Paritian

Developers

Audio Latency Calculator

Buffer latency in milliseconds, round-trip delay, and the distance you would have to stand back to feel the same thing.

Results

Latency of one buffer 2.6667 ms
Round-trip latency 6.8333 ms
Delay from the speakers to you 2.9155 ms
Everything added up 9.7488 ms
Same as standing this much further back 2.344 m
Samples of delay, round trip 328
How that feels to play through Fine — under 10 ms round trip
Buffer that would give 10 ms round trip 240

What this tool does

A buffer of 128 samples at 48 kHz is under three milliseconds, which sounds like nothing until you count it twice for input and output and add the converters. The useful comparison is distance: sound travels about a third of a metre per millisecond, so seven milliseconds of round trip feels like standing two and a bit metres further from the amp — which is why the number stops mattering long before it reaches zero.

Formula

latency = buffer samples ÷ sample rate · round trip = 2 × buffer + converters

Variables

SymbolMeaningUnit
bfBuffer size in samples
srSample rateHz
cvConverter latency, both waysms
dsHow far the speakers arem
LTLatency of one bufferms
RTRound-trip latencyms
ACDelay from the speakers to youms
TTEverything added upms
EQSame as standing this much further backm
SMSamples of delay, round trip
VDHow that feels to play through
BHBuffer that would give 10 ms round trip

Worked example

  • Buffer size in samples128
  • Sample rate48,000 Hz
  • Converter latency, both ways1.5 ms
  • How far the speakers are1 m
  • Latency of one buffer2.6667 ms
  • Round-trip latency6.8333 ms
  • Delay from the speakers to you2.9155 ms
  • Everything added up9.7488 ms
  • Same as standing this much further back2.344 m
  • Samples of delay, round trip328
  • How that feels to play throughFine — under 10 ms round trip
  • Buffer that would give 10 ms round trip240

Limitations

  • The result is an estimate based only on the values you type. Real situations often include factors this calculator does not know about.
  • The calculation runs at full precision and only the display is rounded. If you copy an intermediate value and retype it, small differences can appear.