Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 61937-4: Digital Audio Interface for Non-Linear PCM Encoding — Burst Format and Applications
IEC 61937-4:2003 (with Corrigendum 1:2004) is the international standard that specifies the interface for the transport of non-linear PCM encoded digital audio bitstreams. It defines the burst-payload format used to carry compressed audio formats — including Dolby Digital (AC-3), DTS, MPEG-1/2 Audio, and ATRAC — over the IEC 60958 (S/PDIF or AES/EBU) digital audio interface. This article provides an in-depth look at the burst structure, data rate considerations, and engineering implications for consumer and professional audio systems.
📌 Standard Context: IEC 61937 is a multi-part standard. Part 4 specifically addresses the burst information for the MPEG audio stream format. The standard as a whole enables legacy PCM-based digital interconnects to carry modern compressed multichannel audio without hardware modification.
🔧 Burst-Payload Format and Data Transport Mechanism
The fundamental innovation of IEC 61937 is the burst-payload concept. Unlike linear PCM audio, which uses every audio frame for sample data, compressed audio formats produce variable-length data at irregular intervals. The burst-payload mechanism packs this compressed data into short, high-intensity bursts that fit within the IEC 60958 frame structure.
Each burst consists of the following elements:
- Pa Preamble: A 16-bit synchronization word (0xF872) that signals the start of a burst.
- Burst Information (1-3 audio frames): Contains the data type (indicating the audio coding format), data length, and repetition period.
- Burst Payload: The actual compressed audio data, packed into consecutive IEC 60958 subframes.
- Stuffing: Zero-padding to fill any remaining time slots before the next burst.
⚠️ Critical Requirement: The burst repetition period must exactly match the original audio frame rate to maintain real-time playback. For example, Dolby Digital (AC-3) at 48 kHz has an audio frame duration of 32 ms (1536 samples). The IEC 61937 burst containing one AC-3 frame must repeat every 32 ms, regardless of the actual compressed data size.
| Data Type (Hex) | Audio Format | Max Channels | Typical Bitrate |
|---|---|---|---|
| 0x01 | MPEG-1 Audio (Layer I/II/III) | 2 | 32-448 kbps |
| 0x02 | MPEG-2 Audio (LSF) | 5.1 | 8-160 kbps |
| 0x03 | MPEG-2 AAC | 5.1 | Variable |
| 0x04 | Dolby Digital (AC-3) | 5.1 | 32-640 kbps |
| 0x05 | DTS (Digital Theater Systems) | 5.1-7.1 | 32-1536 kbps |
| 0x0A | ATRAC/ATRAC2/ATRAC3 | 2 | 66-292 kbps |
| 0x0C | DTS-HD High Resolution | 7.1 | Up to 6 Mbps |
🎚️ Data Rate Management and Timing Constraints
One of the most critical engineering challenges addressed by IEC 61937 is managing the mismatch between the constant bitrate of the IEC 60958 transport (fixed at the audio sampling rate) and the variable bitrate of compressed audio.
The key timing parameters for a compliant implementation are:
- Burst Repetition Period: Must equal the encoded audio frame duration. For a given sampling rate Fs and audio frame size N samples, the period is N / Fs seconds.
- Minimum Delay (Burst-to-PCM): The time from receiving an IEC 60958 frame containing burst data to presenting decoded audio samples. This is typically 1-3 audio frame periods.
- Buffer Management: The decoder must buffer incoming burst data to smooth out the instantaneous high data rate (the burst) into a continuous audio output.
