Samples and milliseconds
Convert a sample count into time, or a time offset into samples, at the exact sample rate you select.
Did this calculator give you what you needed?
What could be better?
The sample rate is the conversion factor
At 48 kHz, 48 samples last 1 millisecond and 48,000 samples last 1 second. At 44.1 kHz, the same sample count represents a different duration.
Fractional sample results
A requested time may fall between discrete samples. The calculator shows the exact result and the nearest whole sample so you can match the precision your editor accepts.
Buffer sizes in milliseconds
The buffer setting in an audio interface is a sample count, and this is the conversion that turns it into the delay you actually feel. At 48 kHz a 128-sample buffer is 2.67 ms and a 512-sample buffer is 10.67 ms. At 44.1 kHz the same numbers are 2.9 ms and 11.6 ms.
One caveat worth knowing before you blame the buffer: round-trip latency is larger than a single buffer. The signal is buffered on the way in and again on the way out, and the converters add their own delay, so a 128-sample setting typically measures somewhere near three times its own length by the time a sound has been through the computer and back.
Frequently asked questions
How do I convert my buffer size to latency?
Enter the buffer in samples and read the milliseconds. At 48 kHz, 64 samples is 1.33 ms, 128 is 2.67, 256 is 5.33 and 512 is 10.67. That figure is one buffer, not the round trip, so treat it as the number that changes when you move the slider rather than as the delay you will measure.
Why does my editor ask for samples instead of milliseconds?
Because a sample is the smallest unit the file actually has, and a nudge in samples is exact. Milliseconds have to be rounded to the nearest sample anyway, so working in samples avoids a rounding step and keeps two takes that should line up phase-aligned.
What happens to my offsets if the sample rate changes?
They move. A 1,000-sample offset is 22.68 ms at 44.1 kHz and 20.83 ms at 48 kHz, so an alignment copied between projects at different rates drifts by nearly 2 ms. Convert through time rather than carrying sample counts across, or the error compounds along the timeline.
Which sample rate should a project use?
Match the delivery. Music released to streaming and CD is 44.1 kHz; anything going to video, film or broadcast is 48 kHz, because that is what the video standards use. Higher rates are a workflow decision about processing headroom, not a requirement of the format you deliver.
What if the conversion lands between two samples?
It often does, and the page shows both the exact figure and the nearest whole sample. At 44.1 kHz a round 10 ms is 441 samples exactly, but 7 ms is 308.7, so something has to round. Take the whole number for an editor that only accepts integers, and keep the exact value if you are chaining further calculations.
More tools
Estimate BPM from the average gap between your taps.
Delay timeConvert BPM to straight, dotted and triplet times.
Reverb timingBuild tempo-synced pre-delay and decay spans.
Audio file sizeEstimate PCM or constant-bitrate audio data.
Semitone calculatorMeasure or apply a musical frequency interval.
Speaker impedanceCombine two to four speaker loads in series or parallel.