Hacker News Reader

Comment by ricardobeat | Hacker News Reader

Awesome project! I was really hoping to read about the amp/cab simulator and effects running on the Rasperry Pi as the title mentions, but the software side is entirely missing from the story.

Note that from a cursory look at the links, there hasn't been updates since 2018, and the KS project never launched.

View on HN

Sorry if it was short on software details.

It's realtime raspbian. Headless ardour (lua implementation). A mix of guitarix and other amp sims. Proprietary cab sim IRs. Various other effects packages like rkr, ardour-native plugins.

Person speaks to alexa, alexa calls a series of lambdas (basically the not-yet-public API), and sends MQTT messages to the device which is tied to the user's Thingamagig account which is linked to their Alexa account.

https://github.com/raspberrypi/linux/tree/rpi-4.19.y-rt

Let me know if you have any other questions.

Looks like he used Ardour [1] and Guitarix [2].

I am curious about the latency of this setup, having used jack and Ardour in the past for recording. I would be very surprised if this rig yields a latency below a couple dozen milliseconds, which, at least for me, is absolutely unbearable when playing guitar.

Even with a dedicated DSP, effects processors in the early 2000's (like the otherwise excellent Vox TonelabSE) greatly struggled with latency issues.

[1] https://ardour.org/

[2] https://guitarix.org/

I was able to get 48000/256/3 working.

i get 48000Khz sample rage, 256 sample buffer size. what is the the '3'?

Thanks for this. DSPs that accurately simulate hardware are not as trivial as one might initially think. I've done some thinking on this front, and the DSP based approaches that simulate hardware are likely making many shortcuts. Take a zener limiter circuit for example. You could just rail samples with an amplitude comparison, but that's a bad sound. So what I think most limiter DSPs do is apply a shaping filter to that. It's an approximation, but not really capturing the characterization of the circuit. The simplest way to get an accurate response is with SPICE simulation. I'd love to see a DSP that specialized in realtime SPICE simulation. Failing that, you have to sit down and do the math for each circuit yourself to establish the relationships between component values and signals, then code that into your DSP. It's not an unreasonably large amount of work, but judging from the people I think are selling audio software, I would be shocked if anyone is actually doing that.

"please visit our kickstarter campaign launching Tuesday April 21st, 2020"

Reply on news.ycombinator.com