Probably you wonder how Made in China iPod Nano clones are sold at $10. They work in the way people used to build clap switches with 555 IC once. IC and sometimes microcontroller with manufacturer distributed hex files or binary files. It was proprietary mixed with community effort. Arduino and ESP32 are indeed costly. But they have made programming microcontrollers completely open source for the programming part. However, learning how to build a clap switch or blink an LED with 555 IC has value in learning the basics plus they do have production value since learning hardware is the only way for mass production in certain segments.
Decoding and Playing MP3 Files is Easy
There are $1 MP3 decoder boards with 2W amplifier IC such as the GPD2846A module which can easily create an MP3 player by connecting a memory card, speaker and battery. This particular MP3 decoder board utilises the GPD2856A MP3 decoder IC and has an onboard 2-watt L/R mixed mono audio amplifier IC (8002). These ICs are designed in China and their datasheets describe their available functions in China-English language. GPD2846A IC is not only the one of its kind. There are many other similar ICs available. There are JQ6500 MP3 modules, DFRobot’s DFPlayer Mini etc.
Commonly we can optionally add a microcontroller to control it. Also, there is a thing named MP3stick. It uses the microcontroller Atmel AVR ATmega128. MP3 decoding is done by a VLSI VS1011b decoder IC. It includes an MMC/SD card reader which stores MP3 files, playlist files and skin files. It is designed to use a LiIo/LiPo battery for which there is a charger circuit, based on MAX1811. There is a Nokia colour LCD.
---
If you read this webpage of Microchip:
1 | https://www.microchip.com/en-us/solutions/consumer/audio-and-speech/pic32-digital-audio/pic32-audio-decoders |
You’ll understand the proprietary business model. License fees and who purchased their license are written here:
1 2 | https://www.via-la.com/licensing-2/aac/ https://www.via-la.com/licensing-2/aac/license-fees/ |
People who create MP3 players for different reasons, usually use different amplifier ICs to create the MP3 player. For example, this is a circuit diagram:

This is extremely hard way but resulting products are often serviceable at lower cost.
How They Add Bluetooth, Voice etc
They add more ICs or circuits to add the functionalities. As for voice feedback (Connected, Disconnected etc announcements), custom MCUs are sold. In the same way, they add spectrum analyzers and other features.
There is a difference in approach in real-life manufacturing planning since end price is a factor. Usually, audio device manufacturers give more importance to the amplifier and speakers.
That is How They Sound Typical
Yes. These ICs easily produce 8-bit to 16-bit audio. With IC-based amplifiers, they sound not like the CDs. It is not the MP3 codec you can blame. But this way gives more control on the sound amplification in a cost-effective manner.
Many of the readers read articles like this one Best Earphone/IEM, DAC & Music Player for Samsung S24 Ultra and then plans to build ESP32 based or Raspberry Pi based 32 bit high-end FLAC players. That is good for building one unit as a hobby using SBCs. For mass sales, one has to go through the hard way of hardware. Probably most of the manufacturers avoid both ESP32 and Raspberry Pi for these purposes.
Stop buying too costly modules with latest DACs, instead give importance to the hardware. Even ES9038Q2M chip costs mere $10. A module or board with ES9038Q2M chip will cost near $100. It is super complicated to think them from programming point to view since they are proprietary. It is easier to use them with SBCs such as Raspberry Pi.