VAEDecodeAudioTiled - ComfyUI Built-in Node Documentation

This node converts a compressed audio representation (latent samples) back into an audio waveform using a Variational Autoencoder (VAE). It processes the data in smaller, overlapping sections (tiles) to manage memory usage, making it suitable for handling longer audio sequences. The decoded audio is also normalized to keep its volume level consistent.

Inputs

Parameter	Description	Data Type	Required	Range
`samples`	The compressed latent representation of the audio to be decoded.	LATENT	Yes	N/A
`vae`	The Variational Autoencoder model used to perform the decoding.	VAE	Yes	N/A
`tile_size`	The size of each processing tile. The audio is decoded in sections of this length to conserve memory (default: 512).	INT	Yes	32 to 8192
`overlap`	The number of samples that adjacent tiles overlap. This helps to reduce artifacts at the boundaries between tiles (default: 64).	INT	Yes	0 to 1024

Outputs

Output Name	Description	Data Type
`output`	The decoded audio waveform, including its sample rate information.	AUDIO

The output sample rate is taken from the input samples when it contains one; otherwise, it is read from the VAE model (defaulting to 44100 Hz).

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 5ddedf218ba27ab9f463646c1e5288091172f2d7fae8f2980bb2b5e4d3dca89c

​Inputs

​Outputs

Inputs

Outputs