Audio Codecs for Voice and Instrumentation
Product Specifications
Product Overview
An audio codec is chosen by the interface, the converter performance and what is already inside the package. A voice terminal needs 8 kHz, a programmable gain amplifier and a speaker driver, while an instrument needs 96 kHz, a flat passband and a converter that does not add noise below the signal. The parts here cover both ends, and the difference between them is mostly how much of the analogue chain is on the die.
The TLV320AIC3104 family is the voice and telephony shape. It has six inputs and six outputs, a stereo headphone amplifier, programmable gain amplifiers with up to 59.5 dB of analogue gain and an AGC for the microphone, a 92 dB signal to noise ratio on the ADC at 48 kHz, and it runs from 2.7 V to 3.6 V on the analogue rail with a 1.525 V to 1.95 V core and a 1.1 V to 3.6 V I/O rail. The 3101 adds a speaker amplifier, the 3105 makes every input single ended, and the 3106 brings ten inputs and seven outputs, all in the same 5 by 5 mm QFN and the same register map.
The WM8960 is the portable player shape: 24 bit converters with 98 dB on the DAC and 94 dB on the ADC, a stereo class D driver that puts 1 W into 8 ohms per channel from a 5 V rail at 87 percent efficiency, a capless headphone driver, no charge pump and an on chip PLL, all in a 5 by 5 mm QFN. The ADAU1361 is the low power shape, with stereo 48 kHz record and playback at 14 mW from a 1.8 V analogue supply, which is what a battery powered recorder or a headset needs. Both take I2S and TDM, and both are programmed over I2C.
Key Specifications
| Parameter | Voice and telephony codecs | Portable and low power codecs |
|---|---|---|
| Converters | 24 bit stereo ADC and DAC | 24 bit stereo ADC and DAC |
| Sample rate | 8 kHz to 96 kHz | 8 kHz to 48 kHz |
| Signal to noise | 92 dB on the ADC, 93 dB dynamic range | 98 dB on the DAC, 94 dB on the ADC |
| Analogue gain | Up to 59.5 dB with AGC | Programmable, with ALC and noise gate |
| Speaker driver | Class D or class AB depending on part | Class D, 1 W into 8 ohms at 5 V |
| Supply | 2.7 V to 3.6 V analogue, 1.525 V to 1.95 V core | 2.7 V to 3.6 V analogue, 1.71 V to 3.6 V digital |
| Digital interface | I2S, left and right justified, DSP, TDM | I2S, left and right justified, PCM |
| Package | VQFN-32, 5 by 5 mm | QFN-32, 5 by 5 mm, and WLCSP options |
Included Models and Order Codes
The table lists the order codes most often quoted for intercoms, headsets, portable recorders and instrument front panels. State the sample rate, the number of inputs and outputs, the speaker power, the supply rails and the digital audio format, and the part and its packing are confirmed.
| Part Number | Channels | Sample Rate | Signal to Noise | Drivers | Package |
|---|---|---|---|---|---|
| TLV320AIC3101IRHBR | Stereo codec, 6 inputs, 6 outputs | 8 kHz to 96 kHz | 92 dB ADC, 93 dB dynamic range | Headphone and speaker amplifier | VQFN-32, 5 by 5 mm |
| TLV320AIC3104IRHBR | Stereo codec, 6 inputs, 6 outputs | 8 kHz to 96 kHz | 92 dB ADC, 93 dB dynamic range | Headphone amplifier | VQFN-32, 5 by 5 mm |
| TLV320AIC3105IRHBR | Stereo codec, all inputs single ended | 8 kHz to 96 kHz | 92 dB ADC, 93 dB dynamic range | Headphone amplifier | VQFN-32, 5 by 5 mm |
| TLV320AIC3106IRGZR | Stereo codec, 10 inputs, 7 outputs | 8 kHz to 96 kHz | 92 dB ADC, 93 dB dynamic range | Headphone and line outputs | VQFN-48, 7 by 7 mm |
| WM8960CGEFL/RV | Stereo codec, 3 stereo inputs | 8 kHz to 48 kHz | 98 dB DAC, 94 dB ADC | Class D 1 W per channel, capless headphone | QFN-32, 5 by 5 mm |
| ADAU1361ACPZ-R7 | Stereo codec with SigmaDSP core | 8 kHz to 48 kHz | Not quoted as a single figure, 14 mW record and playback | Headphone amplifier | LFCSP-32, 5 by 5 mm |
Key Features
Converter performance stated separately – the ADC and the DAC are quoted with their own signal to noise figures, because a recorder and a playback path do not need the same number.
