Professional Audio On An ESP32

Audiophiles have worked diligently to alert the rest of the world to products with superior sound quality, and to warn us away from expensive gimmicks that have middling features at best. Unfortunately, the downside of most high quality audio equipment is the sticker price. But with some soldering skills and a bit of hardware, you can build your own professional-level audio equipment around an ESP32 and impress almost any dedicated audiophile.

The list of features the tiny picoAUDIO board packs is impressive, starting with a 3.7 watt stereo amplifier and a second dedicated headphone amplifier. It also has all of the I/O you would expect something based on an ESP32 to have, such as I2S stereo DAC, an I2S microphone input, I2C GPIO extenders and, of course, a built-in MicroSD card reader. The audio quality is impressive too, and the project page has some MP3 files of audio recorded using this device that are worth listening to.

Whether you want the highest sound quality for your headphones while you listen to music, or you need a pocket-sized audio recording device, this might be the way to go. The project files are all available so you can build this from the ground up as well. Once you have that knocked out, you can move on to building your own speakers.

53 thoughts on “Professional Audio On An ESP32

    1. I don’t know why anyone would make something like this when amazon sells nice sounding bluetooth speakers with an internal battery charger. They also connect up to phones or whatever and sound great. I must be missing something here. I see audio boards for raspberry pi’s and beagle bones and they cost a lot. by the time you buy the cpu board the audio board the power supply get all the sw working oh yea and a case and speakers the cost is sky high. And at the end of the day all you have a cheap sounding piece of crap.

      1. If this makes no sense to you, then why do u come to this post/page any way? jst close the page, your problem is solved. But our hunt to do most of the things ourselves continues invariant of the cost.

      2. Seems kind of obvious, but the attraction here is to make your own devices, not simply a bluetooth speaker (something this would hardly even do). You can use this as a brain for any musical device. Off the top of my head, a musical horn, musical doorbell, 2 way audio communication, guitar tuner, and the list goes on and on. Honestly, I think you’re looking at the wrong website if that stuff isn’t immediately obvious to you.

      3. Besides the previous and other obvious responses, how about doing things like adding audio I/O to a motion detector or camera module, for security communication or to have sound follow you around the house when moving from room to room, or adding custom intercom capabilities to an existing setup inside a boat where installing a closed, pre-determined and programmed intercom system is impractical. Doing things like this yourself can be far more rewarding, not to mention customized to your needed application, cheaper (especially in the marine environment where anything with the prefix “marine” carries a price tag that is 10x what it would be for any other environment. The possibilities are endless, unless you’re satisfied with allowing technology based on the limitations of mass-sales to dictate your life choices.

  1. “Audiophiles have worked diligently to alert the rest of the world to products with superior sound quality, and to warn us away from expensive gimmicks that have middling features at best.”

    What? Audiophiles work diligently to convince us that digital audio signals need to be “reclocked”, that CD players need special isolation feet and spindles made out of a pound of metal, that you need to spend $500 on speaker cables, etc.

    And that we need to spend $500 on special DACs, when it turns out that many of them can’t do better than the $10 Apple USB-C dongle.

    1. But given the price tag of $45 for this device totally follows the audiophiles philosophy. Although it might sound “cheap” – it’s more than five times the price you pay for a combination of uda1334a board and class d amp at your chinese web shop of choice.

      1. That’s somewhat unfair. Chinese stuff is obviously cheaper, but this isn’t a bad price for a small run of a hobby product. I think if it were ‘Audiophile Grade’ it’d have to be $400.

      2. It’s definitely not cheap, but it’s also not unreasonable. Looking at Adafruit as a comparison;

        UDA1344 codec: $6.95
        Stereo AMP: $8.95
        MEMS Mic: $6.95
        SD Card Reader: $7.50
        IO Expander: $5.95 (Sparkfun)
        Headphone Amp: $9.95 (AK74 – Can’t find anyone else making them)


        And it will look something like this:

      1. You can go a lot higher than that. Several speakers in the 40 grand a pair range going up into the hundreds of thousands. Of course you can spend tens of thousands on a wrist watch also.

