Traffic Frame (Full Rate Speech Traffic Channel)

Description

A traffic frame, as defined in 3GPP specifications, represents the fundamental unit of information transport on the full rate speech traffic channel (TCH/FS) in GSM and its evolved packet-switched successors. Each traffic frame consists of exactly 260 information bits, which are derived from the speech encoder output and subsequently subjected to channel coding, interleaving, and modulation before transmission over the air interface. The 260-bit block size is intrinsically linked to the GSM full rate speech codec (GSM 06.10, originally specified in GSM 05.03), which produces 260 bits every 20 milliseconds, corresponding to a gross bit rate of 13 kbit/s for the speech information itself.

The structure of the traffic frame is critical to understanding the entire GSM speech processing chain. The 260 bits are classified into three sensitivity classes based on their perceptual importance: Class 1a (50 bits) are the most significant and are protected by a 3-bit cyclic redundancy check (CRC) plus convolutional coding; Class 1b (132 bits) are also convolutionally coded but without CRC protection; and Class 2 (78 bits) are transmitted without any error protection. This classification ensures that the most perceptually critical bits receive the strongest protection against channel errors, thereby maintaining acceptable speech quality even under degraded radio conditions.

In the physical layer processing chain, the 260-bit traffic frame undergoes block coding (for Class 1a bits), convolutional encoding (rate 1/2 for Class 1 bits), and interleaving across eight consecutive TDMA time slots. The interleaving spreads the bits of a single traffic frame over 40 milliseconds, which provides time diversity and mitigates the effects of fast fading and burst errors. After interleaving, the bits are mapped to the GSM burst structure using Gaussian Minimum Shift Keying (GMSK) modulation and transmitted on the assigned physical channel.

The traffic frame concept extends beyond the original GSM full rate channel. In later 3GPP releases, particularly from Rel-17 onwards, the definition was broadened to encompass the Wideband Full Rate (TCH/WFS) and Wideband Half Rate (TCH/WHS) speech traffic channels, where the traffic frame may contain between 132 and 397 information bits depending on the codec mode and channel type. This reflects the evolution towards adaptive multi-rate (AMR) and wideband AMR codecs, which produce variable-size frames that are dynamically selected based on channel conditions and network configuration.

In the broader 3GPP architecture, traffic frames are relevant to the circuit-switched domain, particularly in the context of the A interface (between BSS and MSC), the Abis interface (between BTS and BSC), and the Iu-CS interface in UTRAN. The specifications 26.093 and 26.193 define the adaptive multi-rate speech codec frame formats, while 46.031, 46.041, and 46.081 define the full rate, half rate, and enhanced full rate speech processing functions respectively. Understanding traffic frames is essential for engineers working on speech codec implementation, channel coding design, and radio resource management in GSM/EDGE and legacy UTRAN networks.

Purpose & Motivation

The traffic frame concept was created to provide a standardized, fixed-size unit of speech information that could be reliably processed through the entire GSM transmission chain, from speech encoding to radio transmission. By defining a precise 260-bit block corresponding to 20 milliseconds of speech, the standard ensures interoperability between equipment from different vendors and enables consistent channel coding and interleaving strategies. This fixed framing structure was essential in the early 1990s when GSM was being developed, as it allowed deterministic timing and processing requirements to be established for both mobile stations and base station equipment.

The primary problem the traffic frame solves is the need to protect perceptually important speech bits against transmission errors while maintaining low latency and efficient use of the radio spectrum. Before the traffic frame concept was standardized, speech transmission over radio links suffered from unacceptable quality degradation due to burst errors and fading. The classification of bits within the traffic frame into sensitivity classes, combined with convolutional coding and interleaving, directly addresses this problem by providing unequal error protection tailored to the perceptual significance of each bit.

Historically, the traffic frame definition has remained remarkably stable since its introduction, reflecting the maturity and robustness of the original GSM design. The only significant revision occurred in Rel-17, when the definition was extended to cover wideband speech channels with variable frame sizes (132 to 397 bits), accommodating the introduction of wideband AMR codecs that offer improved speech quality. This evolution demonstrates how the original traffic frame concept has been adapted to support newer codec technologies while preserving backward compatibility with existing GSM infrastructure.