AC-3 (Dolby Digital) and E-AC-3 (Dolby Digital Plus)

Published: 2026-06-05 · Reading time: 14 min read · Author: Nikolay Sapunov, CEO at Fora Soft

Why this matters

If you run an OTT service, an Internet-TV app, or any product that streams to TVs and soundbars, you cannot avoid the Dolby Digital family — it is the audio format the hardware in the room actually understands. This article is written for a product manager, founder, or operations lead with no audio background: by the end you will understand what AC-3 and E-AC-3 are, how DD+ carries Dolby Atmos without breaking older devices, what bitrates the major platforms use in 2026, and the common mistakes that corrupt a Dolby stream in a delivery pipeline. Every number traces back to the controlling standard — ETSI TS 102 366 and ATSC A/52 — or to the platform's own published figures, not a secondhand blog.

The codec that put surround sound in the living room

Before 1991, multichannel sound was an analog trick. Cinemas painted optical stripes on the edge of film; the format that mattered was a matrix that folded four channels into two and unfolded them at the projector. There was no clean way to store five separate, full-quality channels and a deep bass channel on consumer media, because the bits did not fit.

Dolby Laboratories solved the storage problem in 1991 with a codec called AC-3 — "Audio Coding 3", the third in a line of Dolby coding systems — and marketed it as Dolby Digital. AC-3 was the first widely deployed audio codec built on a mathematical tool called the Modified Discrete Cosine Transform, or MDCT, which converts a slice of sound into a set of frequency ingredients and then spends bits only where the ear will notice (the same family of math later used by MP3 and AAC, covered in how audio compression works). That efficiency is what let five full channels plus bass fit on a DVD and inside a broadcast signal.

The thing to hold onto for the rest of this article: AC-3 is the original 5.1 surround codec, and E-AC-3 is the same idea, extended. Everything else is detail layered on those two facts.

Figure 1. Thirty-five years of one codec family. AC-3 owns the disc-and-broadcast era; E-AC-3 with Joint Object Coding owns the streaming-Atmos era.

AC-3 in plain terms: five channels, one bass channel, 640 kbit/s

A surround mix is usually described as 5.1. The five is five full-range speakers — front left, front center, front right, surround left, surround right. The ".1" is one extra channel that carries only the deep bass, called the Low-Frequency Effects channel, or LFE; it feeds the subwoofer. So "5.1" means six channels of audio, five of them full-range and one of them bass-only.

AC-3 packs that 5.1 mix into a stream with a hard ceiling on its data rate. The standard allows a maximum coded bit rate of 640 kbit/s (ATSC A/52:2012, §5.3.2). It supports three sampling rates — 32, 44.1, and 48 kHz — and in practice almost all video uses 48 kHz, the rate the rest of the video world settled on, explained in sample rate: 44.1, 48, 96 kHz.

AC-3 chops sound into fixed blocks. Each AC-3 frame holds 1,536 samples per channel. At 48,000 samples per second, the arithmetic for one frame's duration is:

frame duration = samples per frame ÷ sample rate
frame duration = 1,536 ÷ 48,000
frame duration = 0.032 seconds = 32 milliseconds

So an AC-3 frame is 32 milliseconds of sound. That number matters when you sync audio to video or worry about latency — a topic frames, packets, granules covers for codecs in general.

For two decades that was enough. AC-3 became the mandatory surround codec for the DVD, for the ATSC digital-television system used across North America, and for cable and satellite. If a device made in the last twenty-five years can decode any surround format at all, it can almost certainly decode AC-3. That ubiquity is AC-3's lasting value: it is the format you fall back to when you do not know what the device supports.

E-AC-3 (Dolby Digital Plus): the same idea, with room to grow

By the early 2000s, AC-3's ceilings were a problem. The 640 kbit/s cap limited quality at the top end, and 5.1 was no longer the most channels anyone wanted. Dolby's answer, standardized in 2004 and known as E-AC-3 (Enhanced AC-3) or Dolby Digital Plus (DD+), kept the AC-3 core and pushed every limit outward.

Three changes matter to a product team. First, the data-rate range widened enormously: DD+ runs from 32 kbit/s all the way to 6.144 Mbit/s (ETSI TS 102 366 V1.4.1, Annex E). The low end means DD+ can be efficient at modest bitrates where AC-3 would sound strained; the high end means it has headroom AC-3 never had. Second, the channel count grew from AC-3's 5.1 ceiling to as many as 15 full-range channels plus LFE — written 15.1. Third, DD+ added substreams, a way to pack several audio programs — a main mix, a foreign-language dub, a director's commentary, a described-video track for blind viewers — into one stream, which connects to multi-language audio.

Under the hood, DD+ adds coding tools that AC-3 lacks: the Adaptive Hybrid Transform for steady tones, Spectral Extension to rebuild high frequencies cheaply, Enhanced Coupling to preserve stereo phase, and transient pre-noise processing to clean up sharp sounds. You do not need to track these by name. The takeaway is that DD+ squeezes more quality out of each bit than AC-3, which is exactly why broadcasters and streamers adopted it as a higher-efficiency replacement.

Figure 2. What E-AC-3 adds to AC-3: more channels, a far wider bitrate range, multi-program substreams, extra coding tools, and the Atmos-carrying JOC layer.

Here is the same comparison as a table you can paste into a planning doc.

Table 1. AC-3 versus E-AC-3 at a glance. Channel, bitrate, and substream figures from ETSI TS 102 366 V1.4.1 and ATSC A/52:2012.

How DD+ carries Dolby Atmos without breaking old devices

Dolby Atmos is object-based audio. Instead of mixing sounds permanently into fixed channels — "put the helicopter in the rear-left speaker" — an Atmos mix stores each sound as an object with a position in 3-D space, plus a set of bed channels for the ambient layer. The decoder in the room then renders those objects to whatever speakers are actually present, whether that is a 5.1 living room or a 7.1.4 home cinema.

The problem: an object scene is a different kind of data from a channel mix, and the millions of DD+ decoders already in TVs and soundbars only know how to read channels. Shipping a brand-new format would have stranded that hardware.

Dolby's solution is Joint Object Coding, or JOC. Think of JOC as a recipe card tucked inside an ordinary meal. The DD+ stream still carries a normal 5.1 mix — the meal — that any existing decoder can play. Riding alongside it as side data is the JOC payload: a compact set of instructions plus per-object metadata that an Atmos-capable decoder uses to reconstruct the original object scene from that 5.1 downmix (Dolby Developer Knowledge Base, "What is Dolby Digital Plus JOC?", accessed 2026-06-05).

The result is one stream that does two jobs. An old 5.1 decoder reads only the channel bed and ignores the JOC side data, so it plays correct surround sound. An Atmos decoder reads both and rebuilds the immersive mix. This is why streaming services can label a title "Dolby Atmos" while sending a single DD+ stream that still works on a decade-old soundbar. The marketing name for this combination is Dolby Digital Plus with Dolby Atmos, sometimes written "DD+ JOC".

Figure 3. One stream, two outcomes. The 5.1 bed plays everywhere; the JOC side data lets an Atmos decoder rebuild the object scene. Old hardware never sees the difference.

What bitrates the platforms actually use in 2026

Numbers anchor this better than adjectives. The clearest public example is Netflix, which delivers its surround and immersive audio as DD+ and has published its bitrates.

For 5.1 surround, Netflix raised its DD+ bitrate from 192 kbit/s to as high as 640 kbit/s, and stated that DD+ at and above 640 kbit/s is perceptually transparent — meaning a listener cannot reliably tell it from the studio master (Netflix Technology Blog, "High Quality Audio", 2019). For Dolby Atmos, Netflix raised the delivery rate from 448 kbit/s to 768 kbit/s. These are delivered adaptively: a device on a fast connection gets the higher rate, a constrained device gets a lower one, the same way video adapts in audio adaptive bitrate ladders.

To make the storage side concrete, take a 90-minute feature with a 5.1 DD+ track at 640 kbit/s:

size = bitrate × duration
size = 640,000 bits/s × (90 × 60) seconds
size = 640,000 × 5,400 = 3.456 × 10^9 bits
size = 3.456 × 10^9 ÷ 8 ÷ 1,000,000 ≈ 432 megabytes

So one high-quality 5.1 audio track for a feature is roughly 432 MB — small next to the video, but it multiplies fast once you add an Atmos rendition and several language tracks, the math worked out in multi-language audio storage and CDN cost.

As of 2026, DD+ is the codec behind the Dolby Atmos badge on Netflix, Disney+, Apple TV+, and Amazon Prime Video, because DD+ with JOC is the practical way to send Atmos over a stream to consumer hardware. Plain AC-3 remains the safe surround fallback for devices that predate DD+. Newer broadcast systems such as ATSC 3.0 add Dolby's next-generation codec, AC-4, covered in AC-4 explained, but for streaming today the workhorse is still DD+.

The loudness metadata you must not strip: dialnorm

Every Dolby Digital stream carries a small but critical piece of metadata called dialnorm — short for "dialogue normalization". It is a single number that tells the decoder how loud the dialogue in this program is, on a perceptual loudness scale, so the decoder can adjust playback to a consistent reference level. It is the mechanism that stops the audio jumping in volume when a viewer changes channels or moves from a show to an ad, and it underpins loudness-compliance regimes such as the CALM Act, detailed in loudness normalization: EBU R128, BS.1770, ATSC A/85.

DD+ also defines decoder operating modes that apply different amounts of dynamic-range compression for different rooms: a Line mode at a −31 LKFS reference for home cinema, an RF mode at −20 LKFS for TV speakers, and a Portable mode at −11 LKFS for phones and headphones (LKFS is the loudness unit ATSC uses; it is the same scale as LUFS used in streaming). The point for a product team is simpler than the numbers: this metadata is part of the bitstream, and if your pipeline mangles it, loudness goes wrong.

