96000 Hz

96000 Hz (96 kHz) is a high-resolution digital audio sample rate: the signal is measured 96,000 times per second, so it can represent frequencies up to 48 kHz, more than twice the upper limit of human hearing.

In studios, 96 kHz is used mainly for its processing advantages: cleaner filtering, less aliasing, lower latency and more room for extreme pitch-shifting. Whether listeners can hear it on finished music is a separate, and much more debated, question.

What does 96 kHz mean?

An analogue-to-digital converter working at 96 kHz takes a measurement of the incoming voltage every 10.4 microseconds, twice as often as at 48 kHz. Each measurement is stored with a certain precision, the bit depth, which at this rate is almost always 24-bit. That is why you will often see the combination written as 24/96.

According to the sampling theorem, a digital system can represent frequencies up to half the sample rate, the Nyquist frequency. At 96 kHz that is 48 kHz. Nobody hears 48 kHz, but that extra space above the audible band is exactly what makes the rate useful in production, as explained below.

A simple way to picture it: 96 kHz is like filming at a much higher frame rate than you need for normal playback. You will not see a difference at normal speed, but you have far more material to work with if you slow things down or process them heavily.

How it is defined

The Audio Engineering Society’s AES5 recommended practice keeps 48 kHz as the main professional rate, and allows 96 kHz for applications that need a bandwidth wider than 20 kHz or a wider transition region for anti-alias filtering. Consumer “hi-res” definitions, covered below, also frequently refer to 96 kHz/24-bit.

One number, two meanings

In production, 96000 Hz always means the sample rate. As a physical frequency, a 96,000 Hz tone would be ultrasonic, in the range used by bats or by ultrasonic measurement equipment, and far outside anything musical.

A short history of 96 kHz

• 1980s–1990s: the filter problem. CD (44.1 kHz) and professional (48 kHz) rates left little room between 20 kHz and the Nyquist limit, so early converters needed very steep anti-aliasing filters, which designers and mastering engineers worked hard to make transparent.
• 1997: 24/96 products appear. British company dCS, whose 900 had been presented as the first 24-bit A/D converter back in 1989, introduced the 972, a 24/96 digital-to-digital converter, in 1997 (dCS timeline). High-end studio converters running at 96 kHz followed quickly.
• 2000: DVD-Audio. The DVD-Audio format brought 96 kHz/24-bit multichannel and up to 192 kHz stereo to consumers. It never became mainstream, but it fixed 24/96 in the public idea of “high resolution”.
• 2014: hi-res definitions. The Japan Audio Society’s Hi-Res AUDIO logo set minimum requirements of 96 kHz/24-bit for digital processing (and 40 kHz response for analogue stages), while the US definition agreed by the DEG, CEA (now CTA), the Recording Academy and major labels described hi-res as lossless audio mastered from better-than-CD-quality sources (High-resolution audio, Wikipedia).
• 2021 onward: streaming. Apple Music added Hi-Res Lossless streaming up to 24-bit/192 kHz in 2021, and other services offer similar tiers, so 96 kHz masters now reach listeners directly.

How 96 kHz works and what it really improves

A wide, relaxed anti-aliasing filter

Every converter needs an anti-aliasing filter to remove frequencies above the Nyquist limit before sampling. At 44.1 kHz that filter has to pass 20 kHz and block 22.05 kHz: a very narrow transition. At 96 kHz the transition can stretch from 20 kHz all the way to 48 kHz, so the filter can be gentle, with its side effects pushed far above the audible range. This is the original engineering argument for high sample rates. Modern oversampling converters also solve the problem internally at 44.1 and 48 kHz, which is one reason the audible difference is now small.

Less aliasing from plugins

Aliasing does not only happen at the converter. Non-linear processing inside the computer, such as saturation, distortion, amp simulation, clipping and fast limiting, creates new harmonics. If those harmonics land above the Nyquist limit, they fold back into the audible range as aliasing: inharmonic, gritty tones that were never in the music. At 44.1 kHz the Nyquist limit is 22.05 kHz; at 96 kHz it is 48 kHz, so many harmonics that would alias at a lower rate are simply represented correctly and removed later. Many plugins get the same benefit through built-in oversampling.

More room near the top of the EQ

Some digital EQ designs change shape as they approach the Nyquist frequency, an effect often called cramping: a high shelf or bell near 16–20 kHz does not behave like its analogue model. Working at 96 kHz moves the Nyquist limit far away from the audible top end, and many modern EQs also compensate for cramping at lower rates.

