My company's transition away from JUCE

Are y’all using Electron or Chromium Embedded Framework or something else?


The idea of professional poets complaining about being unable to make money from poetry because any old fool could acquire the materials and toss off a few trite lines precedes mass media. This was already the case 200 years ago. It was the case 2000 years ago, see Satyricon. William Rowan Hamilton, the genius mathematician who discovered quaternions and contributed to many fields of science, also tossed off depressed love poems from time to time, and his friend William Wordsworth supposedly gently tried to encourage him to focus on his great gifts of mathematics (because his poetry was awful). He didn’t need a pro to satisfy his need for poetry.

I am not at all saying that Gucci Gang sounds better than anything by Beethoven. I am saying two other things:

  1. There was plenty of music in Beethoven’s time that was far worse than Beethoven, and there is plenty of music today that is FAR better than Lil Pump. There most definitely is.
  2. People need music/art/entertainment, but they have never needed elite specialized classes who do only that. What percentage of the population today can listen to a Beethoven piece and actually notice and identify what makes it special? How many would notice if you reversed two movements or ended a piece in the wrong key (which has been known to happen in front of sophisticated audiences)? Would you? Even if you only include the European population, I wager less than 1% have any idea.

How do you plan to get around the memory limitations imposed by the browser?

e.g. Bandlab runs into problems with projects holding more than 10 minutes of audio due to the memory limit of V8 on a single browser process. Are you holding everything server side? If so, how do you deal with network latency?

This is interesting stuff you bring up here. I had never heard of it before, so I cannot form a reliable counterargument against it. All I can say, is that animals including humans organise their societies according to dominance heirarchies. There will always be people looking to “win” and be better than everyone else. Therefore, they will create a demand for the best ideas. You can compete to become the new elite selling the hottest algorithms and song ideas to them.

Music is a subjective experience largely because the brain likes a mix between the familiar and the new. Beethoven is actually too new for most people’s brains because they’re not used to listening to that type of music. They can’t appreciate his works because of this, in addition to the fact that he doesn’t create repetitive music. Repetitive music becomes familiar very quickly, but classical styles of music do not generally do this, aside from a few exceptions.

So I think we’re comparing apples to oranges on this point.

The songs will not require 10 minutes of audio. The server will convert each synthesizer into a sampler and download the samples for each note that is expected to be played by the user or in the composition. The browser will then play the samples when needed. There will not be live processing of effects. The effects will be baked into the synthesizers.

In Satyricon, there’s a poor but devoted poet who complains to the main characters about his talentless rival who is adored by all and throws huge parties. Later, he himself rises to be revered the same and also throws such parties. At the end, he dies, and his will states that his belongings should be divided among his friends, provided they devour his body. They do. This is also shown in Fellini’s version. Regardless, true artists were already complaining about no-talent hacks muddying the waters!

I agree completely. You have to compete, but complaining that the means of production (of music) are easier to come by for people with little experience or budget doesn’t change the reality. I’m all for making more tools available to more people; where technology is concerned, this tends to be good for progress.

People make music for the love of it and will continue to do so whatever the machines are doing. It is one of many ways we express our emotions; art is how we connect with others transcending linguistic and physical barriers. Human creativity is virtually limitless, and there are now virtually no barriers to entry for musicians. I don’t really see a problem here that needs solving, but it’s a very interesting computer scientist perspective and I’ll be interested to see how this work contributes to the art.

That’s because we live in a society where our political discourse is obsessed with inclusiveness but what they’re focusing on is only the barrier: whether they’re talking about national borders, walls, immigration, pay gaps etc. But there are more complex issues that exist after you gain access. The music industry in particular has an incredible amount of politics and this has led to widespread mental health problems. What is needed now is a focus on minimising the barriers to success.


What are your thoughts on Soul. To me it appears to be a return to simplicity.
C++ programmers can be proud to be smart enough to be using C++, but there are many kinds of smart that are not comparable with the faith and patience required to learn C++. No high functioning, fast firing ADHD brain likes that sort of torture. C++ has a ridiculous entry barrier that will not doubt kill itself and discourage many possible future geniuses.
Isn’t Soul going to gel well with AI, as it will be dynamic and therefore easily dynamically reconfigured by the AI?
Cheers and good luck

1 Like

To be honest, I had been away from this forum for a while, so I only learned about soul two days ago and I haven’t used it yet.

I may have misunderstood but from what I can see it is a positive step in the right direction but it is still a different approach to the problem because it is essentially seeming to make it easier for people to build things that have already been built a thousand times. There are so many EQ plugins for example. Do we need more of this? Why not have more plugins that transform midi compositions? That’s what I’m starting with, for drums specifically.

I prefer not to work with soul though because I already tried making a business relationship with Roli after they granted me a free pass to the ADC (paid for by Google) but they strangely refused to respond to me emails. On the other hand, I work down the road from Google’s headquarters in Sydney and Google have always been supportive so I would prefer to partner with them.

