k2k audio Sound enters and leaves k2k here. Like plugging a cable into a mixing desk — this is where you load your raw material and send the finished work back out.
Load Audio
What it does Opens an audio file from disk (WAV, AIFF, or FLAC), runs spectral analysis on it, and feeds the result into your node graph as one output for mono files or two outputs (Left and Right) for stereo.
When you’d reach for it You have a recording, a sample, or a rendered stem on your hard drive and you want to bring it into k2k to start working on it.
Quick example
- Drop a Load Audio node onto the graph — it shows an empty drop zone.
- Drag a vocal stem (
lead_vocal.wav, 48 kHz stereo) onto the drop zone. - Leave FFT Size at 4096 and Normalize off (the stem is already gain-staged).
- Two outputs appear — Left and Right — ready to feed into a Transient Extractor or Slicer downstream.
Parameters
| Parameter | What it controls | Range | Sweet spot hint |
|---|---|---|---|
| Normalize | Scales the audio so the loudest peak sits at 0 dB | On / Off | Turn on when comparing files recorded at different levels. Leave off if your gain staging already makes sense. |
| FFT Size | Size of the analysis window — larger values give sharper frequency detail but blurrier timing | 512, 1024, 2048, 4096, 8192 | 4096 is a solid default. Drop to 1024 for percussive material where timing matters more than frequency precision. |
| Hop Size | How far the analysis window advances between snapshots — smaller hops mean more overlap and smoother results | 128 to half the FFT Size | The default (one quarter of FFT Size) works well for most material. Lower it for denser, higher-resolution analysis at the cost of processing time. |