Introducing DSPi | A powerful, user friendly and open source DSP for less than a cup of coffee

[5th element]

Because they are simple to create (just two clicks in REW with an existing measurement), and they produce great results for room and speaker correction. Better than any PEQ generation in my tests

Yes but you can do the same with IIR filters with much less processor requirements and no latency.

 

[Weeb Labs]

Major efficiency enhancement in the most recent build. Bypasses are now handled at both channel and band levels, along with bypass flags replacing inefficient coefficient checks.

As a result, baseline core 0 utilization has been drastically reduced to just 4% with all bands bypassed.

This will make its way to the repo once it has been thoroughly tested.

 

[andrewskaterrr]

Stupid question but is there a possibility for a 2xRCA IN, 2xRCA OUT type configuration? Basically like creating your own MiniDSP 2x4 HD. Example, I'd love to be able to use this to time align and EQ my 2 subs that are hooked up to my WiiM Amp's single sub out.

 

[BR52]

Hi, I guess we are now in the phase of defining a project. I see here a lot of aspects, for example:
Price for the HW, ability to build the HW! (skills of the DIYer) How many channels, sampling rate, general perfomance, can it be a real commercial project like a HAT, and so on?
I see here a great skilled SW developer in a forum with people with a very diverse background.
Depending on the intention, I highly recommend a poll to find out the path to go.
A well-done definition in the beginning can avoid a lot of frustations.
Again, depending on the intentions.

 

[DrCWO]

Why do people think they need FIR filters?

... because crossover networks without FIR always exhibit a measurable and audible error in the group delay.

The diagrams show the step response of a "normal" crossover and one using FIR. Image two has a small amplitude error (I didn't have another example handy), hence the slight pre- and post-ringing. And yes, this is audible. I offer FIR room correction filters that can be used in Roon for example. None of my customers ever will get back to uncorrected nor to IIR correction.

 

[FraPia-62]

Fantastic project, thank you for sharing!
A quite interesting use case for this DSPi could be in a high quality monitor loudspeaker. The device could be installed inside the loudspeaker cabinet and implement x-over, mass breakpoint equalisation of the compression driver (if used), 6th order alignment for the woofer (again, if used) and fine eq of the overall frequency response, in a similar way you find in many JBL, Neumann or Genelec monitors.
For such application however it would be necessary to implement an analog input (balanced line amplifier+ADC, rather easy, but it's extra hardware), an AES3 (SPDIF) digital input and possibly also AES67/Ravenna (again, extra HW and SW). Device control via USB would be impractical for a loudspeaker, AES70? I understand that it is a lot of work and I'm sad that I cannot help since I cannot write code, unfortunately.

 

[Kingsnake]

As simple as this?

 

[Weeb Labs]

As simple as this?

View attachment 509664

Yes. That’s all you need.

 

[Barrelhouse Solly]

I'm excited. I have, uh, several Pis, ranging from Zero to 5 with the exception of 1. I've been doing Pi audio for over 10 years and have long passed the point where my wife can stand listening to me talk about it. I've implemented CamillaDSP on Moode and PiCorePlayer. I just ordered my first Pico. It has headers. Would a breakout board make the hardware implementation easier?

 

[MCH]

When the project is more advanced and the definitive I/O defined, it will be worth, in my opinion, leaving the pico behind and designing a custom board for the RP2040/2350, with footprints for connectors and a proper usb-c. Shouldn't be difficult, an advantage vs other raspberry pi based projects, that force you to live with the board as is.

 

[Weeb Labs]

When the project is more advanced and the definitive I/O defined, it will be worth, in my opinion, leaving the pico behind and designing a custom board for the RP2040/2350, with footprints for connectors and a proper usb-c. Shouldn't be difficult, an advantage vs other raspberry pi based projects, that force you to live with the board as is.

I completely agree. A custom board with the appropriate connectors, affordable DACs and a couple of ASRCs would be a great idea when the project is more mature. At the same time, feature creep is something that I want to avoid. There is a plethora of important functionality that still needs to be implemented but at the same time, I don't want to fall down the hole of trying to make it "everything to everyone".

 

[awdelft]

I completely agree. A custom board with the appropriate connectors, affordable DACs and a couple of ASRCs would be a great idea when the project is more mature. At the same time, feature creep is something that I want to avoid. There is a plethora of important functionality that still needs to be implemented but at the same time, I don't want to fall down the hole of trying to make it "everything to everyone".

Great project.
But when going to a consumer box with dac I doubt you can do better/cheaper than the wondom boxes with adau1701?

 

[Tell]

Great project.
But when going to a consumer box with dac I doubt you can do better/cheaper than the wondom boxes with adau1701?

How's the software experience of that though? I've used a chinese amplifier with built in DSP and the software really is horrible, and from the looks of OPs software it looks like heaven compared to the chinese stuff.

 

[Weeb Labs]

Sigma Studio in conjunction with the ADAU1701 is certainly functional and I use it frequently but it is not at all designed with the end user in mind. The hardware is important but if the software is poor or unnecessarily complex, most people won't use it.

 

[yewneek]

So, with the hardware assembled, you
Use a calibrated microphone to do the measurements with your Windows/mac software
Transpose that via USB to the assembled hardware.
Unplug the assembled hardware, plug in the cables/SPDIF/sub.
Play music.

Correct?

The DSP hardware sits inbetween where and where?

Looks interesting

 

[Kingsnake]

So, with the hardware assembled, you
Use a calibrated microphone to do the measurements with your Windows/mac software
Transpose that via USB to the assembled hardware.
Unplug the assembled hardware, plug in the cables/SPDIF/sub.
Play music.

Correct?

The DSP hardware sits inbetween where and where?

Looks interesting

My intended signal chain is:

PC → DSPi via TOSLINK → DX5 II DAC (speaker PEQ applied) → Cambridge Audio amplifier → Wharfedale Diamond 12.1 loudspeakers
PC → DSPi RCA subwoofer output → BK Electronics XXLS400-FF subwoofer

The missing link would be that my current TV TOSLINK into the DX5 II wont have bass management. The Source has to be a PC from my understanding.

 

[TurtlePaul]

Quick question? Does this appear as a UAC1 audio device (and therefore work with consoles) or UAC2? I now have my Pico2 in hand and am going to solder a motherboard optical header.

 

[Weeb Labs]

Quick question? Does this appear as a UAC1 audio device (and therefore work with consoles) or UAC2? I now have my Pico2 in hand and am going to solder a motherboard optical header.

Yes, it is a UAC1 device. That was one of several reasons for choosing the older protocol.

 

[Weeb Labs]

Absolute legend

Sounds quite nice.
I'm working on ITD at the moment.

 

[eric00886]

@Weeb Labs

This is an excellent project. Which board would be more suitable: the Raspberry Pi Pico 2 or the Pico 2 W?

Thanks,