Dolby Digital (AC-3)
Dolby Digital, also called AC-3, is a lossy audio format supporting mono, stereo, and surround presentations.
A 5.1 Dolby Digital track carries five main signals and LFE inside a lossy coded stream. AC-3 and Dolby Digital name this same format. Its frame structure describes the sound, while metadata helps a decoder present it for the listening situation.
1,536 samples per channel in one AC-3 frame
A standard AC-3 synchronization frame represents 1,536 samples per channel. At 48 kHz, that is 1,536/48,000 = 32 ms of sound.
For three selected encoded rates:
| Bitrate | Bytes for a 32 ms frame at 48 kHz |
|---|---|
| 384 kbit/s | 1,536 |
| 448 kbit/s | 1,792 |
| 640 kbit/s | 2,560 |
The samples represented by a frame and the bytes storing it are different quantities. More encoded data gives the codec a different budget for describing the same timed sound.
The format supports mono, stereo, and surround modes. Common coded rates include 32, 44.1, and 48 kHz. The standard maximum bitrate is 640 kbit/s, while DVD-Video applies a 448 kbit/s limit. An application's constraint is narrower than a format capability.
Reconstruct audio using its coding choices
AC-3 represents audio in transformed blocks and uses allocated precision and related coding tools to fit its rate. The decoder reads those choices, reconstructs the frequency-domain data, and returns timed channel samples.
That approximation is the lossy part. The receiver's later volume or speaker processing is another stage and can change what the listener hears without changing the stored AC-3 bitstream.
Metadata changes the presentation, not the source recording
Three useful metadata roles are:
| Role | Decoder use |
|---|---|
| Dialogue normalization | Align programme playback level to the relevant reference |
| Dynamic-range information | Supply level-dependent adjustments for supported listening modes |
| Downmix information | Guide combining channels for fewer outputs |
For an illustrative dialogue-normalization report of −27 dB, compensation toward a −31 dB reference is −4 dB. Its amplitude factor is 10^(−4/20), approximately 0.631. This is a decoder-level adjustment, not a claim that every piece of spoken audio occupies one isolated channel.
A decoder applying dynamic-range information can reduce loud/quiet contrast for a chosen mode. A downmix can incorporate center dialogue and surrounds into stereo instead of losing them by taking only the front pair.
Place the decoder in the session
An Emby player receiving a compatible AC-3 track can decode and output PCM, or pass a supported encoded presentation to a receiver's decoder. In either case, source coding and selected decoder processing explain the output.
If two AC-3 paths sound different in level or dynamics, inspect which decoder and metadata/listening mode were used before attributing the difference to compressed bitrate. The related E-AC-3 format requires its own bitstream handling; it is not obtained by changing the AC-3 name.