Updated October 2026 · Audio basics for beginners
Latency is the delay between a sound going into a system and coming back out. It comes from the time the computer and the interface need to process the audio. Latency makes monitoring feel strange and makes singing or playing hard. Reduce it with a smaller buffer size and a suitable driver, and monitor the microphone directly where possible.
Many beginners notice an odd effect when they record: they speak into the microphone and hear their own voice a moment later in the headphones. The delay is called latency, and it can make the recording experience strange, tiring or impossible for singers and instrumentalists. The good news is that latency has clear causes, and most of them can be reduced with a few settings and habits.
This post explains where latency comes from, why it matters for different kinds of work, how to reduce it and when a delay is simply part of the process.
Latency is the time it takes for a sound to travel through the system: from the microphone into the interface, into the computer, through the software and back out to the headphones. Each step takes a small amount of time, and the total is the delay you hear. The computer needs time to collect audio into blocks before it processes them, and that collection is one of the largest causes of delay.
Latency is not caused by the distance the sound travels, which is negligible. It comes from the processing time of the system.
For speech, a small delay is annoying but often manageable. For singing and playing, a delay makes it hard to stay in time, because the voice or the instrument seems to lag behind what you hear. Many singers find that even a few milliseconds of delay is distracting, and the effect can make the performance feel out of sync.
Latency can also affect editing and overdubbing, where a recording must line up with existing material. Compensation tools exist for that, but it is simpler to avoid the problem during recording.
The buffer size is the amount of audio the computer collects before processing it. A smaller buffer gives lower latency, but it demands more from the computer, and it can cause crackles if the system cannot keep up. A larger buffer is more stable, but it adds delay. The usual approach is to use the smallest buffer that runs without crackles, and to increase it only when the system needs it.
Your audio driver also affects latency. Dedicated drivers from the interface maker are usually better than general drivers, because they are designed for low delay.
Many interfaces offer direct monitoring, which sends the microphone signal straight to the headphones without passing through the computer. That removes most of the latency from monitoring. The recording still goes through the computer, but you hear the live voice without the delay. Direct monitoring is one of the most useful features for singers and speakers.
When you use direct monitoring, remember that you are hearing the input, not the recording. Any effects added in the software may not be heard during monitoring.
Some delay is unavoidable, especially with network connections, video calls and some streaming setups. A call or stream may add seconds of delay, which is a property of the network and the platform rather than the audio setup. Know which kind of delay you are dealing with before trying to fix it.
Video calls add their own delay, which comes from the network, the server and the software on each side. Speakers on a call often hear each other a fraction of a second late, and that delay can make conversation feel awkward. Keep the audio path short, use headphones to avoid echo and speak with a small pause between turns. None of these reduce the network delay itself, but they make the conversation easier to manage.
If the delay becomes a problem during a recorded call, record locally on each side. A local recording is not affected by the network, and the two tracks can be aligned later.
A simple check for latency is to clap once and listen for the echo in your headphones. If you hear a faint second clap, the monitoring path is adding delay. Try the same test with direct monitoring turned on, and compare. The difference is often clear enough to decide the setting in a few seconds, without any measurement.
Settings and menus for latency differ between interfaces, computers and software, and the exact steps change. The principles are stable: smaller buffers reduce delay, and direct monitoring removes most of it. Check the documentation for your own setup.
A short test shows whether the delay is acceptable. The Vocal Checker shows clipping, noise and level problems in a recording you own. The Tunedzy Go page covers cleaning a phone recording. For the wider setup, read the beginner basics post in this series.
Spot clipping, noise, rumble and level problems. Runs in your browser, nothing is uploaded.
The system takes time to process the audio. Use direct monitoring or a smaller buffer.
Only if the system can keep up. Too small a buffer causes crackles.
The recording is usually aligned, but it can make performing harder. Direct monitoring helps.
Yes, from the network and the platform. That delay is different from interface latency.