Common mistake: treating a Dolby stream as if it were PCM

The single most common way teams corrupt Dolby audio is to let a tool treat the compressed bitstream as ordinary uncompressed audio — called PCM, the raw sample format. Much consumer and even professional gear does not recognize DD+ as an encoded format and will pass it through a S/PDIF cable or store it in a WAV file as though it were PCM. That is fine if the bytes are passed through untouched. But any operation that is harmless to real PCM — a volume change, a sample-rate conversion, a "normalize" pass — will scramble the Dolby bitstream and destroy it. The fix is discipline in the pipeline: carry Dolby streams as opaque, never apply gain or resampling to them, and only ever decode them with a real Dolby decoder. A second frequent error is stripping dialnorm or substream metadata during repackaging, which silently breaks loudness and secondary-language tracks.

Where Fora Soft fits in

We build OTT and Internet-TV products, video-conferencing and e-learning platforms, telemedicine systems, and surveillance and AR/VR software, and audio delivery to living-room devices is a recurring part of that work. When a client's catalogue has to reach smart TVs, soundbars, and set-top boxes, AC-3 and E-AC-3 are usually the formats those devices decode, so the engineering questions are practical: which renditions to package, how to keep dialnorm and substream metadata intact through transcode and repackage, and when an Atmos rendition earns its extra storage and CDN cost. We have handled those trade-offs across streaming and conferencing builds since 2005.

What to read next

• How audio compression works: the four ideas behind every modern codec
• AAC family: AAC-LC, HE-AAC v1, HE-AAC v2, xHE-AAC
• Audio in HLS, DASH, CMAF: how a streaming player picks an audio track

Call to action

• Talk to a audio engineer — book a 30-minute scoping call to talk through your dolby digital plus explained plan.
• See our case studies — 250+ shipped projects across video streaming, WebRTC, OTT, telemedicine, e-learning, surveillance, and AR/VR.
• Download the Dolby Digital Family — cheat sheet — One page: AC-3 vs E-AC-3 comparison, how DD+ carries Dolby Atmos via JOC, 2026 platform bitrates (Netflix 5.1 and Atmos), and the 'a Dolby stream is not PCM' pitfall.

References

1. ETSI TS 102 366 V1.4.1 (2017-09), Digital Audio Compression (AC-3, Enhanced AC-3) Standard. The controlling specification: the main body defines AC-3; Annex E defines E-AC-3, including the 6.144 Mbit/s ceiling, 15.1 channel support, and substream structure. https://www.etsi.org/deliver/etsi_ts/102300_102399/102366/01.04.01_60/ts_102366v010401p.pdf
2. ATSC A/52:2012 (17 December 2012), Digital Audio Compression (AC-3, E-AC-3) Standard. The North American broadcast specification; source for the AC-3 640 kbit/s maximum and the 1,536-sample frame. https://www.atsc.org/wp-content/uploads/2015/03/A52-201212-17.pdf
3. Dolby Developer Knowledge Base, What is Dolby Digital Plus JOC (Joint Object Coding)? (accessed 2026-06-05). First-party description of how a DD+ Atmos decoder reconstructs the object scene from a 5.1 bed plus side metadata. https://developerkb.dolby.com/support/solutions/articles/16000067758-what-is-dolby-digital-plus-joc-joint-object-coding-
4. Dolby Professional, Dolby Digital Plus technology page (accessed 2026-06-05). Vendor overview of DD+ deployment in streaming and broadcast. https://professional.dolby.com/technologies/dolby-digital-plus/
5. Netflix Technology Blog, High Quality Audio: Studio-quality sound on Netflix (2019). Source for Netflix's DD+ 5.1 bitrate (192→640 kbit/s, transparent at 640) and Atmos (448→768 kbit/s). Used for platform bitrate figures; where it discussed perceptual transparency, cross-checked against the codec ceilings in ETSI TS 102 366. https://netflixtechblog.com/
6. Andersen, R. L., et al. (2004), Introduction to Dolby Digital Plus, an Enhancement to the Dolby Digital Coding System, Journal of the Audio Engineering Society / AES Convention Paper. The original technical paper introducing E-AC-3's coding tools (AHT, SPX, ECPL, TPNP). https://web.archive.org/web/20161119192949/https://www.dolby.com/us/en/technologies/aes-convention-paper-intro-to-dolby-digital-plus.pdf
7. Library of Congress, AC-3 Compressed Audio (Dolby Digital), Revision A — Sustainability of Digital Formats (accessed 2026-06-05). Independent archival summary of the AC-3 format and its history. https://www.loc.gov/preservation/digital/formats/fdd/fdd000209.shtml
8. ETSI TS 101 154, Specification for the use of Video and Audio Coding in Broadcasting Applications based on the MPEG-2 Transport Stream. The DVB carriage specification for E-AC-3, referenced for how DD+ rides in broadcast transport streams. https://www.etsi.org/standards
9. IEC 61937-3, Interfaces for non-linear PCM encoded audio bitstreams — Part 3: Non-linear PCM bitstreams according to the AC-3 and enhanced AC-3 formats. The interface standard governing carriage of AC-3 / E-AC-3 over S/PDIF and HDMI; cited for the PCM-passthrough pitfall. https://webstore.iec.ch/