How to Test Audio Latency: 5 Reliable Methods

You can’t fix what you can’t measure. Whether you’re tracking vocals, monitoring through headphones during recording, or streaming, testing your actual latency reveals where the delay is coming from and tells you if your setup is fast enough for the job.

Latency testing sounds technical, but it’s straightforward: send a signal out, record it coming back in, measure the gap. The time between the two is your round-trip latency—the total delay from input through your system back to output.

Here are the five methods that work, from easiest to most precise.

Method 1: Browser-Based Loopback Test (Easiest)

The simplest way to test latency is using a browser tool. No setup, no cables needed if you’re testing relative delay between devices.

How It Works:

Load our free online sound latency test in your browser. The tool plays a click through your speakers (or interface headphone output) and listens through your microphone (or loopback input). When you tap the screen in sync with the sound, the tool measures the delay in milliseconds.

Setup (2 minutes):

1. Connect your audio interface or headphones if you have one.
2. Go to soundlatencytest.com in your browser.
3. Click “Start Test” and tap your screen in sync with the click sound.
4. The tool reports your latency in ms.

When to Use This:

For quick checks on any device—laptop, phone, desktop. Great for Bluetooth latency testing too, since it measures round-trip in real-world conditions. You can test multiple devices and compare.

Limitations:

Measures relative latency, not absolute. Web Audio API adds a baseline buffer (typically 1024 samples or ~21ms at 48kHz) that the tool tries to account for but can’t fully eliminate. Use this as your “worst case” baseline.

Method 2: DAW Buffer Readout (Fastest for Recording)

Most DAWs display your current latency right in the settings. It’s the quickest way to know what you’re working with.

How It Works:

Your DAW calculates latency based on buffer size, sample rate, and driver info. Open audio settings and look for “Overall Latency” or “Round-Trip Latency.” The number you see is theoretical—it assumes your drivers report accurately.

Step-by-Step:

Ableton Live: Preferences > Audio > Look at the “Overall Latency” number. At 128 samples and 48kHz, this is typically 7.88ms.

FL Studio: Options > Audio Settings > ASIO Settings > “Buffer length.” FL Studio shows both buffer latency and total latency. At 512 samples and 48kHz, total is around 12ms.

Logic Pro: Settings > Audio > Devices > I/O Buffer Size. Logic shows the resulting latency for each buffer setting. 64 samples at 48kHz typically shows 13ms roundtrip on modern Macs.

Reaper: Options > Preferences > Audio > Device > Adjust block size and watch the latency readout update in real-time.

When to Use This:

During recording or mixing sessions. It’s reliable if your audio interface drivers report accurate latency, which most modern interfaces do. Good for testing after changing buffer size or driver.

Limitations:

Assumes driver honesty. Some older or generic audio interface drivers underreport latency—the number you see might be 5ms but actual latency is 12ms. If your recordings sound slightly out of sync even though the DAW says latency is low, test with Method 4 or 5 to verify.

Method 3: Physical Cable Loopback with Audacity (Most Accurate)

For precise round-trip latency measurement, you need a physical cable looping audio from output back to input, plus a tool to measure the gap.

What You Need:

• An audio interface with both line output and line input (or use onboard soundcard if available)
• A short cable (3.5mm or XLR depending on your interface)
• Audacity (free)

Setup (5 minutes):

1. Plug a cable from your interface’s audio output jack into an input jack.
2. Open your DAW and generate a click or tone on a track set to output through that interface.
3. Open Audacity and set it to record from that input.
4. Play the click in your DAW for 5 seconds while recording in Audacity.
5. Stop. You now have the original click and the looped-back click on the same timeline.

Finding the Latency:

1. Zoom in on the Audacity waveform.
2. Find where the original click starts and where the looped-back click appears.
3. Count the samples between the two peaks or use Audacity’s selection to measure. Use the formula: (Sample Offset / Sample Rate) × 1000 = Latency in ms.

