this one looks interesting

[musictop]

Some interesting ideas and tools in this one

https://kodasampler.com/

[Lindon]

@musictop yes apparently these people have signed up for it:

soniccouture, soundiron, realitone, and joshua bell

[David Healey]

@musictop I guess the figma integration is appealing to a lot of people.

[dannytaurus]

Tangentially related - over time I believe more and more that HISE will be left in the dust in the coming years, as AI-powered software development brings a large number of apps and services to the market, aimed at creating plugins and tools for music production.

It's depressing and exciting at the same time.

[Oli Ullmann]

@dannytaurus
Another possibility:
The production of audio software that goes beyond a simple sample player is very complex. From knowledge and skills regarding DSP technology to knowledge regarding the tools, frameworks, or third-party software to be used, a lot of know-how is required to produce a plug-in that is truly competitive.

I don't think this will be achievable in the next few years through AI alone, as we are not currently seeing any real development in AI technology. The only development is: “Let's build bigger data centers.” I think it's clear that this cannot work in the long term and does not lead to any real further development.

Therefore, the future could be that AI is used as an additional tool, but in the end, there is still a programmer behind it who has to put everything together. And HISE is wonderfully suited for this. HISE is great at combining code from multiple directions (e.g., RNBO, Faust, etc.) and creating an end product from it. Therefore, it could also be that HISE will continue to grow with the support of AI workflows.

[dannytaurus]

@Oli-Ullmann I agree that humans will still be required to make competitive plugins. When I said "AI-powered software development" I meant software that is developed with the enhancing (speed and breadth) power of AI, not that AI should make the software itself.

Up until a year ago, if I had an idea for a sample management app, or plugin organiser app, etc, it was very difficult to tackle those ideas from the ground up as a solo developer. But now those apps are starting to appear, and they're being made by one person with the help of AI.

With music production being a niche software market (compared to things like social media, finance, etc) the potential ROI for an app had to very large for a company or team to consider building it. Just because our market is niche, it doesn't mean the software is any easier to build.

But now, those ideas and be explored and taken to fruition by a single dev.

HISE has an unusually large barrier to entry that new AI-minded developers won't endure.

Everything will still probably be based on JUCE, until a new audio framework comes along, but the abstraction wrappers around JUCE will get a lot better.

[Oli Ullmann]

@dannytaurus said in this one looks interesting:

HISE has an unusually large barrier to entry that new AI-minded developers won't endure.

What exactly do you mean by that?

[dannytaurus]

@Oli-Ullmann The requirement to use GitHub to clone the source, including the new submodule requirement (which even as a git user of 15+ years I've never used), and to compile HISE with all the necessary libraries, using at least 3 different apps (terminal, Projucer and build environment Visual Studio/Xcode, etc).

When you've finally managed to compile HISE, you're faced with either letting it slowly rot out of date, or do regular updates which involves repeating the compilation process and all it's complexities.

Then you face a similar, if less severe, situation of exporting plugins.

You only have to look at the recent forum posts to see how many people are struggling hard with compiling HISE and then exporting plugins.

[David Healey]

@dannytaurus Yeah the submodule thing is a pain, as long as Christoph never updates it it isn't too bad, but he did update it the other day so I had to mess around with submodules again :p

The barrier to entry is definitely quite high, but I think once you're over it then it's mostly not too difficult to keep up to date.

And if you build a Rhapsody instrument there's not need to compile HISE at all, but I need to work on making Rhapsody development a bit more accessible.

[dannytaurus]

@David-Healey And all I'm talking about there is getting set up to use HISE. Let alone the challenges in actually developing in HISE.

I'm not trying to sh*t on HISE. I love HISE, and for the most part I find it very enjoyable to develop plugins. The possibilities are vast compared to where I came from - Maize Sampler.

But when you hit a wall, it's often difficult to move forward. It might be missing, outdated or incorrect documentation. It might be hitting the barriers of Christoph's vision for certain aspects of HISE. It might be, that after investing dozens or hundreds of hours in HISE, you finally realise that you can't achieve you goal without plunging into custom C++, which makes HISE a convenient UI playground at best.

At least with something like Maize, you know the limitations from the start. And all the rest is easy.

Compare the whole process of cloning, configuring and compiling HISE (and then repeat the configuration and compilation process, PLUS notarisation, to export a plugin) with Maize Sampler - you download the app, you chuck in a load of samples, you add simple UI elements and then your plugin is compiled on Maize's own servers, fully notarised and ready for distribution.

I'm not saying Maize is in any way comparable to HISE in features. But I've built a real business, making comfortably more than minimum wage, just selling Maize-based plugins.

PS I can't upvote any more of your posts today!

[David Healey]

@dannytaurus said in this one looks interesting:

It might be, that after investing dozens or hundreds of hours in HISE, you finally realise that you can't achieve you goal without plunging into custom C++, which makes HISE a convenient UI playground at best.

This one I see people hitting a lot. My advice is plan your project around HISE's capabilities rather your idealised outcome. But of course if you don't know HISE's limitations then how can you plan, which is why my other advice is to built lots of little projects in HISE to get to know it before you dive into a big project.

@dannytaurus said in this one looks interesting:

PS I can't upvote any more of your posts today!

I hit the same bug