This year marks the 30th birthday of the MP3 format. Back in 1995, scientists at Fraunhofer IIS in Erlangen chose the file ending .mp3 for MPEG-1 Layer 3 — and unknowingly changed audio consumption forever.
What many don’t realize: broadcasting was one of MP3’s first real-world applications.
Today, codecs like AAC, Opus, and xHE-AAC have evolved far beyond what MP3 offered — but its legacy lives on. So the question remains:
Which audio codec should I use in my workflow today?
Audio codecs are essential in modern media workflows – from digital broadcasting and livestreaming to remote contribution and archiving. But audio codecs differ significantly in performance, latency and efficiency. Selecting the right one depends on your needs: quality, latency, bandwidth, or compatibility.
What is Audio Coding?
Before choosing a codec, it helps to understand how raw audio (like PCM) becomes manageable for transmission and storage. A stereo 48 kHz/16-bit signal generates 1.5 Mbit/s – about 691 MB per hour. Multiply that by a thousand listeners, and the need for compression becomes obvious.
There are two types: lossless and lossy. Which one fits best depends on your use case:
|
Lossless (e.g. FLAC)
|
Lossy (e.g. MP3, AAC)
|
|---|---|
|
No audio information is lost – great for archiving and mastering
|
Uses perceptual models to discard inaudible details for better compression
|
Typical audio data rates for 1 hour of stereo audio:
Uncompressed PCM
FLAC (Lossless)
MP3 / AAC / Opus
Audio Coding Technologies
Different codecs are optimized for different tasks — from speech to music, from low latency to maximum efficiency. Here’s a quick overview of common codec families.
Speech Codecs
Examples: G.711, G.722, SILK
Low bitrate, low latency – optimized for efficient transmission of human speech.
Perceptual Codecs
Examples: MP3, AAC, Vorbis
Use psychoacoustic models to discard sounds that are less likely to be perceived, achieving compression without hurting audio quality.
Predictive Codecs
Examples: ADPCM, SILK, aptX
Work by encoding only the difference between audio samples. This results in lower CPU usage and stable quality
Adaptive Codecs
Examples: Opus, xHE-AAC
Can dynamically adjust bitrate and behavior depending on content and network conditions.
Perceptual Audio Coding
Not all sounds are heard equally, and perceptual audio coding takes full advantage of that fact.
Perceptual coding relies on how we hear: louder sounds mask nearby quiet ones — both in time and in frequency. These masking effects allow encoders to discard audio information that would be inaudible to human ears, saving bits without significantly harming perceived quality.
The most famous codec based on this principle is MPEG-1 Layer 3 — better known as MP3.
It revolutionized the way we consume audio and remains one of the most widespread formats to this day. However, MP3 is now outperformed by newer codecs like Opus and the AAC family.
At the same bitrate, these formats offer better audio quality — or to put it another way: modern codecs achieve comparable quality at lower bitrates than MP3. See PDF
Codec Delay – Why Latency Matters
When choosing a codec, bitrate and audio quality often take center stage — but latency can be just as critical.
In live broadcasts, remote contribution, or intercom systems, even a small delay can disrupt timing, interaction, and listener experience. Some codecs, like Opus or AAC-LD / AAC-ELD, are optimized for low-latency streaming. Others, such as HE-AAC v2, prioritize compression efficiency — but introduce more delay as a trade-off.
|
Codec
|
Delay
|
Efficiency
|
|---|---|---|
|
AAC-ELD
|
Low (~20–32 ms)
|
High
|
|
Opus
|
Very Low (~5–30 ms)
|
High
|
|
Enhanced aptX
|
Very Low (<5 ms)
|
Very High
|
|
FLAC
|
Low (~5–10 ms)
|
Lossless (heavy)
|
Which Codec for Which Application?
The best codec depends on what you’re trying to achieve — low delay, high quality, mobile efficiency, or infrastructure compatibility.
Here’s a practical overview based on common broadcast and streaming scenarios.
|
Use Case
|
Recommended Codecs
|
Why
|
|---|---|---|
|
Remote contribution, intercom, live-link
|
AAC-ELD, Opus, Enhanced aptX
|
Low latency
|
|
Mobile streaming, backup transmitter feeds
|
HE-AAC v2, xHE-AAC, Opus
|
Good quality at low bitrates
|
|
Studio-to-transmitter link (STL), DVB-Backbone
|
MPEG-1 Layer 2, AAC-LC, Enhanced aptX
|
Infrastructure compatibility
|
|
Music streaming, archiving
|
AAC-LC, FLAC, Opus (fullband)
|
High audio quality
|
|
Standard-compliant broadcasting (DAB+, DVB, DRM)
|
MPEG-1 Layer 2, HE-AAC v2, xHE-AAC
|
Required by standards
|
|
Adaptive streaming (HLS, DASH, mobile)
|
xHE-AAC, Opus
|
Dynamic bitrate switching
|
Built-In Quality: No Codec Compromises
MPEG specifies the bitstream format and decoder — but encoder quality can vary.
Our products use the Fraunhofer professional codec libraries for MPEG codecs, including MPEG Layer 2 and 3, as well as the complete AAC family.
This ensures high performance and superior audio quality across all supported workflows.
Isn’t DAB+ just HE-AAC?
DAB+ uses the HE-AAC v2 profile, but it’s not identical to what’s used in typical streaming workflows.
Specific adaptations have been made to meet the transmission and error protection requirements of digital radio systems.
We’ll explore the technical differences in an upcoming article.
Want to Dive Deeper?
Here’s a downloadable summary of the core codecs used in today’s broadcast and streaming environments — all of which are supported across the 2wcom portfolio: