What is sample rate?
Sample rate is the number of times per second that a digital audio system measures an analog signal. It is usually written in hertz (Hz) or kilohertz (kHz). A rate of 48 kHz means the system takes 48,000 samples every second for each channel. Those samples are the time-based snapshots that make it possible to store, edit, process, and play audio in a digital workstation, recorder, phone, camera, or game engine.
Higher sample rates create more measurement points per second. That can extend the theoretical frequency range and give filters more room to work, but it also increases the amount of data your computer, recorder, interface, and storage need to handle. Sample rate is therefore a project decision, not just a quality slider. The best setting depends on what you are recording, where the audio will be delivered, how much processing you expect, and whether you need to stay aligned with video or broadcast standards.
What is Nyquist frequency?
Nyquist frequency is half the sample rate. It is the highest theoretical frequency that can be represented by a sampled signal before aliasing becomes a problem. If a session runs at 44.1 kHz, the Nyquist frequency is 22.05 kHz. If a session runs at 48 kHz, the Nyquist frequency is 24 kHz. If a session runs at 96 kHz, the Nyquist frequency is 48 kHz.
This does not mean every system perfectly captures sound right up to Nyquist. Real converters use anti-aliasing filters so frequencies above the useful band do not fold back into the audible range as aliasing. That filter behavior is one reason 44.1 kHz and 48 kHz leave a little space above the upper edge of typical human hearing. The calculator’s Nyquist readout gives the theoretical ceiling; the real-world result still depends on converter design and filtering.
Sample rate vs bit depth
Sample rate and bit depth solve different problems. Sample rate affects the frequency range that can be represented. Bit depth affects the number of amplitude steps available for each sample, which is usually discussed as dynamic range and noise floor. A higher sample rate does not automatically give you more headroom, and a higher bit depth does not extend the highest frequency that can be captured.
For example, 48 kHz / 16-bit audio has a 24 kHz Nyquist frequency and a theoretical dynamic range of about 98 dB. 48 kHz / 24-bit audio has the same 24 kHz Nyquist frequency but a much larger theoretical dynamic range of about 146 dB. In recording, 24-bit capture is often useful because it gives more room for conservative levels and later processing. For final delivery, 16-bit can still be perfectly appropriate when properly dithered and mastered.
44.1 kHz vs 48 kHz
44.1 kHz is historically tied to CD audio and remains common for music delivery. It has a Nyquist frequency of 22.05 kHz, which covers the generally accepted upper range of human hearing with a small amount of extra room. If your final release is CD-style music delivery or a streaming master that will be encoded from a music-focused workflow, 44.1 kHz can still be a sensible target.
48 kHz is the normal choice for video, film, broadcast, games, and many modern production pipelines. It has a Nyquist frequency of 24 kHz and aligns better with common camera, video editing, broadcast, and game audio expectations. If your audio will be synced to picture, embedded in a video timeline, delivered to a broadcaster, or placed in a game engine, 48 kHz is usually the safer default. The audible difference between 44.1 and 48 kHz is rarely the deciding factor; workflow compatibility usually matters more.
48 kHz vs 96 kHz
96 kHz doubles the sample rate of 48 kHz and raises the Nyquist frequency from 24 kHz to 48 kHz. That extra ultrasonic bandwidth can help in specific workflows, especially sound design, pitch shifting, extreme time stretching, scientific capture, or plug-in chains that behave better at higher internal rates. Some engineers also prefer 96 kHz because anti-aliasing filters can be placed farther away from the audible band.
The cost is straightforward: 96 kHz uses about twice the uncompressed audio data of 48 kHz at the same bit depth, channel count, and duration. It also increases CPU load for many plug-ins and can reduce track counts on limited systems. For ordinary dialogue, podcasts, video editing, music delivery, and most location recording, 48 kHz is usually more practical. Use 96 kHz when the project benefits from the extra processing margin, not because the number is larger.
Should I record at 192 kHz?
192 kHz gives a 96 kHz Nyquist frequency, which is far beyond normal human hearing. It can be useful for specialist sound design, ultrasonic capture, measurement work, and DSP research where material may be slowed down dramatically or analyzed outside the audible band. It can also be useful when a specific production brief requires it.
For normal recording, mixing, video, podcasting, and music delivery, 192 kHz is usually unnecessary. It creates large files, increases disk bandwidth, pushes CPU load higher, and may expose limits in interfaces or plug-ins. A well-recorded 48 kHz or 96 kHz session will usually beat a poorly managed 192 kHz session. Choose 192 kHz only when the extra ultrasonic range has a clear job.
How sample rate affects file size
For uncompressed PCM audio, file size scales directly with sample rate. The formula is sample rate × bit depth × channels × duration ÷ 8. A mono file uses one channel. A stereo file uses two channels. A surround or multichannel recording multiplies the size by the number of channels. WAV and AIFF headers add a tiny amount of metadata, but the audio payload is dominated by this formula.
As a practical example, 48 kHz / 24-bit stereo audio uses 2,304,000 bits per second, or 288,000 bytes per second. Five minutes of that audio is about 86.4 MB in decimal units. If you move the same project to 96 kHz while keeping 24-bit stereo and the same duration, the file size is roughly 172.8 MB. If you also increase channel count, the storage requirement rises again. This is why sample-rate decisions matter for long recordings, multitrack sessions, portable recorders, and shared project drives.
