What is the minimum bitrate for acceptable voice audio?

Opus at 24 kbps mono is the absolute floor for intelligible speech; 32 kbps mono is the modern minimum for clean voice. For AAC, 64 kbps mono is the lower bound. For MP3, 64 kbps mono is acceptable but sounds worse than Opus or AAC at the same bitrate because MP3's psychoacoustic model is older. Below these floors, sibilants (s, sh, t) start to fizz and break up. Podcasts traditionally use 128 kbps MP3 not for audible quality reasons but for universal player compatibility — Opus at 32 kbps would sound the same.

Does higher bitrate always mean better quality?

Only up to the transparency threshold — the bitrate at which the encoder reaches the limit of human perception. Above that, additional bits are spent on content that is already inaudibly faithful, so the file gets bigger without sounding better. For AAC, transparency lands around 192 kbps for music. For MP3 it is closer to 256 kbps. For Opus, around 128 kbps. Above those numbers, doubling the bitrate doubles the file size with essentially no audible improvement. The exception is critical mastering work where extra headroom matters for further processing.

What bitrate does Spotify use?

Spotify Free streams at up to 160 kbps OGG Vorbis on desktop. Spotify Premium streams at up to 320 kbps Vorbis on desktop and tablet, and 96/160/320 kbps on mobile depending on the network setting. The Premium tier uses Vorbis q9 (~320 kbps) which is essentially transparent for most listeners. Spotify HiFi (lossless FLAC at 1,411 kbps) was announced in 2021 but has not rolled out broadly. Apple Music and Tidal currently offer lossless and hi-res lossless at higher tiers; Amazon Music HD does as well.

What bitrate should I use for a 1-hour recording to minimize file size?

For voice content, Opus at 32 kbps mono produces a 14 MB file with excellent intelligibility — the smallest practical option. AAC at 64 kbps mono is ~29 MB and works on more players. MP3 at 64 kbps mono is ~29 MB and plays everywhere. For music, MP3 at 128 kbps stereo is ~57 MB and reasonable for casual listening; 192 kbps stereo is ~85 MB and approaches transparency. The general approach: pick the codec your destination supports, encode at its transparency threshold, and accept that file size as the lower bound for that quality level.

What bitrate is required for ACX / Audible audiobook submission?

Audible's ACX platform requires MP3 CBR mono at 192 kbps with a 44.1 kHz sample rate. Loudness must fall between -23 dB and -18 dB RMS, peak below -3 dB, and the noise floor below -60 dB. Each chapter must be its own MP3 file, with 1 second of room tone at the start and 2–5 seconds at the end. ACX rejects VBR, stereo, and any other bitrate. The encoder must be LAME or another high-quality MP3 encoder — older or buggy encoders sometimes produce files that fail ACX's automated quality check.

Audio Bitrate Guide: Right Settings for Every Use Case

Bitrate is the audio engineer's most-asked question and the one with the most context-dependent answer. The right bitrate is whatever produces transparent audio at the smallest file size for the destination platform — which means the answer changes for podcasts, music, VoIP, audiobooks, video, and game audio. This guide collects practical numbers that work in production. Each section gives the codec, the target bitrate, and the platform-specific reasons why.

Voice and Speech

| Use Case | Format | Bitrate | Notes | |----------|--------|---------|-------| | Podcast (voice only) | MP3 | 128 kbps CBR mono | Apple Podcasts recommended | | Audiobook (ACX) | MP3 | 192 kbps CBR mono | ACX required specification | | Voice call (VoIP) | Opus | 32-64 kbps | WhatsApp, Discord use range | | Voice memo | AAC | 64-128 kbps | iPhone Voice Memos default | | Webinar recording | AAC | 128 kbps | Zoom, Meet default | | Voice call (HD) | Opus | 64 kbps mono | Discord standard, Google Meet | | Audiobook (general, non-ACX) | MP3 | 64-96 kbps mono | Smaller, fine for non-Audible distribution | | Conference broadcast | AAC | 128 kbps stereo | Webcast platforms | | Broadcast contribution | AAC | 192-320 kbps | Studio-to-studio with headroom for post |

Voice is spectrally simpler than music — Opus and AAC both reach perceptual transparency on speech well below 64 kbps. The reason podcasts standardized on 128 kbps MP3 is universal player compatibility, not audio quality. Modern voice content can use Opus at 32 kbps mono and sound essentially identical to 128 kbps MP3 at one-quarter the size. To re-target an existing voice MP3 to a lower bitrate from this table, use the MP3 compressor.

