I think I'll feel better once we have a few ppl testing so I can get all the harsh feedback I really need

I wish I could afford a UX person tho. This project is already a mess and I've been trying to sketch out the next things with Estelle and I'm just worried this is all about to get even more wildly complicated for the user,

So I am just gonna do my own thing, self fund, see how far I can get with community and OSS help. Until I can't anymore

And I feel like this was maybe a thing valued more in the beginnings of audio programming, at least that is what I can gather from the stories being told, and the folks people look up to in the industry (all white men btw, so maybe that wouldn't have worked out for us anyway),

Honestly I feel like a C++ hack, but everyone starts somewhere right. It'd be interesting to get an audio dev job but I feel like most studios/orgs are probably looking for past experience with good DSP backgrounds. I learn v quick on the job & to me its balancing learning skill w working with ppl,

C++26 has "^^" aka. the "cat ears operator"

Anyway, it sounds like you're speaking from insider knowledge/experience or being acquainted with it from the very beginning don't mind my rambling about it lol. It's been on my mind since I've been tasked w writing a bot very recently and we also started using IRC again recently

I think u still can. I always thought it might've been the other way around, (tech they put an IRC wrapper on like Slack used to do), but given it's been around so long maybe it really is IRC core w/ proprietary bespoke extensions & REST/other thing built on top of.

We're currently prepping for travel and stuff but we're gonna reach out to u later on Discord if that's ok :3

Doing the envelope and sample editor was also an experiment with using QPainter. I have a plan to migrate the tracker UI to QPainter, as it's currently on being drawn using ImGui, and we have reached the limits of immediate mode for a long time already, esp with us drawing waveforms on everything

It's not something we use heavily in .IT, & when we do it's mostly for doing drum kits. Peeking at samplers in various other DAWs etc for inspo, tho I might settle for a simple list/columns like what OpenMPT has for now.

We're still missing a lot of UI paradigms to make a good key zone map UI rn

The last basic feature missing from fully supporting .IT in UI (besides a few effects that I removed/didn't re-add support for), is the key zone mapping between samples and instruments. It supports them in the engine (in import/saving), but just no UI for it yet,

And a basic sample editor (reusing the operations code from chibitracker, which is nice.. only I put a very rudimentary UI over it lol).

Anyway, I accidentally stayed up all night making these two things, ooopppsss.…

(CW: tiny flashing light)

First time I tried making an envelope editor. For now it only supports the standard .IT features (which means no beziers or control curves yet.. limited to 25 nodes, etc.)

When we& don't have confidence and we're viewing ourselves in a very negative light.. even when we are self aware of it it's really hard to just cut through that.

Hard to know how much of this is really being back on hormones and medication, or just relief at the fact we are, or placebo, or onset of hypomania, or having gotten thru some challenging dev work, or a mixture. But I am feeling generally calm and refreshed and more motivated

I couldn't help myself and implemented VST3 instruments anyway…

Excerpt of a Soulver notebook showing various calculations of how far you could count with 64-bit or 32-bit ints 64-bit ints: • Can count in microseconds for up to ~586,549 years • By samples (at 44,100 sample rate) for up to ~13,300,440 years • In seconds by 586,549,402,018 years 32-bit ints: • Can count in microseconds for up to ~1.2 hours • By samples (at 44,100 sample rate) for up to ~27 hours • In seconds by ~136 years (yep, the 2038 problem)

This is not really a surprise, but curious what it looked like quantified side to side and with different time units. It does alleviate any lingering worry I had about just eternally increasing a uint64 counter even at microsecond precision.

(props to @estelle.rosemade.art for making it)

Anyway these are all just loose thoughts, I haven't properly formalised any research yet w Val I think the next round of UI stuff which I've started mocking up will help us figure things out. No radical changes yet

Could be interesting tho, it would lead to us having VST and AU support in Godot

Other wild ideas were making this in Godot, which would be a weird full circle given the current tracker base is an old tracker written by one of the original Godot devs.
But there is a lot of heavily lifting is in audio code and systems that would essentially need to all live in native extensions,

Desktop app innovation seems mostly driven by creative tools, of which many are built atop proprietary/long term developed systems. Either way it takes a shit ton of work to make something look and feel nice while also feeling like a desktop app,

Desktop Apps are in this weird state thanks to much of US tech going all in on web and thus all corporate funding and dev focus abandoned desktop. Native mobile has a bit more focus bc it is a powerful platform to them

QML also makes me feel nervous about the amount of additional design and implementation work that needs to be done. Muse Group had to create an entire components library and without a dedicated design team that could take us over a year

Multi window, docking, etc, lots of little design & UI interactions that are specific to each OS. Much we would have to reimplement from scratch to make a webview rendered app feel native.

I am also unsure about Qt Widgets in the long run, vs, exploring QML. At that point is looking at Webview2 a better option? But also we use Qt because it abstracts away a lot of native OS stuff like file open dialogues, menus, accessibility, lots of subtle patterns that all add up,

My other big task is needing to port all the ImGui stuff to QPainter (I have only been using it for 2D drawing and we've outgrown immediate mode stuff by now)

This will make us even more entangled with Qt but eh

Then I'm gonna have to do it all again for Audio Unit.. and possibly CLAP… if VST3 isn't enough for Linux

And how to shift/restructure the current audio engine towards that. Long way yet to go still but the road is clearer

Screenshot of a experimental work in progress tracker/DAW showing an assortment of VST plugins with their editor windows open: chipsynth SFC, Xfer DJMFilter, and Fabfilter plugins: Pro-L, Pro-C, Timeless

I worried this little adventure was gonna be a waste of time or at least, a huge distraction from other things I need to do in this app. Bc we don't have the UI to quite support these yet. But the deep dive into VST3 hosting has given me a better idea of how I want to structure events/audio graph,