True single-chip MP3 encoder/decoder

According to Flint, the Micronas MAS3587F is the first true single-chip MP3 encoder/decoder, bringing an unprecedented level of integration and performance to any speech/audio design.

According to Flint, the Micronas MAS3587F is the first true single-chip MP3 encoder/decoder, bringing an unprecedented level of integration and performance to any speech/audio design.

Key opportunities include all memory-based recording/playback systems such as portable MP3 recorder/players and in car systems, plus a wealth of new applications in kiosks, recorded public address systems, in-store audio equipment, telephones and speech-enabled equipment (lifts, security, access control).

While MP3 encoding and decoding are possible with generic DSPs and appropriate software, the MAS3587F provides optimised silicon in an application-specific device, At the heart of the IC is a RISC-based DSP engine with embedded RAM and ROM, which provides audio baseband features and MP3 encoding/decoding functionality, together with optional software downloads and flexible digital interfaces for serial and S/PDIF audio data input and output.

The DSP core has access to two RAM banks, and all RAM addresses can be accessed in 20 or 16bit mode via I2C bus.

For fast access of internal DSP states the processor core has an address space of 256 data registers, which can be accessed by the device's I2C bus.

Firmware implemented in the program ROM provides MPEG 1/2 Layer 3 encoding and decoding of MPEG 1/2 Layer 2 and MPEG 1/2 Layer 3.

Selection of either the MPEG audio encoder or decoder is activated on start-up.

The MPEG decoder provides an automatic standard detection mode.

If all MPEG audio decoders are selected, the Layer 2 or Layer 3 bitstream is recognised and decoded automatically.

For general control purposes, the operation system provides a set of I2C instructions that give access to internal DSP registers and memory areas.

An auxiliary digital volume control and mixer matrix is applied to the digital stereo audio data.

This matrix is capable of performing the balance control and a simple type of stereo basewidth enhancement.

Additional functionality is achieved via downloadable software, for example the Micronas SC4 encoder/decoder.

SC4 is an exclusive Micronas speech codec technology based on ADPCM.

The codec can be downloaded to the MAS 3587F to allow high quality speech recording and playing back at various sampling rates.

Using the built-in Fast Program Download software, up to 4K x 20bit of program code and additionally up to 4K x 20bit of coefficients can be downloaded into the internal RAM - switchable from data memory to program memory.

A sophisticated set of audio convertors and sound features is incorporated in the audio codec to comply with various kinds of operating environments up to high-end equipment.

In encoding mode, audio data is input via the integrated A/D convertor, serial PCM, or S/PDIF interface.

The compressed digital data stream is sent via the parallel interface.

In decoding mode, compressed digital data streams are accepted in the parallel or serial format.

The audio data stream is output via a high-quality D/A convertor.

A digital output is also provided catering for both serial PCM and S/PDIF formats.

A pair of A/D convertors is provided for recording or loop-through purposes.

In addition, a microphone amplifier including voltage supply function for an electret type microphone has been integrated.

The several baseband functions are applied to the digital audio signal immediately before D/A conversion, including standard bass, treble, and loudness, the Micronas Dynamic Bass system (MDB), automatic volume control, balance and volume.

MDB extends frequency range of loudspeakers or headphones below the cutoff frequency of the speakers, both by dynamically amplifying the low frequency bass signals, and also adding harmonics of lower frequency components to give the illusion that these bass notes themselves are being heard.

Parametric implementation of the MDB allows it to be customised to create different bass effects adapt to various loudspeaker characteristics.

A pair of Micronas multibit sigma-delta D/A convertors is used to convert the audio data with high linearity and a superior signal-to-noise performance.

Internal lowpass filters are included to attenuate high-frequency noise caused by noise-shaping.

A single crystal-controlled clock drives the MAS3587F with a frequency of 18.432MHz.

Other reference clocks can also be used.

For compressed audio data reception, the IC may act either as the clock master (demand mode) or as a slave (broadcast mode) under software control.

In both modes, the output of the clock synthesiser depends on the sample rate of the decoded data stream.

The DSP clock has a separate divider, which for power conservation is set to the lowest acceptable rate of the synthesiser clock that allows the processor core to perform all tasks.

The two embedded DC/DC convertors of the MAS3587F generate optimum power supply voltages for the DSP core and also for state-of-the art flash memories that typically require 2.7 to 3.3V supply.

High integration of the MAS3587F means only a simple controller and the external flash memories to complete a full portable audio recorder device.

By means of the embedded A/D convertor, the MAS 3587F supports both speech and FM radio quality audio encoding.

CD-quality encoding/decoding is achieved by using digital inputs and embedded D/A-conversion.

Flint supplies the MAS3587F in 64-pin plastic low-profile QFP packaging.

Back to top