Aurora FM - New Win(32/64) VSTi FM synthesizer - Introductory pricing

[briefcasemanx]

Damn, I'm very glad to support such a cool dude! Thanks for developing cool stuff!

[RyFi]

Damn, I'm very glad to support such a cool dude! Thanks for developing cool stuff!

Well thanks again and you are very welcome and the pleasure's all mine! I'd still be developing my synth even if I was the only person in the world who used it, but let's face it, that would be pretty sad. It is so incredibly rewarding to create this thing knowing that other people like it and get enjoyment from it and put it to good use. And you fine folks have in turn helped make it even better for me and for everyone else with your suggestions and feedback and yes, even criticism. I look forward to many many more years of updates and improvements. (I just wish I could program faster and that the time scales were a tad more condensed...... :-/)

[briefcasemanx]

Damn, I'm very glad to support such a cool dude! Thanks for developing cool stuff!

Well thanks again and you are very welcome and the pleasure's all mine! I'd still be developing my synth even if I was the only person in the world who used it, but let's face it, that would be pretty sad. It is so incredibly rewarding to create this thing knowing that other people like it and get enjoyment from it and put it to good use. And you fine folks have in turn helped make it even better for me and for everyone else with your suggestions and feedback and yes, even criticism. I look forward to many many more years of updates and improvements. (I just wish I could program faster and that the time scales were a tad more condensed...... :-/)

Very good to hear! Do updates at whatever pace you can, don't worry about having to be mega fast with everything.

how long have you been doing DSP to get to the level you're at now?

[RyFi]

Spoiler: I really know very little about DSP in the formal sense; a tad of practical, sure, but very little in the theoretical knowledge area and certainly not capable of doing anything that hasn't been done better a million times by much smarter people than I. I'm pretty good at math (algebra and geo/trig yes, calculus hell no), and I'm getting to be a pretty solid programmer, and apparently my understanding of performance/efficiency/SIMD is better than average, but that's really about it.

I've been programming for about 20 years, and figuring out how to do cool shit with audio and MIDI and graphics only for about the last 10. The reason I started working on an FM synth for a hobby project 10 years ago was because I knew I knew the math for FM synthesis and that I most certainly did NOT know the DSP/math for subtractive synthesis, bandlimited generators, filters, all that complicated stuff!

So my synth has mostly been just a lot of math (some I knew, some I had to figure out), reading a lot and learning and getting informed about the DSP stuff that was beyond my novice understanding (like reverb design - I could never have written one if the amazing resources and papers I've learned from weren't available, and wow that's definitely been one of the most fun and rewarding things to work on), seriously digging into performance/optimization stuff, and being really motivated and serious and disciplined about trying to become a slightly better programmer every day.

Basically just a lot of dedication and stubbornness over a long slow and steady period of time. (I sure as shit wasn't some brilliant kid writing cool software in his teens and then got an EE or CS Master's degree/PhD or anything like that.)

Are you interested in DSP and audio / synthesis programming? I'd expect the person asking that kind of question to probably be doing so because they have at least some interest in it. Anything else I can answer for you? If you're asking because you yourself are interested, I'd actually really like to answer anything else I can if it helps you in any way, because I highly encourage it to anyone even slightly inclined.

[briefcasemanx]

Spoiler: I really know very little about DSP in the formal sense; a tad of practical, sure, but very little in the theoretical knowledge area and certainly not capable of doing anything that hasn't been done better a million times by much smarter people than I. I'm pretty good at math (algebra and geo/trig yes, calculus hell no), and I'm getting to be a pretty solid programmer, and apparently my understanding of performance/efficiency/SIMD is better than average, but that's really about it.

I've been programming for about 20 years, and figuring out how to do cool shit with audio and MIDI and graphics only for about the last 10. The reason I started working on an FM synth for a hobby project 10 years ago was because I knew I knew the math for FM synthesis and that I most certainly did NOT know the DSP/math for subtractive synthesis, bandlimited generators, filters, all that complicated stuff!

So my synth has mostly been just a lot of math (some I knew, some I had to figure out), reading a lot and learning and getting informed about the DSP stuff that was beyond my novice understanding (like reverb design - I could never have written one if the amazing resources and papers I've learned from weren't available, and wow that's definitely been one of the most fun and rewarding things to work on), seriously digging into performance/optimization stuff, and being really motivated and serious and disciplined about trying to become a slightly better programmer every day.

Basically just a lot of dedication and stubbornness over a long slow and steady period of time. (I sure as shit wasn't some brilliant kid writing cool software in his teens and then got an EE or CS Master's degree/PhD or anything like that.)

Are you interested in DSP and audio / synthesis programming? I'd expect the person asking that kind of question to probably be doing so because they have at least some interest in it. Anything else I can answer for you? If you're asking because you yourself are interested, I'd actually really like to answer anything else I can if it helps you in any way, because I highly encourage it to anyone even slightly inclined.

