Description
TDMA-EFR is a speech coding algorithm standardized as TIA IS-641, adopted by 3GPP for interoperability and reference. It is an Enhanced Full Rate codec designed for Time Division Multiple Access (TDMA) digital cellular systems, such as those based on the IS-136 standard. The codec operates at a net bit rate of 7.4 kbit/s, but with channel coding for error protection, the total rate transmitted over the air interface is higher. It is based on the Algebraic Code Excited Linear Prediction (ACELP) coding paradigm, similar to the GSM EFR and AMR codecs.
The codec works by analyzing 20 ms frames of speech signal (sampled at 8 kHz). For each frame, it extracts parameters representing the vocal tract filter (Linear Predictive Coding coefficients) and the excitation signal (a fixed codebook and an adaptive codebook for pitch). These parameters are quantized, coded, and transmitted. At the receiver, the parameters are decoded and used to synthesize the speech signal through an ACELP synthesis filter. The key to its enhanced quality is the more efficient and robust quantization of these parameters compared to earlier full-rate codecs, resulting in speech quality that is close to wireline toll quality even under moderate channel error conditions.
Its role in the network is as a voice service component within the User Plane. It is implemented in the speech processing units of both mobile stations and network transcoders. The codec defines the precise bitstream format for compressed speech frames, which are then carried over the traffic channels of the TDMA air interface. Specifications like 3GPP TS 26.093 define the codec's bit-exact implementation, test sequences, and compatibility requirements, ensuring consistent voice quality across different manufacturers' equipment.
Purpose & Motivation
TDMA-EFR was created to address the need for significantly improved voice quality in digital TDMA (IS-136) networks, which initially used older, lower-quality speech codecs. The primary problem was that while digital TDMA provided better spectral efficiency and security than analog systems, the voice quality of early full-rate codecs was often perceived as inferior to analog AMPS or wireline calls. This was a competitive disadvantage.
The motivation was to develop a codec that could deliver 'wireline-like' or 'toll-quality' speech within the constrained bit-rate budget of a digital cellular channel. The ACELP technology, which was proving successful in other standards (like GSM EFR), was adapted for the North American TDMA market. TDMA-EFR solved the quality problem by employing more advanced analysis-by-synthesis search procedures and sophisticated quantization methods, making the compressed speech more natural and resilient to bit errors. Its creation was driven by the desire to enhance customer satisfaction and compete effectively with other evolving digital technologies, such as CDMA, which also offered high-quality voice services.
Key Features
- Based on ACELP (Algebraic Code Excited Linear Prediction) speech coding algorithm
- Operates on 20 ms speech frames at a net bit rate of 7.4 kbit/s
- Delivers enhanced voice quality approaching wireline toll quality
- Includes robust channel coding for error protection on radio links
- Defined by TIA standard IS-641 and adopted for 3GPP interoperability
- Provides a standardized test sequence for verification of implementations
Evolution Across Releases
Defining Specifications
| Specification | Title |
|---|---|
| TS 26.093 | 3GPP TS 26.093 |