Lower latency for the same buffer size

Audio interfaces process audio in buffers measured in samples. At a higher sample rate, the same number of samples passes in less time, so a fixed buffer size gives lower latency:

Buffer size	44.1 kHz	48 kHz	96 kHz
64 samples	1.45 ms	1.33 ms	0.67 ms
128 samples	2.90 ms	2.67 ms	1.33 ms
256 samples	5.80 ms	5.33 ms	2.67 ms
512 samples	11.61 ms	10.67 ms	5.33 ms

These figures are per buffer; real round-trip latency also includes converter and driver delays. The catch is that the computer has to do twice the work in the same time, so a system that runs smoothly at 128 samples at 48 kHz may need a larger buffer at 96 kHz. This latency advantage is one reason many digital live consoles run their audio engine at 96 kHz.

Storage and data rates

Uncompressed audio size is simply sample rate × bit depth × channels:

Format	Data rate (stereo)	Approx. size per minute (stereo)
44.1 kHz / 16-bit (CD)	1,411 kbps	10.6 MB
48 kHz / 24-bit	2,304 kbps	17.3 MB
96 kHz / 24-bit	4,608 kbps	34.6 MB
192 kHz / 24-bit	9,216 kbps	69.1 MB

In practice, a 96 kHz session takes twice the disk space and roughly twice the CPU of the same session at 48 kHz.

Can you hear 96 kHz?

This is where the debate lives, and it helps to separate production (where 96 kHz has clear technical benefits) from playback of finished music (where any audible benefit is small at best).

• Meyer and Moran (2007): in 554 trials, listeners could not detect a CD-quality 16-bit/44.1 kHz stage inserted into high-resolution playback, scoring 49.8% (JAES, September 2007).
• Reiss (2016): a meta-analysis of 18 studies, with more than 400 participants and over 12,500 trials, found a small but statistically significant ability to tell high-resolution audio from standard quality, only slightly above chance overall, with considerably better results among listeners who had been trained for the task (JAES, June 2016).

Taken together: if there is an audible difference in playback, it is subtle, depends on training and on the playback system, and it is far smaller than the difference made by the recording, the mix, the monitors or the room.

The “time resolution” argument

One popular claim is that 44.1 kHz cannot capture timing finer than its sample interval (22.7 microseconds), so it must smear transients that the ear could resolve, citing research such as Milind Kunchur’s 2007 experiments in which listeners detected time misalignments of about 6 microseconds (Electronic Journal “Technical Acoustics”). The research on hearing is real, but the conclusion about sampling is a misunderstanding: a band-limited sampled signal can represent timing differences far smaller than the interval between samples. What a higher sample rate really adds is bandwidth above 20 kHz, not finer timing within the audible band.

96 kHz in practice

Sound design and extreme pitch-shifting

Sound designers often record at 96 kHz or higher for one practical reason: pitching down. Pitch a 48 kHz recording down one octave and everything that was between 12 and 24 kHz ends up between 6 and 12 kHz, with nothing above: the result sounds dull. Pitch a 96 kHz recording down an octave and content that was between 20 and 40 kHz lands in the audible 10–20 kHz range, keeping the result bright and detailed. The benefit only exists if the microphone actually captures ultrasonic content; many studio microphones roll off well before 48 kHz, which is why field recordists pair high rates with extended-range microphones.

Recording and mixing

• Acoustic, classical and jazz recording: engineers who work at 96 kHz often cite space, air and natural decay. Producer Nick Patrick, describing the orchestral sessions for the Elvis Presley and Royal Philharmonic Orchestra albums, said “the recordings were all 96/24” and that for that kind of project it definitely made a difference (Sound & Vision, 2016).
• Heavily processed pop, hip-hop and electronic music: 96 kHz reduces aliasing from saturation and limiting across a whole session, but the same result can often be reached at 44.1 or 48 kHz with plugin oversampling.
• Film, TV and games: sound effects and Foley are often captured at 96 kHz or higher for flexibility, then delivered at 48 kHz.
• Live sound: many digital consoles use 96 kHz internally for lower latency.

Native 96 kHz or oversampling?

You do not need to run a whole session at 96 kHz to avoid aliasing. Oversampling inside individual plugins, such as saturators, clippers and limiters, gives much of the same benefit where it matters, while the rest of the project runs at 44.1 or 48 kHz. Reaper can even oversample individual plugins or whole FX chains. A common balanced approach is to record and mix at 48 kHz, use oversampling on non-linear plugins, and reserve 96 kHz for sound design or acoustic projects where its benefits are clearer.

