need help: low latency wireless audio transmission for guitar?

Hello! for a guitar effect device we want a setup like this: ADC->(I2S)->MCU (audio processing, simple) -> wireless -> MCU-> (I2S)->DAC Amp.
We need mono, 16bit, 32khz or better, with low latency ( <10mecs). We tried to achieve this with a pair of ESP32 with audio codec chips connected via I2S.
We used the proprietary ESP-NOW protocol, but with no luck, the latency that can be achieved is too high and the wireless protocol doesnt work well in environments with a lot of traffic in the 2.4ghz band (routers, phones, BT etc.).
However, there are affordable guitar tranceiver pairs which can do just that ( low latency, high quality mono audio) without problems. But i cant see the hardware that used, they scraped off the labels from the chips...
There is a simple audio tranceiver module pair available (from China), which uses the nrf24L01 chip, in conjunction with a ATMEL MEGA48 processor (very low spec 8bit) and Wolfson ADC and DAC chips.
And they work just great! How is this even possible? the MEGA48 doesnt even have a I2S interface. the latency is very low and the audio is 16bit with a sample rate >=32khz.
I need advice or even support on how to achieve this, so we can use it in our guitar effect. Can you help ? Any ideas are appreciated,
Cheers!
 
Have a Line 6 Relay G30. Looks to be about the same as wireless mic technology. Not that low cost, however.

Regarding I2S, there doesn't have to be an interface of that type. In the old days there were various non-I2S bus ways of sending digital audio between devices.

May I ask what kind of budget there is for such a device? Is it a one-off, or for production? Why is the particular topology required? Why can't your effects processor or whatever it is do the A/D conversion and then send digital audio to a transmitter? What distance does it need to operate over? What reliability is needed, the same as professional wireless mics? How long does batter life have to be? Any size constraints?

EDIT: Are most of the low cost Chinese units Bluetooth-based? Seems likely. Here's one that claims not to be: https://www.aliexpress.us/item/3256804212158288.html?spm=a2g0o.productlist.main.95.3491ZkHDZkHD0P&algo_pvid=c3168307-fd0e-43bd-a496-84c47271264b&algo_exp_id=c3168307-fd0e-43bd-a496-84c47271264b-47&pdp_npi=4@dis!USD!39.83!28.28!!!39.83!!@210318c916972374804966393e47c4!12000029036948301!sea!US!744813149!&curPageLogUid=y1CxuZjYWJyC
 
Last edited:
I'd use the Sennheiser EWD1 digital wireless system, which has the absolute lowest latency and doesn't run in the clogged up 2.4ghz band. It also sounds the best compared to all other wireless systems under $600.

I run this system with all my basses if I need to be wireless, and it has the best frequency response of most wireless systems in this price range. I have a few transmitter packs to switch basses on the fly between songs if needed, so I don't have to fuss around with cabling. I even have a transmitter pack holder on my upright bass. Its a very versatile system and worth every penny.

It also.has selectable cable capacitance simulation for passive pickup instruments.

In terms of latency, you can notice just a few ms, which is a problem if you're playing alot of 16th notes and rhythm parts. I can't deal with more than 2 or 3 ms.

Thats why I prefer a wired setup if its a critical situation ie. when recording, going direct with a DI.

Of course you need to figure out what frequency band you need based on your location and country having different frequencies.

Not sure if the EWD1 has spdif out, so hopefully that will be the case.
 
Thank you for your valuable input. I have to explain further what we are trying to achieve: there will be a foot controller, which takes the analog instrument input , processes it a bit digitally i(compression mainly) and sends the processed digital audio stream wirelessly to a small receiver device, which is attached to the headstock of the instrument.
So the wireless solution should be embedded into the two devices. the receiver amplifies the digital audio (tiny I2S class D amp) and sends it to a small transducer, which is pressed on to the surface of the headstock. The vibration, when in phase, make the instruments strings resonate, which causes a infinite loop, -> harmonic feedback 🙂
This is can be achieved at any loudness level, because we don't need the guitar amp speakers creating unbearable volume to make the instrument resonate. This allows to create sustain and harmonic feedback like the old masters did: Hendrix, Santana, Page, etc. , but at headphone level.
Latency is critical, although the fed back audio is not heard by the player. It can be felt though, since it makes the instrument vibrate. There is also a timing agorithm that we are working on: the audio is -depending on the note played- more or less out of phase. If the latency time is known and stable, a shift (delay) can be calculated at realtime (using pitch detection) , in order to compensate for that. I created "frankenstein" test set ups for this, using similar items like the Line6 Relay, connecting the the analog audio, etc. it works great as a proof of concept. Now we want to integrate it further, reducing the parts and digital/audio conversions to the minimum. It should become be a consumer product at the end. bluetooth (even low latency APTX) is too slow.
@Markw4 : the product you found is a good example: they have a proprietary protocol going on, but they protect it by simply scaping the labels from the chips, so I cant figure out what they use. The devices work very well, it would be exactly what I need to embed in our product.
If you are old, like me: in the 80s, there was a product called sustainiac Model C, which did the same thing, but with clumsy big transducers and cables running up the neck. We want to do the same thing, but with new technology.
 
nectite,

Putting a mass on the headstock of a guitar will change the sound, possibly substantially.

