RECORDING & PERFORMANCE · 7 MINUTE READ
Audio interfaces and buffer size: reduce delay without causing dropouts
Audio settings are a trade-off between response and processing time. The right setting is the one that lets your current project run reliably while the performer hears a usable monitoring signal. A smaller buffer is not a quality upgrade for every task.
Choose the route before changing the numbers
Confirm which physical input carries your microphone or instrument and which output feeds your headphones or speakers. A mono microphone usually needs the matching mono input rather than an arbitrary stereo pair. Select the intended audio device in HyperION’s Audio I/O controls, then check the track input, record arm, monitoring and main output.
Use the interface manufacturer’s supported driver where appropriate; on Windows this commonly means its ASIO driver. macOS and Linux use their platform-specific audio paths. The names and available choices vary by device and operating system. Do not follow a Windows-only screenshot as though it describes every computer.
Understand what buffer latency actually measures
A buffer is a block of audio processed together. Its duration in milliseconds is approximately samples divided by sample rate, multiplied by one thousand. At 48 kHz, 256 samples lasts about 5.33 milliseconds. That is one buffer’s duration, not the complete delay between playing an instrument and hearing it.
Input and output buffering, driver safety buffers, converters and plug-in processing can add more delay. A plug-in that uses look-ahead or other delayed processing may remain noticeable even after lowering the device buffer. Keep the sample rate consistent across the project and device, and change one setting at a time so comparisons are meaningful.
Use different starting points for recording and mixing
For recording, try a moderate-low buffer supported by your system and test the actual performance. Values such as 128 or 256 samples are useful starting experiments, not guarantees. If clicks or dropouts appear, increase the buffer before changing several unrelated settings. During a heavy mix, a larger buffer can give the computer more processing time without changing the musical content.
If your interface offers direct monitoring, it can let you hear the incoming signal without waiting for the full software path. Be careful when direct and software monitoring are both active: hearing both paths can create an echo or coloured sound. Choose the monitoring method deliberately and check how the performer hears software effects.
Diagnose a dropout with a small test
Save the project and test a short recording with one source and no unnecessary effects. If that works, reintroduce demanding plug-ins one at a time. Oversampling, long reverbs and some mastering processors can increase the processing load or latency. Watch whether the problem appears only in the large project or also in the simple one.
Record the operating system, interface and driver version, sample rate, buffer size and the steps that reproduce the issue. Check the resulting file: a monitoring interruption and a damaged recording are different problems. Keep listening levels sensible and use headphones when an open microphone could feed back into speakers. Return to a known stable configuration before recording an important take.
HyperION workflow reference: Official user manual. These are conceptual steps; consult the manual for your installed version for exact controls.