Audio Join 2 Channels

Stereo is two channels, and sometimes you have the two halves separately. Maybe you generated a left-panned voice and a right-panned music bed and want them on one stereo wire, or you're reconstructing a stereo file from two mono stems. Audio Join 2 Channels (SET_AudioJoin2Channels) takes two audio inputs and puts the first on the left, the second on the right, and hands you a stereo signal. It's the pack's "build stereo" node, and it pairs with the "break stereo apart" node, Split 2 Channels.

Inputs

• audio_left - becomes the left channel.
• audio_right - becomes the right channel.

Both inputs are forced to mono first, and the README is upfront about the detail: it uses the plain average downmix method to do it. So if you feed it a stereo file as audio_left, that whole stereo file gets averaged down to mono before it becomes your left channel. If you want a different downmix strategy for that step, the README's advice is to do the conversion yourself with the Channel Converter first. Fair enough.

The alignment magic

Here's where it earns its keep. Two audio inputs almost never match. The node aligns them with the same logic as the pack's Batch Audios: everything is resampled to audio_left's sample rate, shorter clips are padded with silence to match the longer one, and mismatched batch sizes repeat the last item of the shorter batch. You can throw a 44.1 kHz 5-second clip and a 48 kHz 3-second clip at it and get a coherent stereo pair out. That's the difference between this and doing it by hand.

The output, audio_out, is a stereo signal: left channel from audio_left, right from audio_right, both at audio_left's sample rate, both as long as the longer input.

Why it matters

Aside from the obvious stem-mixing use, this is the node that turns "two mono sounds" into "one stereo sound" for any consumer that demands stereo - video pipelines, stereo-only outputs, or just keeping your batch uniform. It's also the natural companion to Split 2 Channels: split a stereo file apart, process the halves independently, join them back. That loop - split, treat, rejoin - is a real workflow pattern and it's why the pack ships both nodes.

Install & gotchas

From ComfyUI Manager, search "Audio Batch", or:

cd ComfyUI/custom_nodes/
git clone https://github.com/set-soft/ComfyUI-AudioBatch
pip install seconohe

Restart ComfyUI. No model downloads.

Two things to remember. First, the silent padding: if your two inputs differ in length, the shorter side gets silence, not stretching - join a 30-second music bed with a 3-second voice and you get a 30-second stereo file where the right channel goes quiet after 3 seconds. Second, the fixed average downmix on multi-channel inputs means this node is a mono-centric tool; feeding it surround sound is asking for a plain average collapse. For anything more careful, downmix explicitly first.