How much mass can you practically add to the headstock with batteries, a transducer, and a receiver (I presume its the receiver that needs to be near the headstock vibrator)?
 
Necite,
Did you had a look at the company LumenRadio, they are specialized in hard environment wireless transmission.
I know that they sell some modules solutions : here
I don't know the costs of such a solution but I think it is what you are looking for.
A lot of fun with your project!
 
Sustainer system already exist.
Done with the pickups to create the loop.
Wireless adds to much complication.

Home users typically use pedal effects.
So normal 1/4" jack still used.
Just plug it in.
Came from Telephone technology, has worked over hundred years now.
Comes standard with all guitars.
 
Latency down to 20ms, is it enough for your application?
I used the CC8520 about 12 years ago, but I'm sure TI must have more modern chips.

(CC8520 - is a radio transmitter/receiver only, with I2S In/Out. It required also ADC at the input and DAC at the output)
Minimum latency is 768/Fs,so using 48kHz it may be 16ms.
 
Bluetooth 100 ms to 40 ms and can be optimized for 20 ms

Dont work for musical instruments.
Needs to be less than 10 ms

20 ms would drive a player insane

I could get my ancient soundcard using windows 98
down to 8 ms

And got plenty sustain / feedback with desk monitors
and plugins when recording.
 
nectite,

Putting a mass on the headstock of a guitar will change the sound, possibly substantially.

How much mass can you practically add to the headstock with batteries, a transducer, and a receiver (I presume its the receiver that needs to be near the headstock vibrator)?
the transducers we plan to use don't add much mass, they use a different concept, like bone conduction speakers in lightweight headphones. the weight added will be around 50 gramm, depending on the size of the rechargeble battery, which is not critical.
 
Last edited:
Latency down to 20ms, is it enough for your application?
I used the CC8520 about 12 years ago, but I'm sure TI must have more modern chips.

(CC8520 - is a radio transmitter/receiver only, with I2S In/Out. It required also ADC at the input and DAC at the output)
Minimum latency is 768/Fs,so using 48kHz it may be 16ms.
thats a good tip. thank you! I will check it out.
 
Sustainer system already exist.
Done with the pickups to create the loop.
Wireless adds to much complication.

Home users typically use pedal effects.
So normal 1/4" jack still used.
Just plug it in.
Came from Telephone technology, has worked over hundred years now.
Comes standard with all guitars.
right, the pickup feedback systems work great ( I own one, built into a fernades guitar), but you have to modify the guitar to make space for electronics and a battery. Also you have to sacrifice a pickup. Our approach would work with any guitar, just clip it on, like one of these small tuners.
 
@nectite: Have you checked latency technical limitations of the wireless digital processing, whether it's actually feasible to use for your phase-sensitive use case? It seems quite unlikely IMO.
the sound does not have to start in the same moment, that would be impossible. For longer notes, where you want to add that sustain effect, the fed back sound needs to be in phase with the strings, meaning the amplitudes are aligned, although they are shifted in time. A delay of 10msecs is acceptable. For example: for a 440 hz A Note, that would be a shift of 4.4 cycles.
 
What is the latency of LC3? bluetooth 5,3? If good you can just buy a cheap earbuds and connect guitar instead of its microphone. Take line out from its speakerconnetion to the dildo thing
CHEERS!
i did that, but its a one off solution for testing and PoC. Latency is still too high. Also, for a commercial product, you have to register your product for the Bluetooth license ($$$) .
 
  • Like
Reactions: Guerilla
Wish I could help...

I think 32KHz bandwidth is a bit ridiculous for guitar; more like 3.2 KHz would do it. That could be covered by transceiver devices intended to send data - not audio - over short wireless distances.

If you're compressing in the floor pedal box, 16 bits may not be necessary either. What, is the wood of the guitar going to hear discrete-ization distortion in a say, 8 bit, 440 Hz signal that's mostly fully modulated - and complain about the SQ? Methinks the neck/body response will ring the same if the audio is 8, 12, 14, or 16 bits.

Chopping those requirements back may open up new chipset possibilities that other systems - like guitar wireless to amp - cannot consider.

I'm old and tried it in the late 70's. Still have the big clunky bolt on transducer, I probably got from Olson's - the one that was widely advertised in the 60's. Actually drilled a hole in my headstock to accommodate the wood/machine screw fastener.