DAW Latency: Buffer Size vs Sample Rate

By Marcus Chen · October 9, 2026

Understanding Audio Latency

Every digital audio workstation introduces latency — the delay between playing a note and hearing it back. For recording, anything above 10 ms becomes noticeable and distracting. For mixing, latency is less critical because the DAW compensates automatically during playback.

How Buffer Size Affects Latency

The audio buffer is a chunk of samples your computer processes before sending audio to your interface. Smaller buffers mean less latency but more CPU load. At 44.1 kHz sample rate, a 128-sample buffer gives you about 2.9 ms of one-way latency. At 96 kHz, the same 128 samples takes only 1.3 ms.

Common Buffer Settings

64 samples: Tracking with virtual instruments (2.9 ms round-trip at 44.1 kHz on modern interfaces)
128 samples: General recording with software monitoring
256 samples: Overdubbing with moderate plugin chains
512-1024 samples: Mixing and mastering with heavy processing

Sample Rate Impact on Latency

Higher sample rates process the same buffer size faster because each sample represents a shorter time interval. At 96 kHz, a 128-sample buffer is processed in 1.33 ms versus 2.90 ms at 44.1 kHz. However, doubling the sample rate roughly doubles your CPU load, since the computer processes twice as many samples per second.

The Round-Trip Latency Formula

Total latency = (buffer size / sample rate) x 2 + converter latency + driver overhead. Converter latency (analog-to-digital and back) typically adds 1-2 ms. Driver overhead varies: ASIO on Windows adds 1-3 ms, CoreAudio on macOS adds less than 1 ms.

Practical Settings by Task

Vocal tracking: 64-128 samples at 48 kHz — prioritize lowest latency
MIDI programming: 128-256 samples — moderate latency acceptable
Mixing: 512-1024 samples — latency does not matter, maximize processing headroom
Mastering: 1024+ samples — maximum processing power, no real-time monitoring needed

Low-Latency Monitoring Workarounds

Many interfaces offer direct monitoring — routing the input signal directly to the output before it enters the computer. This gives zero-latency monitoring of the source, but you hear the dry signal without plugins. For vocalists who need reverb in their headphones, add a hardware reverb unit or use the interface built-in DSP effects.