For example: If the looped click is 2304 samples behind the original at 48kHz: (2304 / 48000) × 1000 = 48ms.

When to Use This:

When you need exact numbers for recording offset or calibration. Professional studios use this to dial in Pro Tools or other systems.

Limitations:

Requires a physical loopback cable and takes time to set up. Most useful if you have multiple setups and need to compare them precisely.

Method 4: The Tap Test (Subjective but Useful)

A simple method for feeling whether your latency is acceptable for real-time playing.

How It Works:

1. Open a metronome or click track in your DAW.
2. Arm a MIDI track with a virtual instrument (drum, synth, whatever responds quickly).
3. Turn on monitoring through your DAW (or direct monitoring if your interface supports it).
4. Play along with the click. Tap a key in sync with the click sound.
5. Ask: Does the note I trigger appear in time with the click, or is there a lag?

Interpreting Results:

Under 10ms: Feels natural, like playing acoustic instruments. Most musicians don’t notice it.

10–20ms: Slightly noticeable for experienced performers, but workable for most genres.

20–40ms: Clear lag. Uncomfortable for real-time playing, but you can adjust to it.

Above 40ms: Very obvious delay. Hard to play in time.

When to Use This:

When setting up a recording session to confirm whether your latency is low enough for the task at hand. Best for musicians evaluating their own comfort level.

Limitations:

Subjective. What feels acceptable to you might not to another musician. Different instruments feel different latencies differently—playing drums feels more responsive at 5ms, but synths can work at 15ms.

Method 5: LatencyMon (Windows, Advanced)

LatencyMon is a Windows-only tool that measures real latency at the operating system level, revealing DPC (Deferred Procedure Call) latency spikes caused by drivers.

How It Works:

LatencyMon runs in the background and measures how long the OS takes to process audio buffer interrupts. High DPC latency causes audio glitches and dropouts, which indirectly reveals latency problems.

Setup:

1. Download LatencyMon from resplendence.com.
2. Run it and let it collect data for 5–10 minutes while your DAW plays audio.
3. Check the “Interrupt to Completion Latency” line. Under 1000µs (1ms) is excellent; under 2000µs (2ms) is good; above 5000µs (5ms) suggests driver or hardware problems.

When to Use This:

When your DAW says latency is low but you hear clicks, pops, or dropouts. LatencyMon identifies which driver is the bottleneck (GPU drivers and WiFi drivers are common culprits).

Limitations:

Windows only. Doesn’t measure absolute latency but instead identifies OS-level issues that cause unpredictable latency spikes.

Putting It Together: Your Test Workflow

For a quick setup check:

1. Use Method 2 (DAW readout) to get baseline numbers.
2. Use Method 1 (browser test) to compare against expectations.
3. If numbers seem high, use Method 4 (tap test) to confirm whether it’s actually a problem for your use case.

For precise calibration:

1. Use Method 3 (cable loopback) with Audacity to get exact numbers.
2. Use those numbers to set recording offset or input compensation in your DAW.

For troubleshooting glitches:

1. Use Method 2 to check your current settings.
2. Use Method 5 (LatencyMon) to diagnose OS-level issues.
3. Update drivers and test again with Method 2.

Real-World Latency Targets

Acceptable latency depends on what you’re doing:

Tracking vocals or guitar: Under 10ms feels natural. 10–20ms is workable. Above 20ms gets in the way.

Monitoring MIDI instruments: Under 15ms. At 20ms+, the timing disconnect becomes clear.

Live streaming or video calls: Network latency dominates. Focus on network quality first (wired internet). Audio latency above 150ms round-trip makes conversation awkward.

Mixing: Latency doesn’t matter. You’re not playing real-time, so buffer size doesn’t affect your mix quality.

The fastest setup we’ve tested with proper drivers and dedicated audio interfaces achieves 3–8ms round-trip. Anything under 15ms is professional-grade.

---