What Is Audio Latency? Complete Guide

Audio latency is the delay between the moment you produce a sound—by playing a note, speaking into a microphone, or triggering a sample—and the moment you hear it back through your speakers or headphones. Even a fraction of a second of latency can feel wrong to musicians and sound unnatural in live situations.

Latency is measured in milliseconds (ms). A latency of 10ms means a quarter-second tenth of a second. Your ear typically notices latency above 20ms in direct monitoring scenarios, though some sensitive listeners detect it at 10–15ms.

How Latency Is Measured

Latency is measured along a path: from input (a microphone or instrument) through your audio interface, DAW, and any processing, back out through your speaker or headphone output. This full path is called roundtrip latency.

Roundtrip latency is the sum of several smaller delays:

• Audio interface latency (the hardware’s inherent delay): 2–5ms for good interfaces
• Buffer latency (time your computer waits before processing audio): 1–10ms depending on buffer size
• Software latency (your DAW and plugins): 0.5–5ms or more
• Monitoring latency (if you’re monitoring through your DAW instead of directly through the interface): adds another 1–10ms

The total you experience is all of these added together. Calculate your roundtrip latency with a latency measurement tool by recording a click or tone and measuring the delay in your DAW.

Roundtrip vs. Input vs. Output Latency

Roundtrip latency is the full there-and-back delay. But latency has layers:

Input latency is the delay from the moment sound enters your microphone to when your DAW receives the audio signal. This matters for live tracking and monitoring.

Output latency is the delay from when your DAW plays audio to when it comes out of your speakers. This matters less for tracking but is noticeable during playback of synced video or in live performance.

Understand the differences between input and output latency to know which one you need to control in your specific workflow.

Why Musicians and Producers Care About It

If you’re recording a vocal track while listening to a beat, latency in your monitor mix creates a feeling of lag. Your brain expects the sound you produce to come back instantly; a delay of even 20ms feels out of sync, like you’re singing ahead of the beat or trailing behind it.

The same issue affects musicians playing virtual instruments. If you hit a key on your MIDI keyboard and the synth sound takes 50ms to reach your ears, it throws off your sense of timing. You’ll unconsciously rush or drag to compensate, and your performance suffers.

For mixing and post-production, latency is usually not a concern because you’re not relying on live monitoring feedback. You can work with higher latency settings to save CPU power.

Where Latency Comes From in Your System

Latency is baked into every stage of audio processing. Your audio interface has inherent hardware latency. Your operating system’s audio driver (WASAPI on Windows, Core Audio on Mac, ALSA on Linux) adds its own buffer delay. Your DAW buffers audio to smooth processing. Each plugin or effect processes audio in blocks, adding a small delay.

On Windows, DPC latency—brief interruptions from the system kernel—can spike latency unpredictably even if your average buffer size is low. On Mac and Linux, latency is more predictable because the OS prioritizes audio differently.

How Much Latency Is Acceptable?

For mixing, mastering, and non-real-time work, latency below 100ms is fine. You’re not relying on immediate feedback.

For live vocal recording or playing virtual instruments, aim for below 20ms roundtrip. Below 10ms is excellent and barely noticeable. Between 10–20ms, most people don’t notice a problem. Above 30ms and timing feels consistently off.

For live performance or DJ work, keep it under 50ms. For video sync work (recording dialogue to picture), keep it under 30ms or you’ll see visible lip-sync drift.

Test your perception of latency thresholds with a latency perception test to understand what’s noticeable in your own hearing.

Examples of Latency in Different Scenarios

When recording a vocal to a backing track, 20ms latency means the vocalist hears their voice delayed by a fraction of a second compared to the beat. This creates a sensation of dragging or rushing and is immediately felt as unnatural.

When playing a MIDI keyboard synth in a DAW, latency adds a tactile delay between key press and sound. At 50ms, this feels sluggish, like playing a piano with sluggish action.

When using a headset mic in Zoom or Teams, latency determines how long a delay there is between someone speaking and the other person hearing them. Call latency is usually 100–300ms and feels conversational but noticeably delayed.

Test your audio latency with the interactive latency test tool to measure your specific system’s delay and see if it’s within acceptable ranges for your use case.

Frequently Asked Questions

Can I eliminate audio latency completely?

No. Some latency is inherent to digital audio processing—your computer has to buffer and process audio in blocks. The goal is to keep it below the threshold where humans notice it (around 10–20ms for real-time work).

Does Bluetooth add latency?

Yes, significantly. Bluetooth headphones typically add 40–500ms of latency depending on the codec (aptX Low Latency is faster than standard SBC). Wired headphones have negligible latency.

Is latency the same as lag?

Latency is the technical term for delay in audio transmission. Lag usually refers to delay in video or network communication, but the concept is similar—a gap between input and output.

Does latency affect audio quality?

Not directly. Latency doesn’t degrade audio quality; it just delays it. But the workarounds people use to reduce latency—smaller buffer sizes, CPU limitations—can sometimes degrade quality if not managed carefully.