Quite a few big hurdles there, I'm sure you are well aware that this hugely slows the likely timescale and may stop this sort of thing really taking off at all.Roger_Linn wrote: ↑Fri May 17, 2019 7:14 pm Yes, I’ve seen those and they’re good steps in the right direction.
I think that software instruments for embedded Linux hardware products won’t have much quality until:
1) a standard hardware platform for Linux-based embedded hardware instruments is agreed upon, so any software instrument will run on any hardware product, and
2) many hardware products exist that support such standard, so the market is big enough for software instrument makers to have incentive to support the format, and
3) software instrument makers are able to implement reliable copy protection on hardware instruments.
I applaud the existing hardware devices that are open in some way, and I wish them well, even though most of them will remain in their own little niche worlds, used by enthusiasts with some great and flexible results, but not mainstream or attractive to commercial software instrument & effects developers.
While wearing both a dev and user hat, I like to take a step back and consider what factors make people want a hardware instrument rather than a computer in the first place. Clearly much hardware these days is a computer really, so what we are mostly talking about is a computer that is dedicated to specific tasks, with a physical interface to match. I wont try to do the whole subject justice here, but its interesting to study what things affect user perceptions. For example, slapping a reasonably large screen on the hardware, and relying on things like navigating menus, starts to head too much towards the computer experience for some people.
Having indulged in quite a lot of hardware over the last 18 months, its definately the physical interface (including many knob per function things etc) that is the main selling point for me. As such, I have become less interested in general purpose platforms that can run multiple soft instruments, and more towards things where the interface can be expertly crafted for that particular soft instrument. Sometimes less is more, and by avoiding having a general platform we avoid some of the hurdles you mention, and also avoid people making direct comparisons between the limits of the platform and what can be done on a traditional computer.
Its interesting, I dont know where its going, and I wish things were further ahead with various competing ecosystems than they are. There are certainly signs of some players in the game, coming from various different angles.
For example there is ELK MusicOS, which have been at the tradeshows a few times now and so far have focussed on making an embedded Linux platform that people will use to turn a particular VST, Reason rack extension etc into a hardware instrument. eg:
http://cdm.link/2019/05/hardware-reason ... xtensions/
http://cdm.link/2019/05/hardware-vst-st ... etrologue/
Not really possible to evaluate this without contacting them to form a commercial relationship, and there are no real consumer products resulting from this stuff yet.
A different angle/different piece of the puzzle (language and certain platform/DSP aspects) is being tackled by the JUCE people, with SOUL:
https://soul.dev
Anyway if I keep talking I will probably keep coming back to the dedicated hardware UI point. Linnstrument is a great example of managing to do a decent UI without a screen, and without the user having to remember tons of obscure things. Doable because of the limited scope for what the instrument does/amount of config options, and I'm sure it was still a challenge to fit all those labels into the available space!
When I look at a classic instrument that is highly regarded for its playability and extra expression, the Yamaha CS-80, I think the physical controls were a big part of this. Aftertouch was a first class citizen, with more than one physical slider to dictate its various influences over the synth. They could do this because it was a beast, but the actual synth architecture and number of parameters the aftertouch can control is limited. But I certainly keep this in mind when I think of what a stunning MPE instrument would be like, there will surely be controls for the MPE dimensions, right there, taking centre stage not an afterthought or menu item.