Audio Quality Settings: Bitrate, Sample Rate, Bit Depth

One reference guide covering audio bitrate, sample rate, and bit depth. What each setting does, how they interact, and which values to use for each use case.

February 24, 2025

Five settings determine how an audio file sounds and how much it weighs: bitrate, sample rate, bit depth, channel count, and codec choice. Encoder presets bundle them. Pick the wrong combination and you ship audio that's too large, too small, too noisy, or incompatible with the destination. This is the one-stop reference covering each setting, the interactions between them, and recommended values for each common use case.

Bitrate: Data Per Second

Bitrate measures how many bits per second the audio file uses, expressed in kbps. It applies to lossy codecs (MP3, AAC, Opus, Vorbis) where the encoder explicitly targets a data rate.

128 kbps MP3: older streaming default; audible artifacts on cymbals and dense music.
192-256 kbps MP3: transparent for casual listening on most equipment.
320 kbps MP3: maximum standard MP3 bitrate; transparent for nearly all listeners.
96-128 kbps Opus or AAC: comparable to higher-bitrate MP3 due to codec efficiency.

For lossless formats, bitrate is calculated, not chosen. CD-quality 44.1 kHz × 16-bit × 2 channels = 1,411 kbps. FLAC compresses this to roughly 700-900 kbps with lossless reconstruction.

The trap: comparing bitrates across codecs is meaningless. 128 kbps MP3 and 128 kbps Opus are not the same audio quality.

Sample Rate: Resolution in Time

Sample rate is how many times per second the analog signal is measured, in Hz or kHz. The Nyquist theorem says a sample rate represents frequencies up to half its value:

8 kHz: telephone-band audio, up to 4 kHz frequency content.
22.05 kHz: old web audio, voice memos.
44.1 kHz: CD quality, captures up to 22.05 kHz, covers all of human hearing.
48 kHz: video and broadcast standard, captures up to 24 kHz.
88.2 / 96 kHz: hi-res audio masters, processing headroom.
176.4 / 192 kHz: ultra-hi-res, mostly studio mastering use.

Above 48 kHz, the additional samples capture frequencies above human hearing. They matter for studio processing (avoiding aliasing during DSP), not for playback. Consumer delivery at 96+ kHz is marketing, not audio improvement.

Bit Depth: Resolution in Amplitude

Bit depth is how many bits represent each sample's amplitude:

16-bit: 65,536 levels, 96 dB dynamic range. CD standard. Adequate for any consumer playback scenario.
24-bit: 16.7 million levels, 144 dB dynamic range. Pro audio standard. The extra range is used for processing headroom, not audible improvement at playback.
32-bit float: effectively unlimited range; used inside DAWs for intermediate calculations. Not used for delivery.

The difference between 16-bit and 24-bit at playback is inaudible in nearly all conditions. The case for 24-bit is during recording and mixing where multiple gain stages can introduce quantization noise — 24-bit's extra headroom prevents that noise from accumulating.

Channel Count

How many independent audio streams the file contains:

Mono (1 channel): voice content, voice memos, podcasts where stereo adds nothing.
Stereo (2 channels): music, most podcast content with music beds, anything with spatial information.
5.1 surround (6 channels): film, home theater. Not supported by MP3; use AAC or Opus.
Higher (7.1, Dolby Atmos): specialty broadcast and film delivery.

Halving channels halves the file size for uncompressed audio and roughly halves the bitrate need for lossy codecs. Voice content in stereo is wasteful — both channels carry the same information.

Codec Choice

The wrapper around the data:

WAV / AIFF: uncompressed PCM. Largest files, zero CPU to decode.
FLAC / ALAC: lossless compression. ~50% of WAV size, bit-perfect.
MP3: universal compatibility, less efficient than modern codecs.
AAC: more efficient than MP3, default in Apple ecosystem.
Opus: most efficient lossy codec for general use, native to web audio.
Vorbis: older lossy codec, mostly replaced by Opus.

Encoder Presets

Many encoders bundle bitrate + complexity + tuning into named presets:

LAME MP3 V0: ~245 kbps VBR average, transparent for nearly everyone.
LAME MP3 V2: ~190 kbps VBR average, the practical floor for music.
AAC iTunes Plus: 256 kbps VBR.
Opus default: 96 kbps VBR (libopus encoder default).

Presets are the easy path. Custom settings make sense only if you have a specific reason.

Recommended Settings by Use Case

Music production master: 48 kHz / 24-bit WAV (or FLAC for archive). Stereo.
Music distribution:
Podcast production: record 48 kHz / 24-bit WAV mono. Master to -16 LUFS. Deliver as 96-128 kbps mono MP3.
Voice memos / voice notes: 16 kHz to 48 kHz mono, 32-64 kbps Opus or 96 kbps MP3.
Web game audio: 48 kHz Opus at 64-96 kbps for music, 32-48 kbps voip mode for voice.
Video soundtrack: 48 kHz / 24-bit WAV through production. Deliver as 192 kbps AAC inside MP4.
Archive of consumer-purchased music: keep AAC/MP3 source as-is or transcode to FLAC if you need lossless wrapper.
Lecture / interview recording: 48 kHz / 16-bit WAV mono recording. Deliver as 128 kbps MP3 or send WAV to transcription service.

Quick Reference Table

| Use case | Sample rate | Bit depth | Codec | Bitrate | |----------|-------------|-----------|-------|---------| | CD audio | 44.1 kHz | 16-bit | PCM | 1,411 kbps | | DVD audio | 48 kHz | 24-bit | PCM | 2,304 kbps | | Studio master | 96 kHz | 24-bit | PCM | 4,608 kbps | | Music streaming (high) | 44.1 kHz | n/a | AAC | 256 kbps | | Music streaming (med) | 44.1 kHz | n/a | Opus | 96 kbps | | Podcast | 44.1 kHz | n/a | MP3 | 128 kbps mono | | WhatsApp voice note | 16 kHz | n/a | Opus | 16 kbps mono | | YouTube audio (high) | 48 kHz | n/a | Opus | 160 kbps |

For deeper coverage of each setting individually, see the audio bitrate guide, the sample rate explainer, and the bit depth deep dive. The bitrate vs sample rate comparison covers the most common confusion between the two settings.