💡 Engineering Insight: The burst-payload mechanism creates a peak-to-average ratio challenge. During a burst, the data rate over IEC 60958 can be 4-6 times the average compressed bitrate. For example, a 448 kbps Dolby Digital stream requires bursts at approximately 2.5 Mbps over the 3.072 Mbps IEC 60958 channel. Designers must ensure that downstream buffers and decoders can handle this bursty traffic without data loss or underflow.
| Format | Frame Size (samples) | Frame Duration (ms) | Burst Size (bytes) | Peak Burst Rate (Mbps) |
|---|---|---|---|---|
| Dolby Digital (AC-3) | 1536 | 32.0 | 1792 @ 448 kbps | 2.46 |
| MPEG-1 Layer II | 1152 | 24.0 | 1044 @ 384 kbps | 2.09 |
| DTS Core | 512 | 10.67 | 2048 @ 1536 kbps | 2.46 |
| MPEG-2 AAC | 1024 | 21.33 | Variable | Variable |
🏠 Practical Implementation and System Design
IEC 61937-4 has been fundamental to the success of home theater and multichannel audio systems. Here are the key engineering considerations for implementing a compliant interface:
Consumer vs. Professional Formats
While IEC 60958 is used in both consumer (S/PDIF — RCA/phono or Toslink optical) and professional (AES/EBU — XLR) domains, IEC 61937 operates identically on both. The key difference is the channel status bits: consumer mode uses different bits to signal non-PCM audio versus professional mode, and the decoder must correctly interpret these to switch between PCM and burst modes.
Latency Considerations
For applications like live broadcast or gaming, the burst-payload processing adds measurable latency. Each compressed frame must be fully received before decoding can begin (codec-dependent latency), and the IEC 61937 burst de-packetization adds a further 1-2 ms. In critical low-latency applications, this may necessitate alternative transport such as I2S or HDMI ARC/eARC.
HDMI and Multi-Channel Evolution
With the rise of HDMI (which inherently supports multichannel audio), IEC 61937’s importance has shifted. However, the standard remains essential for backward compatibility, optical/coaxial S/PDIF connections, and many professional audio routers that still use AES/EBU as their backbone. The burst format also serves as the encapsulation layer for HDMI Audio Return Channel (ARC) in many implementations.
🚨 Common Pitfall: A frequent design error in IEC 61937 receivers is inadequate FIFO buffer sizing. Because compressed audio bitrates can vary dramatically (DTS 1536 kbps vs. Dolby Digital 192 kbps), a fixed buffer strategy may cause underflow or overflow when switching between formats. The recommended approach is a dynamically allocated buffer with at least 4096 bytes of depth plus a watermark-based flow control mechanism.
❓ Frequently Asked Questions
Q: Is IEC 61937 compatible with all IEC 60958 receivers?
A: No. The IEC 60958 receiver must specifically support non-PCM burst mode. Many early S/PDIF receivers would interpret burst data as PCM clicks and pops. Modern receivers use the channel status bits to detect non-PCM mode and automatically switch decoding paths. The standard provides a backward-compatible mechanism: receivers that do not support the burst format simply produce white noise, indicating the need for a compatible decoder.
Q: Can IEC 61937 carry lossless compressed audio formats?
A: Yes, with limitations. Formats like DTS-HD Master Audio and Dolby TrueHD use IEC 61937 burst framing when transported over S/PDIF at their core bitrate. However, the lossless extensions of these formats often exceed the 3.072 Mbps IEC 60958 channel capacity. For full lossless multichannel transport, HDMI is the preferred interface. Some proprietary extensions (e.g., DTS 96/24) fit within the burst constraint.
Q: What is the maximum compressed bitrate that IEC 61937 can support?
A: The theoretical maximum is approximately 3.072 Mbps, which is the maximum data rate of a 48 kHz stereo IEC 60958 link using 24-bit samples. In practice, the overhead of burst headers, preambles, and stuffing reduces the usable payload bitrate to about 2.5-2.8 Mbps. This comfortably accommodates Dolby Digital (max 640 kbps) and DTS (max 1536 kbps) but limits higher-rate formats like DTS-HD MA (up to 24.5 Mbps).
Q: How does the receiver detect the start of a burst?
A: The receiver continuously monitors the IEC 60958 audio frame data for the Pa preamble pattern (0xF872). Once detected, it reads the subsequent burst information fields to determine the data type and payload length. The receiver then extracts the exact payload bytes from the following subframes. Error detection relies on the burst information fields and the CRC protection inherent in the compressed audio format itself.
