Audio Buffer Size & Round-Trip Latency Calculator
Calculate audio interface round-trip latency (RTL in milliseconds), buffer delay, converter filtering overhead, and physical acoustic distance equivalent across sample rates and buffer sizes.
1. Buffer Size & Sample Rate
2. Latency & Acoustic Equivalent
Understanding Audio Latency, Buffers & The Speed of Sound
Audio latency is the time delay between sound entering an audio interface and sound exiting the speakers or headphones:
- The Buffer Calculation Formula: Pure buffer latency is calculated as Buffer Time (ms) = Buffer Samples / Sample Rate × 1000. At 48 kHz with a 128-sample buffer, one-way buffer delay is 2.67 ms.
- Total Round-Trip Latency (RTL): RTL includes the input buffer + output buffer + AD/DA converter chip filter delay + USB/Thunderbolt driver transmission packet buffers.
- Acoustic Distance Comparison: Sound travels through air at roughly 1.13 feet per millisecond (343 m/s). An RTL of 6 ms creates the exact same acoustic time delay as standing 6.8 feet away from a guitar amplifier or stage floor monitor wedge.
- Tracking vs. Mixing Buffer Strategy: When recording live vocals or virtual instruments, set buffers to 64 or 128 samples (RTL under 10 ms to prevent vocal phasing in headphones). When mixing with heavy plugin chains, increase buffers to 512 or 1024 samples to eliminate CPU audio dropouts.
Worked examples: buffer size against round-trip latency
The tool works out the delay of one buffer as the buffer size divided by the sample rate. The round trip is two buffers, one on the way in and one on the way out, plus a fixed overhead for the interface and its driver. The first table uses the tool's starting settings of 48 kHz and a USB interface at 2.5 ms, and goes through every buffer size the tool offers. The rating column uses the tool's own bands: up to 6 ms is ultra low, up to 11 ms is good, up to 18 ms is noticeable, and anything above is high delay.
| Buffer | One buffer | Round trip | Rating |
|---|---|---|---|
| 32 samples | 0.67 ms | 3.83 ms | Ultra low |
| 64 samples | 1.33 ms | 5.17 ms | Ultra low |
| 128 samples | 2.67 ms | 7.83 ms | Good |
| 256 samples | 5.33 ms | 13.17 ms | Noticeable |
| 512 samples | 10.67 ms | 23.83 ms | High delay |
| 1024 samples | 21.33 ms | 45.17 ms | High delay |
| 2048 samples | 42.67 ms | 87.83 ms | High delay |
The overhead is a larger share of the total at small buffers. At 32 samples the two buffers come to 1.33 ms and the overhead supplies the other 2.5 ms. The second table holds the buffer at 128 samples and changes only the interface.
| Interface | Overhead | Round trip | Rating |
|---|---|---|---|
| Thunderbolt or PCIe | 1.2 ms | 6.53 ms | Good |
| USB 2.0 or USB-C | 2.5 ms | 7.83 ms | Good |
| Class-compliant USB | 5.0 ms | 10.33 ms | Good |
The sample rate works the same way. At 96 kHz, 256 samples is 2.67 ms for one buffer and 7.83 ms round trip, which is the same round trip as 128 samples at 48 kHz with the same 2.5 ms overhead.
Working Vocabulary
Key digital audio and latency performance terms:
- Round-Trip Latency (RTL). The total elapsed time for an audio signal to travel from an analog input through AD conversion, DAW processing, DA conversion, and back out to analog speakers.
- Buffer Size. The number of audio sample blocks held in computer memory before being processed by the CPU, trading latency for system stability.
- Sample Rate. The number of discrete audio snapshot measurements taken per second (e.g. 48,000 samples per second for 48 kHz).
- Direct Monitoring. An audio interface feature that routes microphone inputs straight to headphones via hardware analog circuitry with zero computer buffer latency.