Sample rate vs bit depth vs bitrate

The short answer

Recording sound means turning a smooth, continuous wave in the air into a list of numbers a computer can store. To do that, you take measurements: you check the signal's level over and over, very quickly, and write down each value. Every one of those measurements is a sample. All three of the terms people mix up are really just describing that one process from different angles.

• Sample rate is how often you measure, in samples per second. It is resolution along the time axis.
• Bit depth is how finely you can record each measurement, in how many distinct levels a value can land on. It is resolution along the amplitude (loudness) axis.
• Bitrate is how much data all of that adds up to per second. For an uncompressed recording it is simply the first two multiplied together, times the number of channels.

So sample rate and bit depth describe the picture you are capturing, and bitrate describes the size of the file that picture takes. Confusing them is like confusing a photo's width, its color depth, and its file size: related, but not the same thing at all.

Sample rate: resolution in time

Sample rate is the number of measurements taken every second, written in hertz (Hz), which just means "per second." A rate of 48,000 Hz, usually written 48 kHz, means the signal is measured forty-eight thousand times a second. The two rates you will see almost everywhere are 44.1 kHz (the CD standard) and 48 kHz (the standard for video, streaming, and most computer audio).

The reason the numbers are that high is that you need at least two measurements per wave cycle to capture a given pitch. That rule sets a hard ceiling: a sample rate can only capture frequencies up to half its value. At 48 kHz you can capture sound up to 24 kHz, which comfortably covers the roughly 20 kHz top of human hearing with a little room to spare. Going higher, to 96 kHz or 192 kHz, mostly buys headroom for processing rather than sound you can actually hear. For the full picture, see what is sample rate.

The plain takeaway: a higher sample rate means finer resolution in time and a higher top frequency captured, at the cost of more data. It has nothing to do with how loud or how clean each measurement is. That is a different axis entirely.

Bit depth: resolution in amplitude

Bit depth is how precisely each individual sample can be recorded. Every measurement has to snap to one of a fixed set of levels, and the bit depth sets how many levels there are. Each bit doubles the count, so 16-bit audio has 65,536 possible levels and 24-bit audio has more than sixteen million. Think of it as the fineness of the ruler you measure each sample against.

More levels means the recorded value sits closer to the true value, and the gap between them (heard as a faint background hiss called quantization noise) sits lower. In practical terms, bit depth sets your dynamic range: the distance between the quietest detail and the loudest peak you can hold before the signal clips. Every bit is worth about 6 dB of that range, so 16-bit gives you around 96 dB and 24-bit around 144 dB. That extra room is why recording and mixing are done at 24-bit: it lets you leave safe headroom below the ceiling without burying quiet detail in noise. For more, see what is bit depth.

Notice that this says nothing about how many measurements per second you take. You can pair any bit depth with any sample rate. They are two independent dials, one for time and one for amplitude, which is exactly why they should not share a name.

Bitrate: the resulting data rate

Bitrate is the amount of audio data that flows per second, measured in kilobits per second (kbps) or megabits per second (Mbps). For plain, uncompressed audio (the kind in a WAV file) it is not a setting you choose at all; it falls straight out of the other two:

bitrate = sample rate × bit depth × channels. A stereo 48 kHz, 24-bit recording runs 48,000 × 24 × 2 = 2,304,000 bits per second, or about 2,304 kbps. Change the sample rate or the bit depth and the bitrate moves with them, because it is just the total of everything you are storing each second.

The word gets confusing because it means something slightly different for compressed formats like MP3, AAC, or the codecs used by streaming and voice-chat apps. There, bitrate is a target you set, and the encoder throws away whatever detail it must to hit that size. A 128 kbps MP3 is not measuring sample rate times bit depth; it is a promise about file size, and the codec decides what to sacrifice. So for uncompressed audio bitrate is a result, and for compressed audio it is a budget. Either way it is about data, not about the signal's time or amplitude resolution. See what is audio bitrate for the full breakdown.

One picture, three axes

The cleanest way to see that these are three different things is to draw the signal being captured. Sampling lays a grid over the wave. The spacing of the vertical lines is your time resolution (sample rate), the spacing of the horizontal lines is your amplitude resolution (bit depth), and each stored sample snaps to the nearest crossing. The bitrate is just how much all those stored crossings weigh per second.

Once you can see the grid, the common mix-ups fall apart. A "higher bitrate" WAV is not inherently more detailed than a lower one; it might simply have more channels or a higher sample rate. And a crisp, detailed recording that has been squeezed into a low-bitrate stream can still sound worse than a modest one that was left uncompressed, because the bitrate budget forced detail to be discarded.

Which one should you actually care about?

It depends on whether you are capturing sound or delivering it, and the honest answer is that for most day-to-day use the defaults are already right.

• Recording and producing. Sample rate and bit depth are what matter, and 48 kHz / 24-bit is the safe, standard choice. The 24-bit depth gives you room to record without clipping; 48 kHz covers everything you can hear.
• Delivering and streaming. Bitrate is the dial that matters, because it trades file size against how much detail the compression keeps. Higher bitrate means a bigger file and fewer audible compression artifacts.
• Everyday listening and calls. You rarely touch any of them. Your device and your apps pick sensible values, and the numbers only surface when something is misconfigured.

Sample rate, bit depth, and bitrate in patchd

In patchd, a Windows virtual audio mixer, sample rate and bit depth describe how your audio is digitized (the time and amplitude resolution of the signal moving through it), and bitrate is simply the data rate that follows from them. patchd rides the format your audio device is already set to rather than resampling behind your back, and it reports the running clock so you can see the rate it is locked to at a glance.

In practice you almost never touch these beyond picking 48 kHz and 24-bit once, in Windows or your interface panel, and letting everything downstream match. Keeping a single sample rate across your apps avoids silent resampling and the clicks and drift that come with it. The real-time engine works on the samples at whatever rate you set, so the resolution you choose is the resolution it processes, cleanly, without adding a conversion step in the middle. For how the buffer around those samples affects delay, see what is audio buffer size.

The takeaway

Sample rate, bit depth, and bitrate are three views of one digitized signal, not three names for its quality. Sample rate is how finely you slice time; bit depth is how finely you slice amplitude; bitrate is how much data the two produce per second, whether that comes out as a plain multiplication for uncompressed audio or a budget you set for compressed audio. Get those three straight and most of the confusing advice about audio "quality" sorts itself out. For recording, reach for 48 kHz and 24-bit and stop worrying about the rest.

patchd is pre-launch, a real-time Windows mixer that follows your device's sample rate and bit depth instead of fighting them, and shows the clock it is locked to so nothing resamples in secret. If you want a mixer that keeps your audio at the resolution you chose, join the waitlist and get notified when it ships. The paid Studio tier is $39.99 per year, and the real-time engine is identical on both tiers.