Load Audio Pixaroma
Pick the exact piece of a track by looking at it, not guessing
- audio
Load Audio Pixaroma is the audio counterpart to Load Image: pick a sound file, choose the exact piece of it you want, and hand that piece to the rest of the workflow - without cutting the file up in another program first or counting seconds in your head.
The node draws the whole file as a waveform on its face, so you can see where the loud parts are. You drag a window across it to choose your moment: drag either orange edge to trim, drag the middle to slide, or click anywhere to place the play cursor and press play to hear it. It's the rare loader that genuinely shows you what you're about to get before you run.
The one input that matters
seconds(FLOAT, optional) - how many seconds to take. Leave it unconnected to use the whole file or whatever length is set in the node's settings.
The connection people actually want: wire the seconds output of Duration Pixaroma into this input, and the window becomes exactly as long as the video you're about to make, resizing the moment you connect it. Picture and sound literally cannot drift apart, because both derive from the same frame count.
If your window runs past the end of the file, the node fills the rest with real silence or loops back, whichever you chose - and says so on the node, so there's no mystery about the tail.
Output
A single audio output, ready to wire into a save node (Save Video Pixaroma mixes it into an MP4 as AAC), a sync node (H3 Audio Sync Pixaroma), or anything else that takes audio. Upload straight from your computer; the file list refreshes every time you open the node.
Install
It's part of the Pixaroma pack, so the drill is the same as the rest of them: ComfyUI Manager → search Pixaroma → install → restart, or:
cd ComfyUI/custom_nodes
git clone https://gitlab.com/pixaroma/comfyui-pixaroma.git
No models, no extra Python dependencies. Works with any model, since audio is just data flowing through the graph. The usual pack gotcha: after an update, hard-refresh the browser tab (Ctrl+Shift+R / Cmd+Shift+R) or you'll keep looking at the old node visuals and think something broke.
Where people get burned with audio-in-video generally, it's almost always a length mismatch - the clip is 5.17 seconds and the track is 5.0, and the drift compounds. Wiring Duration's seconds straight in is the fix, and it's the reason this node pairs so well with that one.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| secondsopt | FLOAT | How many seconds to take, normally wired from the seconds output of Duration Pixaroma. Leave it unconnected to use the whole file or the length set in this node's settings. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | The piece of the file you selected, ready to wire into a save node, a sync node, or anything else that takes audio. |