What Is Audio Bitrate?

Bitrate is the number of bits used to encode one second of audio after compression. Measured in kbps (kilobits per second; 1 kbps = 1000 bits/sec) for audio; bps (bits/sec) for theory; Mbps (megabits/sec) for video. A 128 kbps MP3 uses 128,000 bits — 16,000 bytes, 16 KB — for each second of audio it represents. The higher the bitrate, the more data the encoder has to faithfully reproduce the source. For lossy codecs (MP3, AAC, Opus), bitrate is a quality knob: more bits = more frequencies preserved, more transients accurately encoded, less audible artefacting. For lossless codecs (FLAC, ALAC), bitrate is a function of source content (loud, dense music has a higher bitrate after lossless compression than quiet, simple content) — it varies but quality is always perfect. For uncompressed PCM, bitrate is mechanically determined by sample_rate × bit_depth × channels.

Three encoding modes determine how bits are allocated frame-to-frame. CBR (Constant Bitrate): every frame uses the same bitrate. Predictable file size, easy to seek by time, required by some streaming and broadcast specs. VBR (Variable Bitrate): the encoder spends more bits on complex passages, fewer on simple ones, achieving better quality per byte. Ideal for personal libraries and offline listening. ABR (Average Bitrate): the encoder targets a long-run average with controlled per-frame variation; intermediate between CBR and VBR. For most lossy audio, VBR is the right default. CBR is required for ACX audiobooks (192 kbps CBR mono is firm spec), some podcast platforms, and broadcast distribution. Reference: [what-is-vbr](/guide/what-is-vbr) for the full deep-dive.

Voice content (podcasts, audiobooks, voice memos): 64-128 kbps MP3 mono, or 32-64 kbps Opus mono — speech needs little, mono halves it again. Casual music listening on earbuds and laptop speakers: 128-192 kbps MP3 or 96-128 kbps AAC stereo. Quality music listening on good headphones or speakers: 192-256 kbps MP3 or 192 kbps AAC. Audiophile / library archive of compressed music: 320 kbps MP3 CBR or V0 VBR; 256 kbps AAC. Streaming uploads to platforms (Spotify, Apple Music, YouTube Music): deliver lossless [FLAC](/wav-to-flac) where accepted; otherwise 320 kbps MP3 or 256 kbps AAC. Lossless archive: FLAC or ALAC at full source bit depth and rate. Real-time voice (Discord, WhatsApp): Opus 32-64 kbps automatically. ACX audiobooks: MP3 192 kbps CBR mono (firm). See [audio-bitrate-guide-by-use-case](/blog/audio-bitrate-guide-by-use-case) for an extended table.

Size in MB ≈ (bitrate in kbps × duration in seconds) / 8000. Worked examples. 3-minute song at 192 kbps: 192 × 180 / 8000 = 4.32 MB. 60-minute podcast at 96 kbps mono: 96 × 3600 / 8000 = 43.2 MB. 4-minute song at 320 kbps: 320 × 240 / 8000 = 9.6 MB. PCM math (uncompressed WAV): bitrate = sample_rate × bit_depth × channels. CD-quality (44.1 kHz × 16-bit × 2): 1411 kbps = ~10.5 MB/min. Studio (48 kHz × 24-bit × 2): 2304 kbps = ~17 MB/min. Translate per-minute rates by multiplying: a 10-minute file at 1411 kbps is ~105 MB. Use these formulas to estimate before encoding — useful for planning streaming bandwidth, podcast hosting tier, and archive disk needs.

Quality does not scale linearly with bitrate. Below ~80 kbps for music, every additional 32 kbps brings dramatic improvement — going from 64 to 96 kbps removes obvious swirly artefacts. From 96 to 128 kbps: noticeable improvement on cymbals and treble. From 128 to 192 kbps: subtle improvement audible on critical listening. From 192 to 256 kbps: only trained listeners on excellent equipment can reliably distinguish in blind tests. From 256 to 320 kbps: essentially imperceptible. Lossless: fully transparent, no audible difference from 320 kbps in blind ABX testing for the vast majority of listeners on the vast majority of equipment. Practical implication: doubling bitrate from 128 to 256 kbps doubles file size for diminishing returns. The sweet spot for music distribution is 192-256 kbps; for voice, 96-128 kbps. Going higher is rarely audible, almost always wasted.

Each codec has practical bitrate ceilings beyond which they cannot use additional bits effectively. MP3: maximum 320 kbps. Above that, the spec doesn't define and encoders cannot improve quality. AAC-LC: maximum 256-320 kbps practical; specifications allow higher but encoders show no improvement. HE-AAC: optimised for low bitrates (16-64 kbps); above 96 kbps switches behaviour to AAC-LC. Opus: 6-510 kbps; transparent at 96 kbps stereo, perceptually lossless at 256 kbps. Vorbis: 32-500 kbps; transparent at 192-256 kbps. FLAC: variable based on content; CD-quality typically 700-900 kbps after lossless compression. ALAC: similar to FLAC. PCM: fixed at sample_rate × bit_depth × channels (1411 kbps for CD). Don't encode above the codec's practical ceiling — it just inflates file size.

Three quick ways to inspect bitrate. (1) macOS Finder: Get Info on the file shows kbps for music files; works for MP3, AAC, FLAC. (2) ffprobe (universal): 'ffprobe -i input.mp3' prints 'Duration: ...' followed by 'bitrate: 192 kb/s'. For VBR files, this shows the average. (3) MediaInfo (free, all platforms): full per-stream report with mode (CBR vs VBR), nominal vs actual bitrate, sample rate, channels. (4) Audacity File > Track Information for imported files. If the file's bitrate is wrong for the destination spec, convert via [WAV to MP3](/wav-to-mp3), [MP3 to WAV](/mp3-to-wav), or other tools — never re-tag without re-encoding; the bitrate is determined by the actual encoded data, not by metadata. To re-target an existing MP3 to a lower bitrate without changing format, use the [MP3 compressor](/compress-mp3). See [vbr-vs-cbr-mp3](/blog/vbr-vs-cbr-mp3) for verification specifics.