Convert VHS Audio to WAV bytes
Bridge VideoHelperSuite audio into this pack
- vhs_audio
- wav_bytes
This is a bridge node, and it exists purely so VideoHelperSuite's audio and this pack's transcription nodes can talk to each other. If you're already using VHS to load a video and it hands you audio in VHS's own VHS_AUDIO format, you can't wire that straight into Whisper Transcribe - it wants this pack's own WAV_BYTES type instead. Drop this node in between and the conversion just happens. No settings, one input, one output.
Why you'd reach for it over Load Audio from Video
If your workflow already routes through VHS for the video side - frame extraction, video info, whatever else you're doing with the clip - you get its audio as a side output for free. There's no reason to load the same video twice through two different loaders. This converter lets you reuse the audio VHS already pulled out, instead of pointing this pack's own Load Audio from Video at the same file a second time.
How it works
No processing knobs - it re-wraps the audio bytes from VHS's container into this pack's own container, so the type checker on the graph will actually let you connect the wire. That's the entire job.
The input and output that matter
vhs_audio(input,VHS_AUDIO) - from a VideoHelperSuite loader that exposes audio, like Load Video FFmpeg.wav_bytes(output,WAV_BYTES) - wire it into Whisper Transcribe, Whisper Transcribe Batch, or Audio Sink, exactly like any otherwav_bytessource in this pack.
Installing it
You need both packs for this node to be useful: VideoHelperSuite for the VHS_AUDIO producer, and ComfyUI-TranscriptionTools for this converter and everything downstream of it. Comfy.icu already ships VideoHelperSuite in its base image; TranscriptionTools isn't part of the shared image, so install it via ComfyUI Manager (search ComfyUI-TranscriptionTools) or by hand:
cd ComfyUI/custom_nodes
git clone https://github.com/royceschultz/ComfyUI-TranscriptionTools
then restart. If you're self-hosting from scratch and don't already have VHS, that's a second, separate install (Kosinkadink/ComfyUI-VideoHelperSuite) - this pack's own readme credits the VHS team's examples as instrumental to its development, which tells you how closely the two are meant to interoperate.
Common issues & troubleshooting
vhs_audio has nothing to connect to. That means either VideoHelperSuite isn't installed, or the upstream node you're using doesn't happen to expose a VHS_AUDIO output - not every VHS node emits audio, only the loaders that explicitly read or produce it.
Confusing this with VHS's own internal "legacy VHS_AUDIO to AUDIO" converter. That one lives inside VideoHelperSuite itself and converts VHS's old internal audio format into ComfyUI's native AUDIO type - a completely different output from this node's WAV_BYTES. If what you actually want is ComfyUI's native audio type rather than something to feed this pack's Whisper nodes, you want that node, not this one.
Nothing to configure, so a failure here is almost always upstream. If a run breaks at this node, check that the vhs_audio you're feeding it actually contains something - some VHS loaders return empty audio if the source video's audio track failed to decode in the first place.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| vhs_audio | VHS_AUDIO | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| wav_bytes | WAV_BYTES | — |