      1. It’s a add-on board for a TinyPICO ESP32 – not something you go to a high end audio store and pay through your nose at to connect to your home entertainment system – keep some perspective ;)

        24bit / 96kHz Audio DAC
        3.7W Stereo Class D Amplifier with Stereo Terminals
        25mW Headphone Amplifier for 16/32Ω Headphones with 3.5mm Audio Jack
        Individually selectable gain for each amplifier
        24bit / 44.1kHz MEM Microphone
        microSD Flash Card Reader
        8 Port GPIO Expander for Buttons and LEDs
        5 Builtin Test Buttons
        Break Out Headers for Power, SPI, I2C and GPIO Expansion

        I mean, if you’re happy playing wavs through the ESP32’s onboard DAC – more power to you!

        1. The MEMS microphone could be a lot more useful if it was stereo. The 3.7W Class D amp isn’t going to be useful with anything other than desktop PC speakers or obnoxiously efficient speakers like K-horns.

          Not quite up to the professional audio label but a good Swiss Army Knife for digital audio playback at a nice price.

    1. Read the spec sheet for the DAC, it can do 24-bit / 96kHz and has a dynamic range of over 115dB with less than 0.2dB ripple and -85dB THD+N/S, that’s probably good enough to be called professional if the rest of the supporting circuitry is well made.

      1. Yeah, but audio is one of those devil-in-the-details things. 85 dB S/N is good, but not great, and that’s the chip under optimal conditions. Put a noisy microcontroller next-door, add a bad switching power regulator, etc, and you can really mess it up.

        I’m not at all suggesting that this project _has_ messed it up. Just that the 85 dB is a component, not a system, value.

        1. What’s the S/N, dynamic range, and frequency response at your headphones or speakers? That is ultimately what matters. It’s why I continue to hang onto and fix my sansa clip+ as needed, as it was surprisingly good across the board. Of course like all good products it is no longer produced.

      2. dynamic range of over 115dB says who?
        signal to noise is 98dB, distortion is -85dB and this is all A-weighted!
        then it has to go trough an amplifier that also adds noise an therefore removes dynamic
        So more likely something like 90dB real dynamic range best case!

        thats almost half as good as the theoretical 96.33 dB from a CD ( the disk thing from 1982)

    1. I have some reservations on the ESP daughtercard.
      – seems to be full of ceramic caps. I don’t want them in the audio path as they are microphonic. Hard to tell if the audio signal is DC coupling or not.
      – Typical daughtercard don’t have a good record of good signal integrity as they don’t in general have a good ground/signal ratio for return path and loop area especially on a 0.1″ pitch header.
      – You are referencing analog audio on a board with suboptimal ground connection from a noisy digital motherboard. It gets tricky when you try to hook up external power and use external amplifiers. I would rather see the ESP + amplifier on a single 4 layer PCB done correctly.

      The Chinese PCB layout on the other hand is actually quite reasonable from what I can see. So they might have someone that at least follows some app notes.

    2. I actually have one of these boards, albeit a slightly better manufactured one. Right now, it’s hooked up to my PC via the line-in and powering a pair of Polk RT25i’s. Amazingly good sound for how damn cheap these boards are!

      Just make sure to use a decent power supply, it’s great with the 65W Dell laptop supply I’m using now, not so great with the 12V, 2A adapters that are often recommended.

  2. The DAC says ~100dB signal to thats not bad but not “professional”
    its also A-weighted!
    And THD+N is only -90dB

    that not super bad but also noting special.
    Also this is best case from the DAC i highly dough that the system performance comes close to this.

    Digital audio is good enough for consumers since 1983 and this bord dose nothing better then some CD players from this time

  3. yeah some of those numbers suggest that the net performance will be less than “professional”, especially for the higher demands of recording. The good news though is that maybe someone will be inspired enough by the potential to create a better performing board.

    In the meantime, this board is still a great platform for audio experimentation and many applications. Not bad for something that runs on a $5 wifi chip.

    1. Actually I have to walk the above back a little. The thing doesn’t include a decent A/D! So it’s pretty useless for recording.

      For me, the ideal board would drop the speaker amp, and have a decent analog line in plus a respectable A/D. Good analog mic preamps would be a bonus but tricky to include on a low-cost single board. Then you’d have a platform for a respectable field record/playback solution.

  4. Maybe a dumb question, bus is anyone still listening to music from storage?
    I would, however, be very interested in USB “sound card” with 5.1 wireless speakers, all running on ESP32-S2 and ESP-NOW wireless protocol. Now that would have some real world practicality, as such systems are still rare and quite expensive (Sonos offers one at $2000+).

    1. I found this because I am doing this exact thing. I want to make wireless guitar and mic esp32 boards. I already use esp-now to do wireless midi and 1ms latency ( Now time for analog audio!

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.