Setting 96 kHz in your DAW

DAW	Where to set it
Pro Tools	Chosen when creating the session; change it later by saving a session copy at a new rate
Logic Pro	File > Project Settings > Audio > Sample Rate
Ableton Live	Settings (Preferences) > Audio > In/Out Sample Rate
FL Studio	Options > Audio settings > Sample rate
Cubase / Nuendo	Project > Project Setup > Sample Rate
Fender Studio Pro (formerly Studio One)	Song > Song Setup
Reaper	Project settings and audio device preferences; per-plugin and per-chain oversampling is also available

As with any rate: set it before recording, make sure your audio interface is running at the same rate, and convert once, at the end, to whatever your delivery requires.

Gear and tools

• Converters and interfaces: virtually every current audio interface supports 96 kHz; many support 192 kHz.
• Digital live consoles: several current console platforms run their internal processing at 96 kHz for low latency.
• Sample rate converters: iZotope RX, Weiss Saracon and the high-quality converters built into modern DAWs are used to bring 96 kHz masters down to 48 or 44.1 kHz for delivery.
• Extended-range microphones: needed if you actually want to capture ultrasonic content for sound design.

Common mistakes and myths

Mistakes

• Using 96 kHz without the hardware to support it: double the CPU load means dropouts and bigger buffers on an underpowered computer, cancelling the latency benefit.
• Upsampling to “get hi-res”: converting a 44.1 or 48 kHz file to 96 kHz does not add any information. Apple, for example, asks for masters at their original resolution rather than upsampled files.
• Mixing sample rates in one project: importing files at a different rate without conversion changes their speed and pitch.
• Ignoring true-peak headroom: converting down to 44.1 or 48 kHz and encoding to lossy formats can create peaks above 0 dBFS if the master is limited right to the ceiling.

Myths

• “If 96 kHz is better, 192 kHz must be better still.” Not for playback. As Chris “Monty” Montgomery of Xiph.Org explained, ultrasonic content reproduced by real-world amplifiers and speakers can create intermodulation distortion that lands in the audible range, so 192 kHz files offer no playback benefit and can make practical fidelity slightly worse (Xiph.Org, 2012). See 192000 Hz for more.
• “96 kHz has less noise or more dynamic range.” Noise floor and dynamic range come from bit depth, not sample rate.
• “96 kHz captures timing the ear needs that 44.1 kHz misses.” Sampled audio represents timing far more finely than the sample interval; the extra rate adds bandwidth, not timing precision.
• “Hi-res files always come from hi-res recordings.” A 96 kHz file can contain an upsampled 44.1 kHz master. The number on the file does not guarantee what is inside.

96 kHz vs related terms

Comparison	The difference
Sample rate vs bit depth	Sample rate sets the frequency range (up to half the rate); bit depth sets dynamic range and noise floor.
Sample rate vs bit rate	Sample rate is samples per second; bit rate is data per second, and lossy formats can reduce it drastically whatever the original sample rate.
96 kHz vs 48 kHz	96 kHz doubles bandwidth, halves buffer latency and doubles storage and CPU. 48 kHz is the delivery standard for video and broadcast.
Native 96 kHz vs oversampling	Native means the whole project runs at 96 kHz. Oversampling raises the rate only inside a specific plugin, then returns to the project rate.

Standards and definitions

Source	What it says
AES5 (AES5-2018, reaffirmed 2023)	48 kHz is the main professional rate; 96 kHz is permitted where more bandwidth or a wider anti-alias transition is needed.
Japan Audio Society Hi-Res AUDIO logo (2014)	Digital processing of at least 96 kHz/24-bit, analogue response to 40 kHz.
DEG / CEA / Recording Academy / labels (2014)	Lossless audio mastered from better-than-CD-quality sources.
Reiss meta-analysis, JAES 2016	Small but significant discrimination of hi-res in controlled tests, stronger with training.

Related terms

• Sample rate: the general concept.
• 48000 Hz: the professional delivery standard for video and broadcast.
• 192000 Hz: the highest common studio rate and why more is not always better.
• Oversampling: how plugins get high-rate benefits at normal project rates.
• Anti-aliasing filter: the filter that high sample rates make easier to design.
• 24-bit: the bit depth that usually goes with 96 kHz.
• Nyquist frequency: why 96 kHz can represent up to 48 kHz.

FAQ