Audio Latency Calculator – Buffer ÷ Sample Rate

Audio latency from buffer size is straightforward math. The formula is always the same: Latency (ms) = (Buffer Size / Sample Rate) × 1000. That’s it. Once you understand this, you can calculate latency for any setup instantly without needing a calculator.

A 256-sample buffer at 44.1 kHz: (256 / 44,100) × 1000 = 5.8 ms
A 128-sample buffer at 48 kHz: (128 / 48,000) × 1000 = 2.67 ms
A 64-sample buffer at 96 kHz: (64 / 96,000) × 1000 = 0.67 ms

The numbers in the denominator are Hz (cycles per second). 44.1 kHz = 44,100 Hz. 48 kHz = 48,000 Hz. And so on.

Why This Formula Works

When you set a buffer size in your DAW or audio interface, you’re telling the hardware “collect this many audio samples before processing and sending them out.” Your sample rate determines how many samples get captured per second.

If you capture 48,000 samples per second and process 256 samples at a time, each processing cycle takes 256 / 48,000 seconds. Convert to milliseconds: 0.00533 seconds × 1000 = 5.33 ms.

That’s the latency from the buffer alone. Your system also adds hardware conversion time (1–3 ms), driver overhead (2–10 ms), and OS scheduling (0–20 ms on Windows, less on Mac), but the buffer is the core calculation.

Real Latency at Every Common Setting

Here are the numbers you’ll encounter most often. Bookmark this for reference.

At 44.1 kHz:
32 samples = 0.73 ms
64 samples = 1.45 ms
128 samples = 2.9 ms
256 samples = 5.8 ms
512 samples = 11.6 ms
1024 samples = 23.2 ms

At 48 kHz:
64 samples = 1.33 ms
128 samples = 2.67 ms
256 samples = 5.33 ms
512 samples = 10.67 ms
1024 samples = 21.33 ms

At 96 kHz:
64 samples = 0.67 ms
128 samples = 1.35 ms
256 samples = 2.7 ms
512 samples = 5.4 ms
1024 samples = 10.7 ms

At 192 kHz:
64 samples = 0.33 ms
128 samples = 0.67 ms
256 samples = 1.33 ms
512 samples = 2.67 ms
1024 samples = 5.33 ms

Patterns to notice: Doubling the sample rate cuts latency in half. Doubling the buffer size doubles latency. This is why 256 samples at 96 kHz (2.7 ms) matches the latency of 512 samples at 48 kHz (10.67 ms)—almost.

Buffer Latency Is Only Part of the Picture

The formula calculates buffer latency only. Your actual round-trip latency includes:

Interface AD/DA conversion: 1–5 ms (hardware-dependent)
Driver overhead: 2–10 ms (driver quality matters hugely)
Operating system: 0–30 ms (Windows audio stack adds 10–30 ms; Mac adds less)
Plugins: 0–100+ ms (lookahead and delay-compensation in some plugins)
Wireless: 40–300 ms (Bluetooth adds massive latency)
Distance: 0.343 ms per foot (sound travel time, usually negligible)

A 256-sample buffer at 48 kHz calculates to 5.33 ms of buffer latency. Your actual round-trip might be 10–20 ms after accounting for interface and driver overhead. This is why measuring real latency with a loopback test is more accurate than calculating buffer latency alone.

How to Find Your Buffer and Sample Rate Settings

Every DAW displays these settings somewhere. Here’s where to look:

Ableton Live: Preferences → Audio → Look for “I/O Latency” (reported) and “Internal Audio Buffer Size” (your buffer)
FL Studio: Options → Audio Settings → Buffer Length (samples)
Logic Pro: Project Settings → Audio/MIDI → I/O Buffer Size
Reaper: Options → Preferences → Audio Device → Block Size (samples)
Pro Tools: Preferences → Playback Engine → Hardware Buffer Size
GarageBand: Not adjustable in GarageBand directly; uses Core Audio defaults

Once you know your buffer size and sample rate, plug them into the formula. If Ableton shows 512 samples and you’re at 44.1 kHz, that’s (512 / 44,100) × 1000 = 11.6 ms of buffer latency.

