"Prophet 5" project

[IbnCocoon]

Hi, I'm new here...
Some months ago I started with Circle, first with Mt-32pi and later with Mini Dexed, two great projects. The first to say goodbye to my outdated romplers and the latter to carry less weight in live concerts.
I have a poor background in C++, so after a los of trial and error, I managed to compile and modify the miniorgan example. With Chat Gpt help (do not laugh at me, please) I started to isolate basic modules, such as midi serial input, lcd display, sd card file read, i2s dac, etc, so that I can build some simple poly-synth with Sequential Circuits Prophet 5 in mind. I know I cannot recreate the sound nor the nuances of that amazing instrument, so this cannot be named a clone, but the topology is somewhat straightforward to achieve.

If somebody is interested in the code I can share it. I'm sure it has IA nonesense everywhere, because my understanding is very limited. Now I'm implementing oscillator sync, unison mode and spread control. In the future the plan is to add stereo chorus and stereo delay (remember NI's Pro-53??). Now on the project is headless, but I've made a Ctrlr panel. And is running in a single core, so voices are limited to 10.

Here some samples:
https://voca.ro/1lNYdurQmq7T
https://voca.ro/1iYb4PsRFRpm

[rsta2]

Sounds great! 😃

[IbnCocoon]

Hi Rene, please have a look...
Now I'm lost with multicore...mini dexed is too complicated for me yet...

https://github.com/IbnCocoon/LutSynth

[rsta2]

How can I help? BTW. There is a simple multi-core example in sample/17-fractal.

[IbnCocoon]

There is a simple multi-core example in sample/17-fractal.

Yes I've already seen it, but I didn't find out how to make it work not having a screen. My "setup" is a Zero 2W Pi, PCM5102 I2s dac and serial midi. Also I have a I2C 20x4 lcd display, but never managed to use it with Circle, only with Mt32pi and MiniDexed.
My starting point was 29-miniorgan sample. I've tried to "mix" 17-fractal with it, so that miniorgan is running in one core and a simple actled blink in another, but with no success.
Also not having other debug feedback than the ActLed is making things a lot harder...and never managed to get Logger output either. My little sd card is praying for his life being so many times plugged and unplugged jajaja
Sorry if I'm too beginner for this kind of projects...thanks for your help!

[rsta2]

You need a debug output for the logger. For a project like your this is absolutely necessary. If you do not have a HDMI display, the serial interface is used for MIDI and the 20x4 LCD is probably also too small, you can try a soft-serial solution. Please have a look at sample/09-softserial. With it you can use any spare GPIO pins for the TxD and RxD pins. You can configure this in kernel.cpp. Have a look at the sample's README.

I don't know, which kind of computer you are using for development. If it's also a RPi, you can directly connect the wires, otherwise it's likely, that you need an USB-to-serial adapter. They are not expensive.

[IbnCocoon]

Ok, now I've a hdmi display working. It's incredible how easy was to set it up!
I have tried sample programs 03-screentext and 17-fractal, both with success. From the first one I've learned how to integrate the screen to my project via Logger messages. Still I have no clue how to include multicore.
I have a mixer class that mixes all the rendered voices into one buffer. That buffer is later read in my main process loop (CLutSynth::Process) and passed to m_pSound->Write. Can I just render -let's say- 10 voices in one core, 10 in another and mix them into my existing mixer buffer?
Sorry if my questions are too elemental, salutes from Argentina!

[rsta2]

Ok, now I've a hdmi display working.

Very good.

Still I have no clue how to include multicore.

In fact adding it is not difficult. You only have to use CMultiCoreSupport as a base class of your CVoiceHandler class and add a Run() method to it. When you call Initialize() on it, Run() gets called on each core and you can do, what you want there. You have to add the --multicore option, when calling configure before build. sample/17-fractal is a very simple multi-core program. Please read the source code and try to understand, how it works.

Can I just render -let's say- 10 voices in one core, 10 in another and mix them into my existing mixer buffer?

My MiniSynth project is doing that. Have a look the voicemanager.* files. This is the place, where this happens. There is a comment in the header file, which explains the synchronization.

You have to know, that when you write a variable from one core, this can be seen from the other cores. Variables used for synchronization should be defined as volatile, so that the compiler does write them out to memory immediately.

Please try to understand this and ask question(s), if you have one!

[IbnCocoon]

Hi! Now it's working after a lot of trial and error. As you have said, the screen was mandatory.
Understanding MiniSynth's voicemanager helped me a lot.
Now I can return to the audio code to upgrade the bad sounding filter and write a better oscillator code.

Thank you very much for your help!!!

[rsta2]

I'm glad to hear that. You are welcome.

[IbnCocoon]

Hi Rene, happy new year!
This is the updated code working with multicore, now with better filter and oscillator. In the Audio Examples folder you can hear some of its capabilities.
I'm learning a lot with this project. The next features will be patch saving, multi timbrality and a chorus fx.

Thank you for your work and help!!

[rsta2]

Happy new year for you too! Thanks. It' s great to see you making progress. I was listening to your examples and I like it. You are welcome.

[ellisgl]

Do you have a video of this in action?

[IbnCocoon]

Do you have a video of this in action?

Hi, this week I will upload a video.
Although posting audio with fx might be somewhat misleading, I preferred that over dry output because to me it's more important to show the use in real life.
Wave alchemy's Pulse and Tal's chorus are all over the place. I am no a fan of integrating fx in synths, but a simple chorus and delay can make things more easy in real life setup (my goal is to use these Rpi synths live, and keep the expensive gear at home).

[IbnCocoon]

Hi, here are some video examples of the synth.
I'm re-doing the videos because OBS compressed the audio to the point there is a lot of noise. Oscillator and filter demos are currently with Reaper audio. However I'm recording this examples with an onboard motherboard input...the previous audio examples were recorder with an Audient iD, and the sound quality is very different.