Music

| Use Case | Format | Bitrate | Notes | |----------|--------|---------|-------| | Streaming (standard) | AAC or Opus | 128-160 kbps | Spotify free, YouTube | | Streaming (premium) | AAC or Opus | 256-320 kbps | Spotify Premium, Apple Music | | Personal archive | MP3 VBR | V0 (~245 kbps avg) | Near-transparent | | Lossless archive | FLAC | Lossless | ~700-1000 kbps effective | | CD rip | WAV/FLAC | 1411 kbps (WAV) | Bit-perfect from CD |

Video Production

| Use Case | Format | Bitrate | Notes | |----------|--------|---------|-------| | YouTube upload | AAC | 320 kbps | YouTube re-encodes; send high quality | | Broadcast delivery | WAV PCM | 1411 kbps | Lossless, per broadcaster spec | | Social media video | AAC | 192 kbps | Instagram, TikTok, Facebook |

Web and App Audio

| Use Case | Format | Bitrate | Notes | |----------|--------|---------|-------| | Web streaming | Opus | 64-128 kbps | Best quality-to-size for web | | Game audio (music) | OGG Vorbis | 128-192 kbps | Unity, Unreal standard | | Game audio (SFX) | WAV/OGG | Lossless or 128 kbps | Short sounds often kept lossless | | App meditation audio | AAC | 128-256 kbps | 128 kbps voice, 256 kbps music | | HTML5 voice | Opus | 32-64 kbps | All modern browsers | | Audiogram for social | AAC | 192 kbps | Inside MP4 with image | | Notification sounds | OGG or WAV | varies | Short, often kept WAV |

Game engines compress at build time. Source assets should be lossless WAV; the engine produces the compressed runtime variant. For Unity, the Audio Source inspector lets you choose Compression Format per platform. For Unreal, the Audio Asset Editor provides the same control.

The Transparency Thresholds

These are the bitrates at which most listeners in double-blind tests cannot distinguish the audio from the original:

MP3: ~192 kbps (VBR V2 for music, 192 CBR for voice)
AAC: ~128 kbps for voice, ~192 kbps for music
OGG Vorbis: ~160 kbps for music
Opus: ~64 kbps for voice, ~96-128 kbps for music

Opus is the most efficient codec at low bitrates — the reason it replaced older VoIP codecs.

Quick File-Size Math

A useful formula: file size in MB ≈ (bitrate in kbps × duration in seconds) ÷ 8,000.

60-minute podcast at 128 kbps MP3 mono: (128 × 3,600) / 8,000 ≈ 58 MB
4-minute song at 320 kbps MP3 stereo: (320 × 240) / 8,000 ≈ 9.6 MB
1-hour Zoom recording at 128 kbps AAC stereo: ≈ 58 MB
3-hour audiobook at 192 kbps mono: (192 × 10,800) / 8,000 ≈ 259 MB

ACX / Audible Specific Spec

Audible's ACX platform has the strictest consumer-facing audio bitrate spec on the internet:

Container/codec: MP3
Bitrate: 192 kbps CBR (no VBR, no other rate)
Channels: mono (no stereo)
Sample rate: 44.1 kHz
Loudness: between -23 dB and -18 dB RMS
Peak: below -3 dB
Noise floor: below -60 dB
Per-chapter file structure: 1 second of room tone at start, 2–5 seconds at end

Files that fail any of these get rejected by ACX's automated check.

When to Use Lossless

Use lossless (FLAC or WAV) when:

Archiving recordings that may be edited later

Distributing to professional collaborators

Ripping CDs for long-term storage

Recording live performances for permanent archive

A broadcaster or distributor specifies lossless ingestion

For everything that ends in headphones or speakers without further processing, well-encoded lossy at the transparency threshold sounds identical to lossless.

Codec Efficiency Ranked

If you had to pick one codec for everything, modern Opus would be the answer — it consistently outperforms older codecs at every bitrate tested. The ranking from most to least efficient:

1. Opus — best at all bitrates, especially under 128 kbps 2. AAC (HE-AAC for low bitrates, AAC-LC for high) — strong, broad device support 3. Vorbis — competitive with AAC on music 4. MP3 — older, less efficient, but universally compatible 5. WMA Pro — Microsoft's competitor, similar to AAC 6. AC-3 — designed for surround film, inefficient for stereo music

For a 4-minute song at perceptual transparency: Opus needs ~96 kbps, AAC ~128 kbps, Vorbis ~160 kbps, MP3 ~192 kbps. Same audible quality, different file sizes.

Why Bitrate Targets Differ Across Codecs

A 128 kbps MP3 and a 128 kbps Opus do not sound the same. The MP3 at 128 kbps is at the lower edge of acceptable for music; the Opus at 128 kbps is essentially transparent. The reason is encoder maturity and psychoacoustic-model efficiency. Opus, designed in 2012, benefits from 20 more years of audio research than MP3 (1993). It allocates bits more cleverly to perceptually important content and discards more aggressively from masked content.

This is why platform recommendations seem inconsistent. Apple Music uses 256 kbps AAC; Spotify uses 320 kbps Vorbis; YouTube uses 192 kbps AAC plus 160 kbps Opus. Each platform picked the bitrate that matches their codec's transparency point. The numbers differ but the perceived quality is similar.

See lossless audio is it worth it for the deeper analysis, VBR vs CBR for MP3 for the encoding theory, and MP3 128 kbps vs 320 kbps for the bitrate-to-quality curve.