Electrosmith's Daisy


#1

Does anyone experimented with this platform?


#2

nope, probably no-one, as it's still a kickstarter right?


#4

looking at their page they already reached and passed their goal, so guess it shouldn't take too long before they launch it


#5

Impressive hardware at a great price. Somebody should hack the Axoloti patcher to target this device.


#6

Actually, somebody should port the Axoloti firmware to Daisy, and modify the Patcher if absolutely necessary.


#8

Not personaIly but I ordered them a FV-1 DSP module some time ago. They were fast but especially, answered all my questions...


#9

After read this thread, I have some questions for experts.

Elecrosmith's Daisy have a 24bits 192kHz.

  • To my knowledge, high frequency for DAC help to avoid aliasing, but with anti-aliasing algorithms 48kHz are a good compromise between performance and sound quality, no ?
  • 24 bits precision DAC : For example, Mutable Instrument Braids module use 16 bits resolution DAC. DIY Neutron Sound Orgone Accumulator use internal Teensy DAC with 12 bits resolution. Is 24 bits resolution really useful ? For what kinds of applications ?
  • In FV-1 datasheet, I read that internal DAC have 32768 Hz frequency (depending on crystal used). Too low ? But what is the resolution ?
  • If I want to design a digital VCO, is a DAC with a 48kHz frequency and a 10 or 12 bits resolution sufficient ?
  • Finally, with analog devices, having a good precision is difficult, no ? Or it's a misunderstanding in me ? Why is it standard to have high precision with digital chips ?

I don't understand why teensy board (without audio shield) or stm32 nucleo board are not a good alternative to design digital synthesizer. Should we really add a codec for that ?


#10

Sometimes a Codec is the easiest way to get two channels of analog input and output, and the big numbers just come for free.

Example, Teensy 4 has no DAC, so something is needed. A 16bit stereo DAC (pt8211) adds less that $1.

But yes, synth stuff can be done well at 32k/12 bit.

http://www.spinsemi.com/knowledge_base/arch.html

Interesting device, ADC and DAC are 24 bit.


#11

24 bit vs. 12 bit helps a LOT with dynamic range, so if you do volume scaling in the digital domain, you will hear the difference in quiet passages pretty easily. sine waves at low volume are the worst...


#14

Thanks for yours answers ! I did some tests by reducing bit rate and I can hear noise when the volume is low.
I understand why some modular VCO used 12 bits DAC, certainly because they don't use the entire dynamic range (noise seems to be masked by the signal, the ears are logarithmics detectors).
But, yes, considering the price of a standard audio codec, why want to save money...


#15

I believe they nailed it by coming out of the gate with numerous physical formats: pedal, mini-groovebox, eurack,
but of all those programming options, i hate them all. I like the axoloti interface. I think they should license Axoloti from Johannes!

his interface is better because music instruments are mostly patching together existing modules and then spending time crafting details. axoloti's GUI is great for that. PD makes me ill. I can't afford Max/MSP. I don't want to write code for the whole instrument! Only the parts that require special crafting. That's why I won't bother unless Axoloti gets ported to it and Johannes gets paid for it!


#17

Another point of reference: the original Yamaha DX7 has a completely digital audio chain (probably the first affordable synth to be designed that way), and has a 12 bit DAC (albeit with two extra scaling bits external to the DAC chip itself, so in practice, it's a 14 bit DAC).

The codec used in the Axoloti has 24-bit ADCs and DACs. Although, like all other 24 bit ADC/DACs, the resolution may be 24 bits, but the actual dynamic range is much lower, in the case of the ADAU1961 used in the Axoloti, about 100 dB or so, so just a bit more than 16 bits.


#18

I've been a customer of Electro-Smith a few months. Hardware is very well-made and the people nice to deal with

I feel confident they'll deliver hardware on or close to schedule. This has been in the works for a couple of years now. Software will likely be getting fleshed out for a while after.

Axoloti on Daisy would be great. Don't know where a royalty for Johannes would come from, though. It's already priced down to the bone...

High-bit resolution is important for audio input sampling, and for internal processing. At output, the level can be optimized before it goes to the DAC. So less critical there

FV-1 at 32768 and 16 bit IO is great for 'classic' effects such as reverb & delay. That's what it was built for.


#20

This is kinda derailing the daisy thread but of course you guys are right, @GaryInThailand @ricard @tele_player under certain well designed circumstances 12bit are totally sufficient and might even sound good. But as @lokki describes i would think there are too many unknowns in an universal system like a axoloti. You don’t wanna listen to a -24dBFS sine wave at 12bit. And that’s not even the extreme.


#22

Well, properly dithered, a -24 dBFS sine wave at 12 bit will sound just fine, although the noise floor will be noticeable.

But that's not really my point, rather it is that in these days of 24 bit DACs, 12 bit is often classed as decidedly "lo-fi". But properly managed, it can sound fine, as many older digital instruments prove. I'm not suggesting that they sound better than modern systems, just saying that a 12 bit DAC could sound fine in 1983, and it can in 2020 too if needed.


#25

That's a good summary, though I've reached a diffierent conclusion.

First, Daisy has too many development systems, and none of them is complete. My guess - Daisy needs more SW engineers.

Next, for me, the 1Khz noise is a show-stopper. Seems to me that Daisy was designed by amateurs, or just inadequately tested. Or worse, they decided it didn't matter. Note: running on battery also eliminates the noise. BUT, in my opinion, the Seed should be quiet under all power conditions, without requiring external devices.


#27

I understand, but feel utter terror that such an active and talented programmer is having his head turned by this Daisy floozy....she's not good enough for you Smash, her lack of usb midi support is indicative of her ultimately cold heart......

As long as neither the Akso nor the Axoloti are regularly available i will be looking for alternatives.

wish it weren't so but, yes developers who create Ax or Aks related hardware do seem to be stricken with radio silence, I hope it's not a new viral variant in the codebase


#28

Daisy just isn’t ready, and may never be ready. They did a good job of promoting with Kickstarter, promises, and a nice website.

Electrosmith hasn’t made any comment on the 1khz buzz, which isn’t encouraging.

I wish Paul Stoffregen would make a Teensy4 with a well-implemented codec on board.


#29

Curious to know what it would take to get the Axoloti Patcher up and running on a more up to date microcontroller. As far as I can tell, there hasn't been a collaborative initiative to do so yet.

While the hardware like the MIDI din, stereo jacks, etc. are other matters entirely, I think that making the patcher available on a newer chip alone would boost the Axoloti project to a wider audience who could potentially add those things later. Someone in another thread converted their PreenFM2 since it's the same chip, so I'm curious if that extrapolates to something like the i.MX RT1060, so you could plug-and-play with a Teensy 4.x.

Those on this forum may know that my programming journey literally began with the Axo arriving in the mail, so the idea might be absurd, but the Axoloti has really changed my entire philosophy on hardware in a way that I feel should be taken into a new generation, so I had to ask.

There's so much love and energy put into the objects of this thing, I just think that it's within the realm of possibility to direct that energy towards the platform itself, as a community.


#30

Akso is exactly that, Axoloti on a much faster STM32. Unfortunately, it’s creator, Nic (urklang on this forum) has stopped communicating.