Programmable analogue gain – up to 59.5 dB ahead of the converter means a low output microphone does not need an external preamp, and the AGC keeps the level constant.
Speaker driver on the die – 1 W into 8 ohms from a 5 V rail removes the class D amplifier and its filter from the BOM, which is the largest saving on a voice terminal.
Capless headphone output – the WM8960 drives headphones without the large coupling capacitors, which saves board area and removes the pop that a charged capacitor makes at power up.
Flexible digital interface – I2S, left and right justified, DSP and TDM modes let one codec sit on a bus with a DSP, a Bluetooth module and a second codec.
Low power modes – analogue bypass and block level power down keep the current in the microamp range when the product is idle but still listening.
Applications
These codecs are quoted for products that capture or reproduce voice and signal audio, typically where: An audio board is normally quoted with the SUPERB PCB and SUPERB PCBA work we support for the same assembly, so the codec, the clock and the jack protection are handled as one build.
Intercoms and voice terminals – a handset needs a microphone preamp, an AGC, a speaker driver and 8 kHz telephony audio in one part, with the same register map across a product family.
Portable recorders and headsets – a 1.8 V analogue rail, 14 mW of playback power and a capless headphone output are what a battery powered device needs to reach a full day of use.
Instrument front panels – a 96 kHz path with 92 dB of dynamic range is enough to record a sensor tone or a voice annotation without an external converter.
Sourcing and Cross-Reference
Audio codecs are qualified on the analogue path, and the analogue path is where a second source fails. Two parts with the same I2S interface can differ in the microphone bias, in the input impedance, in whether the headphone output needs a coupling capacitor and in the register sequence needed at power up. We quote the input configuration, the supply rails, the driver type and the sample rate range on the confirmation, so a replacement is checked against the acoustic requirements.
The rest of the chain is the crystal or the MCLK source, the decoupling on the analogue rail, the ground split under the codec and the ESD array on the headset jack. We quote the codec with the clock part and the jack protection, and where the board is built in volume the ground and the placement are reviewed before the first lot because a codec that measures 92 dB on a bench will measure 80 dB next to a switching regulator.
Where the audio boards are built in volume, SUPERB Assembly support covers the placement of the QFN package and the wash that the fine pitch and the acoustic ports need.
Q: Can you supply an audio codec that replaces an obsolete Wolfson or TI part?
A: Yes. Send the sample rate, the input and output count, the supply rails and the digital audio format, and the SUPERB Tech sales team will confirm the order code, the packing and the nearest part with the same register interface.
Why Buyers Source Audio Codecs Through SUPERB
- Analogue configuration on the quote: the input type, the microphone bias, the driver and the supply rails are stated per line, so a replacement matches the acoustic requirements as well as the pinout.
- Clock and protection together: the MCLK source, the decoupling and the jack ESD array are quoted with the codec, so the measured performance is repeatable.
- Obsolete Wolfson and TI parts: parts that a product was voiced around are held with date codes, which keeps a headset or an intercom in production while a new board is qualified.
- Board build: the ground split and the placement are reviewed before the first lot, which is where a codec loses the 12 dB it was specified for.
Related Product Families
This page is part of the component range we supply for industrial, telecom, medical and defence builds. Continue with: Analog and mixed signal ICs, DACs and voltage references, 16 bit SAR ADCs, Crystal units and TCXOs, BOM shortage recovery.