> ## Documentation Index
> Fetch the complete documentation index at: https://comfytv.org/llms.txt
> Use this file to discover all available pages before exploring further.
# Score Synth
> Render a performance to real audio with a SoundFont — ComfyTV's built-in SF2/SF3 synthesizer.
## What this node does
**Score Synth** is the audio end of the symbolic-music lane. It takes a
`performance` (from **Score Performer** or **Click Track**) and synthesizes it
into audio using ComfyTV's own SoundFont synthesizer — no external player
required. Point it at a `.sf2`/`.sf3` SoundFont from your resource library and
it plays the notes with those instrument samples.
Press **▶ Run** to render on the backend; progress is reported as it works. The
output is a `COMFYTV_AUDIO` clip you can preview, pass to audio stages, or
Bridge into native ComfyUI.
You choose a default General MIDI instrument (`program`), and optionally
override the instrument per MIDI channel with a small JSON map — handy for
multi-part performances where each channel is a different instrument.
## When to use it
* You have a performance (from Score Performer or a click track) and want to
hear it as audio, not just MIDI.
* You want a specific instrument sound from a SoundFont you've added to the
resource library.
* You need per-channel instruments in one render (e.g. piano + bass + drums).
## Parameters
### soundfont
The SoundFont file name from your resource library (a `.sf2`/`.sf3`). If left
blank, the synth uses its built-in fallback voices. If you name a file that
isn't in the library, the run fails with a "soundfont … not found" message.
### program
The default General MIDI instrument, default **piano**. The options are the GM
program names; this instrument plays any channel you don't override below.
### channel\_programs
An optional JSON map of MIDI channel → GM program, e.g.
`{"0": "piano", "1": "acoustic_bass"}`. Use it to assign different instruments
to different channels of a multi-part performance. Values can be GM program
names or raw program numbers. Invalid JSON is ignored.
### gain
Output level multiplier, **0.1–2.0** (default **1.0**). Below 1 attenuates,
above 1 boosts; back off if the render clips.
### performance (optional input)
The performance to render, type `COMFYTV_TEXT` — wire this from **Score
Performer** or **Click Track**. Running with nothing wired fails with "Score
Synth needs a performance".
## Outputs
| Output | Type | Meaning |
| --------- | --------------- | ----------------------- |
| **audio** | `COMFYTV_AUDIO` | The rendered audio clip |
## Tips
* SoundFonts live in the resource library — add your `.sf2`/`.sf3` there first,
then type its file name into `soundfont`.
* For a multi-instrument mix, set `program` to the most common voice and use
`channel_programs` only for the channels that differ.
* If the result sounds hot or distorted, lower `gain` a little before
re-rendering.
## Related nodes
* **Score Performer** — produces the `performance` this node renders.
* **Click Track** — a metronome whose click also feeds this synth.
* **Chord Accompaniment** — generates a performance from chord symbols.
* **Bridge** — convert `COMFYTV_AUDIO` to/from native ComfyUI audio. See
[https://github.com/jtydhr88/ComfyTV/blob/main/docs/bridges.md](https://github.com/jtydhr88/ComfyTV/blob/main/docs/bridges.md)