, etc. i'm much more interested in focusing on the dsp side. this is, where i know a thing or two. i'm not very good at the other stuff
But you're almost there with multiplatform support, licensing system, gui.
What if I took on the role of designing your custom gui stuff? I need this for my own products, i'd be doing it anyway I guess, from what it sounds like.
If you don't make your library a one-stop-shop for a product from scratch, then I want to!
"Elan's Product RS-MET Extension Libray" only 10,000 eruos per product.
Edit: I don't want a separate library thought, and I don't need credit for doing it for your library.
wait... im confused. The gui stuff was not meant to be part of your library? or part of the deal?
i don't mind, if someone wants to use that, too - but i'm not sure, if my framework is any valuable at all. i'm quite happy with my core dsp-stuff, though. a lot of math problem solving work went into it
The gui stuff was not meant to be part of your library? or part of the deal? that's just for your personal use?
yes. sort of.
Oh, so I know how to override the paint method. How do I add custom filmstrip? Is that pretty easy? If this is all easy then I can go about starting to do custom GUI and add it to your library via pull requests. Or maybe just keep it in my library for now.
Ok what about X/Y pad? That seems difficult to do custom because it requires somehow controlling two values or two RWidgets or i don't know.
By the way you can steal anything from my library that you want. Just let me know and then I can continue to fix issues on it via pull requests. All the jerobeam stuff is ready to go. I also have XoxosOscillator finished, this thing:
and another thing! I feel like there's a lot of basic classes missing from your library. Maybe you just don't realize how difficult things are for people like me who suck at math and need basic building blocks to start out with. The more I think about it, the more I feel like you need me to design the:
"Elan's Wonderful RS-MET Extension for Noobie Developers"
- easy fast gui
- simple building blocks
- etc.
ALSO!!! Does this mean your mod system was created just for your purpose and was not meant to be used by some who is paying for your DSP license?
Ok what about X/Y pad? That seems difficult to do custom because it requires somehow controlling two values or two RWidgets or i don't know.
if i remember correctly, i have functions like assignParameterX, assignParameterY which are analogous to the assignParameter function of the one-parameter-widgets (such as sliders). so, yes, the xy-pad controls two parameters
ALSO!!! Does this mean your mod system was created just for your purpose and was not meant to be used by some who is paying for your DSP license?
yes. actually more for your purpose, but i can make good use of it in my chainer as well. if someone buys a license for my dsp library, i of course don't mind, if they want to use that, too. but the dsp library is actually meant to provide the core dsp/math building blocks, not an application/plugin framework
ALSOO!!! You realize I don't use much DSP from your library? Antialiasing filter and some simple 1-4 pole filters... your noise generator. That's it! I want to get into more complex stuff like antialiased wavetable sawtooth, but... my point is... I am using your framework! not your DSP!
I am using your framework! I thought that was what your library was all about! It's not!
...such that non-math people have easy access to such heavy-duty algorithmic machinery as elliptic filter design
Antialiasing filter
this is actually an elliptic filter - the most math heavy thing in the whole library
I am using your framework! not your DSP!
and that's fine. but i want to sell a dsp library. the framework can be considered an optional add-on
and that's fine. but i want to sell a dsp library.
that's fine, but I wish I realized this a long time ago!
non-math people need simple things like linear ADSR.
nah - linear sounds mostly bad. has to be exponential. non-math people still want good sounding envelopes. they can use my breakpoint-modulator. that can do linear, exponential and more
well......... you can't modulate your breakpoint modulator. I designed my linearADSR to respond to modulation immediately, and it is abusable. It is also sub-sample-accurate.
Why do this? Since it's modulatable you can modulate it into any curve, Sigmoid, Expo, Log, Pow, and if you combine two linear ADSRs, you can further shape it by having one ADSR modulate the other. But I'm not finished with it. I eventually need to modulate the speeds so that it sustains rather than have a hard sustain phase.
true. modulatability was not designed in (so i could do some other optimizations). i didn't think that one would want to modulate an envelope generator. ...maybe someday, i can make a modulatable version
it's all straightforward stuff. gui code generally tends to be verbose
f***, i don't want to use filmstrips, there's a bunch of other design things to deal with
i want a more flexible solution such that client code can use bitmaps, vector graphics, its own algorithmic drawing or whatever it wants. using filmstrips...hmm...i guess that is a nontrivial bit of work. the first thing to do is to delegate the painting to its own object such that client code can do things like mySlider->setPainter(mySliderPainter). then you could just have your own subclass of some (framework-provided) SliderPainter baseclass and in your subclass, you could do whatever drawing you want
you would have to load an image, perhaps read in some metadata (how many images there are, if it's a horizontal or vertical strip, etc.) and then perhaps split the image int a set of images and then select which image to show based on the slider value
Ok how about vector graphics? That would be the easiest. But then what happens if you create a fancy vector thing and then you need some parts to rotate and not otherS? oh you'd need to build a knob out of multiple vectors for the parts that rotate and parts that don't
perhaps one could also rotate a bitmap image, if it's stored in higher resolution than finally needed
What about Drawable class? Is that what I want? Drawable::createFromSVG.
made a thread: https://forum.juce.com/t/how-do-you-use-vector-file-with-graphics-class/24462
Why do this? Since it's modulatable you can modulate it into any curve, Sigmoid, Expo, Log, Pow,
but why achieve these shapes via modulation rather than directly create the shape you want? this seems unnecessarily complicated to me.
non-math people need simple things like linear ADSR
i actually think, non-math people need a library with an easy-to-use interface that does some not-so-simple math under the hood. that's what i'm trying to provide
What about Drawable class? Is that what I want? Drawable::createFromSVG
yes, that's juce's svg rendering facility. i'm not sure, how complete the implementation of the svg spec is though. maybe, if a graphic artist uses one of the more exotic features of the svg spec, it might not be rendered correctly. but i really don't know. maybe juce has a complete svg support
nope - incomplete. this would then have to be checked with any concrete svg file that you create or get from a graphic artist:
https://www.juce.com/doc/classDrawable#a7b89bd78998ffd1f4888a143b15e2258
I just need simple svg stuff anyway.
Or would it be better to create a class that has stuff like: line size, line thickness, outer stroke, inner color, outer color... seems it would be easier just to use a vector program to define all that.
but why achieve these shapes via modulation rather than directly create the shape you want? this seems unnecessarily complicated to me.
Yes maybe it is but recursive functions are better than non-recursive when it comes to musical sound synthesis. Also, probably lighter on CPU than recalculating a curve every sample. But ultimately I'd like to have both. Also, I don't know how to create the shape without modulation and still have reasonable timings and still have nice curves. Also, I would still want to create the sustain stage by modulating the timing down 0 slowly, I guess that's like creating a knee. The sound of transition from decay to sustain is just ugly. Traditional ADSRs suck.