Audio Sample Rate Explained: 44.1 vs 48 vs 96kHz

Sample rate is one of the two fundamental measurements of digital audio quality (the other is bit depth). It defines how many times per second the audio signal is measured and stored. The choice of sample rate affects file size, frequency response, and compatibility with different production environments.

The Nyquist Theorem in Plain English

Digital audio works by measuring the continuous analog waveform thousands of times per second and storing each measurement as a number. Each measurement is one sample. Sample rate, expressed in hertz, is how many of those measurements happen per second.

The Nyquist-Shannon sampling theorem (1928, formalized by Shannon in 1949) sets the rule for how high a sample rate must be to capture a given frequency without distortion: the sample rate must be at least twice the highest frequency you want to represent. Sample at less than 2× and you get aliasing — frequencies above the Nyquist limit fold back into the audible range as new, false frequencies that were not in the original signal. To capture 20 kHz cleanly, you need at least 40,000 samples per second.

Human hearing tops out at roughly 20 kHz when you are young and falls off measurably with age. The CD standard of 44.1 kHz captures everything below 22.05 kHz, leaving a small margin above the audible range for the anti-aliasing low-pass filter to work in. 48 kHz captures up to 24 kHz, and 96 kHz captures up to 48 kHz — both well beyond human hearing.

Why CDs Are 44.1 kHz: The Strange History

The CD's 44.1 kHz number is not arbitrary, but it is not derived from human hearing either. It comes from the early 1980s when digital audio was first being mastered and the only practical way to store the high data rates was on professional video tape. Sony's PCM-1600 system encoded audio as a fake video signal on U-matic tape, which had specific line-rate constraints. The math worked out to 44.1 kHz: for NTSC video at roughly 60 Hz with 245 active video lines per field carrying three audio samples per line, you get 60 × 245 × 3 = 44,100 samples per second. PAL video gave the same number through different math.

When Philips and Sony agreed on the CD format in 1980, they kept 44.1 kHz so masters from the existing PCM-1600 workflow could go straight to CD. Audible-range coverage (up to 22.05 kHz) was a useful side effect. The format then propagated to MiniDisc, DAT, music streaming, MP3, AAC, and FLAC because the entire mastering and distribution chain was built around it.

Why 48 kHz for Video

48 kHz became the video standard for engineering reasons unrelated to hearing. Professional digital video equipment in the 1980s was designed around 48 kHz to provide a clean 1:2 ratio with the 24 kHz Nyquist limit, integer relationships with common video frame rates (24, 30, 60 fps), and small mathematical advantages for synchronization. The DVD specification adopted it. ATSC and DVB broadcast standards adopted it. Every digital video camera ships at 48 kHz today: Sony, Canon, Blackmagic, RED. YouTube, Netflix, and Amazon Prime all expect 48 kHz audio. Theatrical Dolby and DTS encoding is 48 kHz. If your work ends up anywhere near video, start at 48 kHz from the first track.

Mixing 44.1 kHz music into a 48 kHz video project forces a resample. Modern DAWs handle this automatically, but the sample-rate-conversion artifacts (however small) are real and stack up if you cross rates multiple times. Pick the right rate at the start.

When 96 kHz and 192 kHz Matter (and When They Do Not)

Higher rates have legitimate uses, mostly in production rather than playback:

• Non-linear processing. Saturation plugins, tape emulation, distortion, aggressive EQ, and pitch shifting all generate frequencies above the Nyquist limit. At 44.1 kHz those new frequencies alias back into the audible range as harshness. At 96 kHz they have more room to live above hearing before being filtered. Many high-end plugins internally upsample to 96 or 192 kHz for this reason regardless of session rate.
• Pitch and time stretching. Stretching audio down a fifth at 44.1 kHz crowds the highs. Doing the same at 96 kHz preserves more headroom.
• Editing precision. Mathematical operations on audio (gain automation, crossfades, time alignment) are slightly cleaner at higher rates because of finer floating-point representation per second of audio.

What 96 kHz does not do is improve playback for a finished mix. Multiple controlled ABX listening tests have failed to show statistically significant ability to distinguish 44.1 kHz from 96 kHz on properly mastered music played through good equipment. The 2007 Meyer and Moran study (published in the AES journal) tested 60 listeners on hi-rez vs 44.1 kHz/16-bit downsampled and got chance-level results. Subsequent studies have refined the methodology but landed in the same place: most listeners cannot ABX 44.1 vs 96 kHz on music. Sample rate is not where audible differences come from in playback.

For home studios and podcasting, 48 kHz is plenty. The files are 9% larger than 44.1 kHz with no audible improvement. 96 kHz tracking files are double the size of 48 kHz with benefit only during processing, not playback.

App Default Behaviors

Different software defaults for sample rate cause friction in collaborative workflows:

• Logic Pro: 44.1 kHz default for new projects (changeable in project settings)
• Pro Tools: 48 kHz default in newer versions, configurable per session
• Ableton Live: 44.1 kHz default
• GarageBand: 44.1 kHz default
• Audacity: 44.1 kHz default, configurable in preferences
• Adobe Audition: 48 kHz default
• DaVinci Resolve: 48 kHz default (video software)
• iPhone Voice Memos: 44.1 kHz mono
• iPhone video recording: 48 kHz stereo (video alignment)

If you record a podcast in GarageBand at 44.1 kHz and the editor pulls it into a 48 kHz Pro Tools session for a video version, that's a resample. Settle on a rate before recording starts.

Sample Rate vs Bit Depth: Different Things

Sample rate is resolution in time — how often we measure. Bit depth is resolution in amplitude — how precisely each measurement is stored. Both contribute to digital audio quality but do different jobs. 16-bit gives 96 dB of dynamic range; 24-bit gives 144 dB. Bit depth affects the noise floor and headroom; sample rate affects the highest representable frequency. They are independent settings: a 16-bit/96 kHz file is valid and so is a 24-bit/44.1 kHz file. See PCM audio explained for the full picture.

Converting Between Sample Rates Cleanly

When you must convert (delivering 48 kHz mixes to a 44.1 kHz CD master, downsampling a 96 kHz session for streaming), use one good resampler and do it once. Quality varies across tools. iZotope RX, Adobe Audition, Pro Tools, Logic, and Reaper all use high-quality polyphase resamplers that are transparent for normal listening. ffmpeg's default soxr resampler is also excellent — 'ffmpeg -i input.wav -ar 44100 -af aresample=resampler=soxr:precision=28 output.wav' produces studio-grade output. Avoid chaining multiple conversions; each pass adds a tiny amount of artifact, and they accumulate.

For more on adjacent topics, see audio bitrate explained, what is WAV for the format that stores raw PCM at any sample rate, and lossless vs lossy for how compression interacts with sample rate. For converting between formats while keeping the original sample rate, the WAV to FLAC and FLAC to MP3 tools both preserve sample rate by default unless you explicitly downsample.

Practical Guidance

| Use case | Recommended sample rate | |---|---| | Music production for streaming/CD | 44.1 kHz | | Video production (any) | 48 kHz | | Audio-only podcast | 44.1 kHz or 48 kHz | | Video podcast | 48 kHz | | Game audio | 44.1 kHz or 48 kHz | | Professional recording / heavy processing | 88.2 or 96 kHz during tracking | | Streaming delivery | 44.1 kHz (music) / 48 kHz (video) | | Voiceover / spoken word | 48 kHz |