In music production, latency kills vibe. When you hit a note and hear it back 50+ ms later, your brain and hands get out of sync. MIDI timing feels sluggish, vocal phrasing sounds off, and guitar tracking becomes a guessing game. Getting latency under 10 ms transforms recording from frustrating to fluid.
Why Audio Latency Matters in Music Production
Recording is real-time feedback. When you’re singing a vocal, playing guitar, or programming drums in real-time, you’re listening to your own performance through your DAW’s mix. If that feedback is delayed, you compensate instinctively—singing ahead or playing late to match what you’re hearing. When you play back the recording, it’s out of sync with the click track.
MIDI is especially sensitive. When you play a keyboard or guitar controller into your DAW, the note should trigger sound almost instantly. At 50 ms latency, the sound lags behind your finger movement, and you lose the feel of live performance. Your timing instinct fights the delay, and takes suffer.
Even in mixing, latency compounds when plugins add processing. If your audio interface adds 5 ms, your DAW adds 3 ms, and your reverb plugin adds 4 ms, and you’re monitoring through a headphone amp with 2 ms more delay, that’s 14 ms round-trip. Add a click track, and everything shifts slightly out of sync.
Acceptable Latency for Different Recording Scenarios
For recording vocals with minimal effects (just reverb or delay), under 20 ms round-trip is acceptable. Most singers can adapt to latency in that range, though 10 ms or lower is noticeably better.
For recording live guitar or bass (with amp amp simulation or effects), stay under 15 ms. Guitar players are particularly sensitive to latency because their hands feel and react to the vibration and string response. Artificial latency throws off that tactile feedback.
For MIDI and real-time programming (drums, synths, programming a drum machine), target under 5 ms. At 10–20 ms, timing feel degrades. At 30+ ms, humanization and swing become impossible—your playing will sound stiff and late.
For vocal harmonies and overdubs (stacking takes), 20–30 ms is tolerable because you’re syncing to the click and prior takes, not your own real-time feel.
For mixing and effects processing (not recording), latency doesn’t matter if your DAW compensates automatically. Most modern DAWs delay visual tracks to match plugin latency, so mixing stays in sync.
Understanding Buffer Size and Latency
Buffer size is the primary lever for latency control. The buffer is the amount of audio data your audio interface holds before sending it to your DAW for processing. Smaller buffer = lower latency = higher CPU load.
At 48 kHz sample rate:
- 64-sample buffer = 1.33 ms interface latency
- 128-sample buffer = 2.67 ms
- 256-sample buffer = 5.33 ms
- 512-sample buffer = 10.67 ms
- 1024-sample buffer = 21.3 ms
Add your DAW’s engine latency (typically 2–5 ms) and your interface’s AD/DA conversion latency (typically 1–3 ms), and you get your total round-trip latency. A 256-sample buffer becomes about 8–10 ms total.
The constraint: smaller buffers demand more CPU power per audio cycle. At 64 samples, your CPU must complete all processing in 1.33 ms. If a plugin or your DAW’s engine can’t keep up, you get dropouts and crackling. Most home setups can’t sustain 64-sample buffers without degradation.
The compromise: use the smallest buffer size your system can sustain without dropouts. For most modern computers, that’s 128–256 samples. Test during a session—if you hear clicks or pops, increase the buffer size incrementally until it clears.
How to Measure Round-Trip Latency in Your DAW
Set up a loopback test. Use a round-trip latency tester or run the test manually using your DAW and a microphone.
Manual test: Record a sharp click or tone into your DAW’s input, route it through a short-delay effect (5–10 ms), and monitor it back through your interface’s output or headphones. Capture the original click and the delayed output with a microphone placed near your monitors or headphones. Measure the gap in your DAW’s waveform view. That’s your round-trip latency.
Automated test: Use a latency measurement plugin (many are free). They generate a test tone, route it through your I/O, and measure the delay. Run a DAW latency test to establish your baseline.
Note your result and write it down. Many DAWs have a “Latency Compensation” or “Plugin Delay Compensation” setting—enter your measured latency there. The DAW will automatically delay your visual track display to stay in sync with your audio processing.
Latency Compensation and Monitoring Techniques
Modern DAWs (Ableton, Logic, Reaper, Studio One) include automatic latency compensation. When you record a track, the DAW delays the visual waveform to match the round-trip latency you measured. Playback stays in sync with the click and other tracks.
For real-time monitoring, use your audio interface’s hardware mixer if it has one. Route your input directly to your output without going through your DAW. This bypasses all DAW latency and gives you zero-latency monitoring while you record. The recorded track will still be in sync because the DAW compensates for interface latency.
Click track offset: If your click track sounds late relative to your playing, you’re hearing the combined latency of your monitoring chain plus the click. Some DAWs let you offset the click track earlier by a few milliseconds to align with your actual latency. In Ableton Live, adjust the “Metronome Latency Compensation” setting.
Tape machine modeling: Some plugins emulate analog tape machines, which have inherent latency. If you’re recording to a plugin like Waves tape or FabFilter Saturn (in tape mode), expect 10–20 ms of added latency from the plugin. Understand that in your tracking setup.
Direct monitoring (if your interface has it): Connect your input directly to your headphone output on the interface, bypassing your DAW entirely. You’ll hear zero-latency monitoring, and the DAW records the performance cleanly. This is the professional standard in most studios.
Frequently Asked Questions
Should I record at lower buffer sizes than I mix at?
Yes. Record at the smallest buffer size that’s stable (128–256 samples). Once you’ve tracked everything, switch to a larger buffer (512–1024 samples) for mixing. Larger buffers reduce CPU load during the mixing phase when you’re using heavy plugins.
Does 96 kHz sample rate reduce latency?
No. At 96 kHz, 256 samples equals 2.67 ms—half the latency of 48 kHz. But it also demands twice the CPU power. Stick to 48 kHz for recording—it’s the industry standard and offers the best balance of latency and stability.
My DAW shows 50 ms latency in settings. Is that acceptable?
That’s too high for recording real-time performances. 50 ms is typical for mixing or non-real-time work. If you’re tracking vocals or MIDI, reduce your buffer size or check your plugin chain for heavy processors that add latency.
Why does latency increase mid-session?
Latency can spike if your CPU load increases (running more plugins, streaming, background apps). Quit unnecessary apps, freeze non-essential tracks, and render heavy effects to audio to free up processing power.
Can I reduce latency by disabling plugins?
Yes. Each plugin adds latency. Bypass (don’t delete) real-time plugins during tracking, then re-enable them for mixing. Most DAWs have a “record-only” or “monitoring” setting that lets you disable plugins while still recording.
Dalton is an audio testing and latency optimization writer at SoundLatencyTest. He focuses on audio latency analysis, sound delay testing, recording performance, and audio troubleshooting tools for producers, gamers, streamers, musicians, and audio engineers.