Your DAW will also show a total latency figure, which includes buffer latency plus some driver overhead. The calculated figure and the reported figure should be close (within 1–2 ms) but rarely identical.

Practical Latency Targets

Use these targets to decide what buffer size to shoot for:

Music production (recording): Under 10 ms feels responsive. At 44.1 kHz, that means 64–256 samples depending on your CPU. At 48 kHz, 128–256 samples. At 96 kHz, up to 512 samples is fine.

Mixing (no monitoring required): 20–50 ms is fine. You can use larger buffers (512–1024 samples) because you’re not performing. Lower buffer = less CPU load.

Gaming: 40–50 ms or lower is comfortable. Most gamers don’t notice up to 50 ms.

Live performance: Under 10 ms is essential. Musicians notice anything above 10 ms.

Video calls: 50–150 ms is acceptable. VoIP inherently has network latency that dwarfs interface latency.

Casual listening: Anything is fine. You won’t hear latency in passive listening.

The Buffer-to-Latency Conversion Table

If your DAW reports latency in milliseconds instead of samples, use this to convert:

At 44.1 kHz, every 44 samples = 1 ms
At 48 kHz, every 48 samples = 1 ms
At 96 kHz, every 96 samples = 1 ms

So if your interface spec says “5 ms hardware latency,” that’s roughly 220 samples at 44.1 kHz or 480 samples at 96 kHz.

Choosing Your Buffer: The CPU Versus Latency Trade-off

Smaller buffer = lower latency but higher CPU load. Larger buffer = higher latency but more stable performance.

Start with a buffer size your CPU handles without crackling or dropouts. For most modern computers and modest plugin loads, that’s 128–256 samples. Once stable, lower it by one step (128 → 96 → 64) until you hear glitches, then back up by one step. That’s your lowest stable point.

At that lowest stable buffer, check how low your latency is. If it’s 3–5 ms, you’re done. If it’s 10+ ms and you still want lower, bump your sample rate from 44.1 to 96 kHz (which cuts latency roughly in half) and test stability again. The extra CPU demand from 96 kHz might actually let you use a smaller buffer.

Most home studios settle on: 44.1 kHz at 256 samples (5.8 ms) or 48 kHz at 256 samples (5.33 ms). Professional studios tracking multiple instruments often use 96 kHz at 128 samples (1.35 ms).

Testing Your Calculated Latency

The formula gives you buffer latency. Your true round-trip latency is always higher. To measure it:

Run our Sound Latency Test (https://soundlatencytest.com/audio-latency-test/) and note the measured round-trip latency. Compare it against your calculated buffer latency. The difference tells you how much overhead your system is adding.

For example:
Calculated buffer latency: 5.8 ms
Measured round-trip latency: 12 ms
Overhead: 12 – 5.8 = 6.2 ms

If overhead is under 5 ms, your system is efficient. If it’s 10+ ms, driver or OS overhead is eating into performance—update drivers or check your audio settings.

For DAW-specific latency, check your DAW’s reported latency at your working buffer and sample rate. If it says 8 ms and your calculation says 5.8 ms, the extra 2.2 ms is driver + OS overhead.

Using the Formula for Different Workflows

If you record at 44.1 kHz but want to check latency at 96 kHz for comparison:

At 44.1 kHz, 256 samples = 5.8 ms
To get similar latency at 96 kHz: 256 × (44.1 / 96) = 117.75 ≈ 128 samples at 96 kHz = 1.35 ms

So 128 samples at 96 kHz is lower latency than 256 samples at 44.1 kHz. This is why higher sample rates can help when buffer size is limited by CPU.

The Quick Reference: When You Just Need a Number

Can’t remember the formula? Here’s the fast way:

One millisecond of latency at 48 kHz ≈ 48 samples
One millisecond of latency at 44.1 kHz ≈ 44 samples
One millisecond of latency at 96 kHz ≈ 96 samples

So if you want 2 ms of buffer latency at 48 kHz, aim for roughly 96 samples. Want 5 ms? Aim for roughly 240 samples.

---