What is bit depth?

The short answer

Digital audio is a long list of numbers. Many times a second, your system measures how loud the sound is at that instant and writes down a value. Each of those measurements is a sample. Bit depth is how many bits, the ones and zeros a computer counts in, are used to store each sample. It sets how precisely the loudness of every sample can be written down.

The two values you will see almost everywhere are 16-bit and 24-bit. A bigger number of bits means a bigger set of possible values, so the loudness can be captured more finely and the faint background hiss of the format itself sits lower. Bit depth is not about how bright or bassy something sounds. It is about resolution: how many rungs there are on the loudness ladder that every sample has to land on.

How bit depth works

The catch with storing loudness as a whole number is that a sample cannot sit just anywhere. It has to snap to one of a fixed set of allowed levels, the way a staircase only lets you stand on a step and not in the space between two steps. The number of available levels is set by the bit depth: each extra bit doubles them. So 8-bit gives 256 levels, 16-bit gives 65,536 levels, and 24-bit gives 16,777,216 levels. That doubling is why a couple of extra bits makes such a large difference.

When a smooth incoming sound is forced onto the nearest available level, the tiny amount it gets nudged is called quantization error. With only a few coarse levels, the stored signal is a chunky staircase that misses the real curve by a lot, and that error is heard as a faint gritty noise. With many fine levels, the staircase hugs the curve so closely that the error becomes vanishingly small. More bits do not make the sound louder or add detail that was not there. They simply reduce how much each sample gets rounded off.

Bit depth, the noise floor, and dynamic range

The practical payoff of bit depth is the noise floor: the level of the quiet hiss that sits underneath everything, below which real sound gets lost. Finer levels mean smaller rounding, and smaller rounding means a lower noise floor. As a rough rule, every bit buys you about 6 decibels, the standard unit for how loud one thing is compared to another. So 16-bit reaches roughly a 96 dB gap between the loudest sound it can hold and its noise floor, and 24-bit reaches about 144 dB in theory.

That gap between the loudest undistorted sound and the noise floor is the dynamic range. It is worth being honest about the 24-bit number: in the real world the analog parts of an interface add their own hiss, so most converters top out around 120 dB of usable dynamic range no matter how many bits the file holds. The extra bits are still valuable, but as working room rather than as 144 dB of pristine silence you will ever actually hear.

16-bit versus 24-bit: which and when

The two are tools for different jobs, so the honest answer is to match the bit depth to what you are doing rather than always reach for the bigger number.

• Recording and working. Capture at 24-bit. The larger dynamic range lets you record with plenty of room below the point where the signal distorts, so a sudden loud moment does not ruin the take, and quiet detail still sits well above the noise floor. That safety margin is called headroom, and it is the main reason to prefer 24-bit while you capture and edit.
• Final delivery and listening. 16-bit is plenty for a finished file. Its 96 dB range is wider than almost any listening environment can reveal, which is exactly why the audio CD standard is 16-bit. Once the loud and quiet parts are settled, the extra bits no longer earn their file size.
• Voice, calls, and streaming. Whatever your device offers is fine. Many interfaces run at 24-bit internally by default, and you will not hear a downside to leaving it there.

Bit depth is not sample rate or bitrate

These three get tangled together constantly, so it helps to pin down which axis each one controls.

• Bit depth is the loudness axis: how finely each sample is measured.
• Sample rate is the time axis: how many samples are taken each second. For how the two work together, see sample rate vs bit depth.
• Bitrate is the total data per second. For uncompressed audio it is simply sample rate times bit depth times the number of channels, so bit depth is one ingredient of bitrate rather than the same thing.

Bit depth in patchd

In patchd, a Windows virtual audio mixer, bit depth is set by the format your interface or Windows sound device is running at, not by a dial inside the app. patchd works with whatever your device provides and mixes it in real time, so the amplitude resolution and the noise floor you get come from that chosen format. The practical takeaway is the same as in any capture setup: prefer 24-bit while recording so quiet detail has far more headroom than 16-bit would leave.

The engine is real time and runs over ASIO and WASAPI Exclusive, the low-overhead Windows audio paths, adding only single-digit milliseconds of its own delay at a small ASIO buffer. That means you can carry your device format through a full effects rack, a DSP chain of a noise gate, EQ, and compressor, without the extra resolution of 24-bit costing you responsiveness. For how much delay stacks up in a monitoring path, see what is monitoring latency.

The takeaway

Bit depth is how many bits store each sample, which sets how many loudness levels every sample can land on. More bits mean finer resolution, less quantization error, and a lower noise floor, which together add up to more dynamic range. 24-bit gives you generous headroom for capturing and editing, while 16-bit is more than enough for a finished, delivered file. It is a resolution setting, not a quality knob you crank for its own sake.

patchd is pre-launch, with a free effects rack on every channel and a real-time engine that works with whatever bit depth your device is set to. If you want a Windows mixer that keeps your audio clean and your latency honest, join the waitlist and get notified when it ships. The paid Studio tier is $39.99 per year, and the engine is identical on both tiers.