Thank you for telling a little more about your journey! Yes, I got interested in DSP but kinda stopped because some aspect of the way people were teaching the complex sinusoid and fourier theorem wasn't making sense to me and every explanation I found seemed to lack some basic insight or nugget of information that would make it click. Now that I think about it, I wasn't doing any equations on paper, just reading explanations in books. Maybe I would answer my own questions if I tried actually doing equations and solving problems rather than try to keep the theory all in my head.

Or maybe fourier analysis is not a great starting point. How would you suggest getting into DSP?

[RyFi]

Well, I most certainly don't want to overstep my bounds and project any kind of impression of knowing more than I do, and I definitely don't want to assume a role of the blind trying to lead the blind, so I'll just very cautiously give what *I* see to be some safe, unpresuming, hopefully helpful basic suggestions without any kind of proselytizing or claims of authority.

I know next to nothing about Fourier transforms and don't employ them in my synth, and from my perspective, I would not recommend that kind of stuff as a starting point. I'd go way way back and ease into things in bite-sized bits. Build basic building blocks. Create a simple variable frequency sine generator. Modulate the phase of that sine generator with another sine generator (tada, simple incredible FM!). Create an envelope, just a simple 0 > 1 > 0 AR or ADR envelope to run the sine through, with options for linear, logarithmic, and exponential slopes and parameterized times. Extend that, with more flexible slopes, more stages, maybe add looping. Now you've got simple things that were easy to make, but guess what, now you've got basic tools and you can start assembling them in really cool ways.

Interface to a realtime control, like the ModWheel. Route the ModWheel to the various parameters of your building blocks. Notice how a stepped unsmoothed control signal causes various ticks and noises and artifacts in the output when wired to various parameters. Make a smoother, either a trivial linear one or a simple IIR. Now smooth the input from the ModWheel and tune it until you've eliminated the ticks/artifacts. Now you can smoothly change your sine generator's frequency in real time, or control your envelope's amplitude, or run your envelope faster or slower, or ...

Learn about FIRs. Write a generic FIR class. Write a kernel generator. Apply Hamming or Blackman windows to your kernels. Use different kernels with your FIR to filter your signal somewhere, low-pass or high-pass or whatever you want. That FM block you made earlier, set the frequencies / mod indices so that it aliases. Now extend your FIR into a decimation filter with the appropriate low-pass kernel. Double or triple the sample rate of everything (you did parameterize sample rate everywhere, didn't you??) and route everything through the decimation filter and see if how it performs. Tweak and check, learn about the tradeoffs in response vs. kernel size and latency and computation cost and all that.

NOW that you've done all that, you've probably learned a shit ton and you have all these cool useful building blocks. Now you can learn about Fourier transforms and write an FFT and play with all of that stuff, or maybe you really like something else and want to explore that. Write an effect, like a phaser or flanger or vibrato or tremolo or chorus or delay or compressor or limiter or reverb or bit crusher or.......

You see where I'm going? Build small, initially simple, bite-sized pieces that do useful things, that can do cool things when connected together, and gradually incorporate more and more complicated pieces. You'll learn and accomplish and get confidence and satisfaction all along the way. That way it's not all "I jumped into the deep end and tried to understand Fourier transforms but I got stuck and frustrated and now I hate DSP and give up on this" but instead "I slowly but surely learned how to do all these cool little things, in baby steps, and now I can put them all together and make something cool and unique or I can learn more and now start to really challenge myself with some truly heavy stuff and I'm in a way better position for making sense of it or knowing when it's beyond my abilities and then I can just go down a different path and make something else cool that I want instead".

Does this seem at all helpful? That's my basic advice from my own very specific experience that worked out just fine and dandy for me, and nothing more presuming or authoritative than that. Groovy?

[briefcasemanx]

Well, I most certainly don't want to overstep my bounds and project any kind of impression of knowing more than I do, and I definitely don't want to assume a role of the blind trying to lead the blind, so I'll just very cautiously give what *I* see to be some safe, unpresuming, hopefully helpful basic suggestions without any kind of proselytizing or claims of authority.

I know next to nothing about Fourier transforms and don't employ them in my synth, and from my perspective, I would not recommend that kind of stuff as a starting point. I'd go way way back and ease into things in bite-sized bits. Build basic building blocks. Create a simple variable frequency sine generator. Modulate the phase of that sine generator with another sine generator (tada, simple incredible FM!). Create an envelope, just a simple 0 > 1 > 0 AR or ADR envelope to run the sine through, with options for linear, logarithmic, and exponential slopes and parameterized times. Extend that, with more flexible slopes, more stages, maybe add looping. Now you've got simple things that were easy to make, but guess what, now you've got basic tools and you can start assembling them in really cool ways.

Interface to a realtime control, like the ModWheel. Route the ModWheel to the various parameters of your building blocks. Notice how a stepped unsmoothed control signal causes various ticks and noises and artifacts in the output when wired to various parameters. Make a smoother, either a trivial linear one or a simple IIR. Now smooth the input from the ModWheel and tune it until you've eliminated the ticks/artifacts. Now you can smoothly change your sine generator's frequency in real time, or control your envelope's amplitude, or loop your envelope faster or slower, or ...