We do need improved Eqs until the maths is correct and we loose all that horrible digital harshness. Once this is achieved then those Eqs should be aided by AI.
AI assistance for music will be like film editing and special effects for acting. The Matrix live drama production was a disaster, but it had some mighty fine acting. Technology takes acting to a whole new realm. Likewise, AI assistance will enhance and speed up talented musicians in their craft. I shouldn’t have to be able to play an instrument to put down a song that is in my head.

1 Like

Sorry if I was too blunt. It’s just that I come from an electronic music background and I always resented that I was never taught to make more use of the settings built into the synth instead of relying on plugin effects.

If you have a great synth and know how to configure it, you get zero artefacts and therefore hypothetically the sounds will be perfect. No EQ required. EQ is like make up.

We have to consider the possibility that there is no perfect maths for EQ. Maybe a different kind of algorithm is required. My suggestion is to analyse the original audio and then digitally convert it into a synth and render it with the desired EQ characteristics. That will be a perfect outcome and is something I will be working on soon.

I like what you are saying. The perfect maths will still be needed for clear ADC, DAC, image resizing and vocals(maybe?) It may also help analyze the audio(another maybe?).

I think the tricky part will be the human interface. If I learn the keyboard to write that song in my head, will it remain the same song? Music is moving beyond the confines of Western tuning. With AI, I would only need to compose the dominant parts of the song I envisage. AI, like a musician, could fill out the rest that is implied, but not imagined. AI can suggest new note intervals or unusual rhythms that I am not familiar with and educate me as I write.

If you write it in Soul then you would be more connected to the current world of audio developers.


Yeah the perfect math is still needed for analysis so that the AI can limit its focus, instead of trying to do everything.

In response to what you said about the process of making songs, I don’t think there will be one way of making music. I think there will be thousands of methodologies and infinite composites created by combining those methodologies. A real diversity of ideas, processes and approaches. It will be up to us individually to compete against each other in a market of ideas.

I just had a closer look at Soul and I get what you’re saying. But there’s a few problems with this.

Firstly, my company moves a lot faster than ROLI. And Google’s V8 (JavaScript) engine is already ready to go. Its just not possible for me to wait for ROLI to get stuff done especially with the low level of communication that comes from them.

I have a good reputation with investors, and if I make a poor choice here, it could cause me huge problems. Millions of dollars in potential funding could dry up as a result of things that are beyond my control. Very risky.

And lastly… you have to bear in mind that many things that we’re used to and take for granted, exist for arbitrary reasons that don’t make sense anymore. A well known example of this, is the QWERTY layout on our computer keyboards. I believe that our music production process has a bunch of “QWERTY’s” floating around that just make music less accessible and less productive. I’ve been playing around with algorithms for almost a decade now, and have come up with some innovative patent-pending interfaces that take things in a new minimalist direction.

And I’m not too sure that Soul is really useful to me.

I did a quick search for your email address and all the emails on which I was cc-ed had been replied to, including one where you appeared to arrange a skype call with J.B., so it’s pretty disingenuous to come on a public forum and call us “unresponsive”.

Despite the team being overloaded, we make it a priority to go out of our way to engage with people who contact us, even when it’s often not the best use of our time.

Well, best of luck with what you’re doing. Maybe you’re lucky enough to be building an audio app where you won’t hit any of the many obstacles that web-audio will throw at you.

SOUL is in its infancy, and although we’ve got a group of technologically adventurous companies who are getting aboard early, we’re not expecting the average developer to “get” what it’s all about just yet.

But perhaps if your project survives a couple of years and starts to run into performance and complexity problems then you might find you’re the target audience for it.


I have multiple email addresses because after winning the ticket to ADC, I came into several good opportunities which allowed me to start a company. And it was immediately after winning the ticket to ADC in 2016 (or 2017?) that I was talking to JB trying to thank him and saying I was looking forward to talking to him and others at Roli about my plans. And he completely ignored my email. When I attended ADC, I greeted him, but he didn’t make an explanation regarding my previous email at that opportunity. Maybe that means nothing to you, but to somebody just starting out, its a big deal to feel rejected like that by people you look up to, when you’re just getting started.

Somebody else at Roli suggested that I should apply for an internship. My advisers told me that the experience was basically at attempt at “negging”, so that’s how I’ve always viewed the interaction.

And more recently, I emailed looking for premium support regarding user authentication and automatic updates as implemented in the ProJucer. My email was ignored. And then a week later, I followed up to say that I had found another means of getting this work done, that email was ignored too.

I don’t mean to upset you regarding my comments about SOUL. I just don’t have a lot of information about what the goals are. And quite frankly, there are always options out there. If hypothetically Roli is a good company to partner with, it could save me some time and money. Alternatively, I could just raise the money to fund an alternative project.

Name one