Learn about FIRs. Write a generic FIR class. Write a kernel generator. Apply Hamming or Gaussian windows to your kernels. Use different kernels with your FIR to filter your signal somewhere, low-pass or high-pass or whatever you want. That FM block you made earlier, set the frequencies / mod indices so that it aliases. Now extend your FIR into a decimation filter with the appropriate low-pass kernel. Double or triple the sample rate of everything (you did parameterize sample rate everywhere, didn't you??) and route everything through the decimation filter and see if how it performs. Tweak and check, learn about the tradeoffs in kernel size and latency and computation cost and all that.

NOW that you've done all that, you've probably learned a shit ton and you have all these cool useful building blocks. Now you can learn about Fourier transforms and write an FFT and play with all of that stuff, or maybe you really like something else and want to explore that. Write an effect, like a phaser or flanger or vibrato or tremolo or chorus or delay or compressor or limiter or reverb or bit crusher or.......

You see where I'm going? Build small, initially simple, bite-sized pieces that do useful things, that can do cool things when connected together, and gradually incorporate more and more complicated pieces. You'll learn and accomplish and get confidence and satisfaction all along the way. That way it's not all "I jumped into the deep end and tried to understand Fourier transforms but I got stuck and frustrated and now I hate DSP and give up on this" but instead "I slowly but surely learned how to do all these cool little things, in baby steps, and now I can put them all together and make something cool and unique or I can learn more and now start to really challenge myself with some truly heavy stuff and I'm in a way better position for making sense of it or knowing when it's beyond my abilities and then I can just go down a different path and make something else cool that I want instead".

Does this seem at all helpful? That's my basic advice from my own very specific experience that worked out just fine and dandy for me, and nothing more presuming or authoritative than that. Groovy?

Excellent, that makes a lot of sense. Thank you for the response, i'm bookmarking it.

[RyFi]

Edit to be a bit less abrupt and more appreciative:

By the way, thanks so much for asking, which obviously shows a certain level of faith in my abilities or recognition that something I've accomplished is at least a little bit beyond trivial. That really means a lot to me. I'm happy to share my perspective/experience and hope it helps you in any way.

[ElVincente]

I'm ashamed I have so much soft synths I never even touched already but the poly at might make me do the move on that one. I still have to test it though, wondering if it feels different creating sounds on Aurora than with my beloved msoundfactory FM module...

Hey man, please let me know if you find any issues with the Poly AT. It's incredibly hard to test that properly without a Poly AT keyboard and Reaper's atrociously painful manual entry! I'm really not ultra confident that the Poly AT stuff is flawless, but I certainly tried my best... :

Seems to run perfectly well with reaper and Mulab until here. I'll quickly test on Bitwig this evening too.

[RyFi]

Seems to run perfectly well with reaper and Mulab until here. I'll quickly test on Bitwig this evening too.

Awesome, that's encouraging and a big relief. Thanks for putting that functionality to use and letting me know!

[halcyon68]

Glad to see this awesome synth is alive and well.

Not sure if I missed this somewhere and somehow but one of the suggestions I'd like to make for the future is to give each part its own output.

[RyFi]

Glad to see this awesome synth is alive and well.

Not sure if I missed this somewhere and somehow but one of the suggestions I'd like to make for the future is to give each part its own output.

Thanks man. Indeed, that's definitely something I'm well aware of and trying to figure out a plan for. At some point there will be a jump to flexible I/O / internal busses, and that opens up the door for being conducive to supporting more parts, audio-in for an operator's wave source, additional processing (ex: bus comp/lim), all that good stuff. I just haven't decided if that'll come before or after a migration to VST3, but I suspect it's the latter that makes the most sense... I'm in the planning stages now. It may be a little while before I have anything for ya though .

[crickey13]

A built-in soft clip/sine shaper saturation with drive and dry/wet controls would be nice too.

I don't like the built-in overdrive that much, it's more along the lines of a distortion type of effect, so it's suited just for conspicuously distorted stuff.

It would be nice to switch between the conventional overdrive and the more subtle soft clip, it would seriously expand the sound palette IMHO.

A FX bypass button would cool too. That being said, these are just suggestions of course.

[RyFi]

I don't like the built-in overdrive that much, it's more along the lines of a distortion type of effect, so it's suited just for conspicuously distorted stuff.

It would be nice to switch between the conventional overdrive and the more subtle soft clip, it would seriously expand the sound palette IMHO.

You're not alone. It could be much better, but I've always had the need to prioritize other things first.
It's kind of the neglected child of the bunch. But I'll try to make/find the time to give it some love.

[crickey13]

Cheers. FM sometimes sounds "hollow" and subtle saturation can go a long way, giving it more body. I have really fond memories of the Operator's built-in soft clip, which can really make FM sound much warmer, punchier and push it in the more analog direction.

An internal soft clip or a sine shaper would really complement the sound of Aurora FM IMHO. But obviously, first things first, the VST3 version has a higher priority for obvious reasons. Aurora FM is very usable as it is.