Audio Latent Info ⚡🅡🅞🅣🅘

Another debug/inspector node for ACE-Step latents, and the detail that tells you which generation it targets is the default frame_rate: 10.77, which doesn't match the 25 fps used by this pack's ACE-Step 1.5 nodes. That points to this one being built for the earlier ACE-Step generation, where ACEStepAnalyzeLatent looks like the more direct 1.5-era equivalent.

Part of ryanontheinside's ACE-Step utility set.

How it works

Feed it a latent and a frame rate, and it does the arithmetic to tell you how long that latent actually represents in real time, plus its batch size and frame count - all without needing to decode anything to audio first.

The inputs and outputs that matter

• audio_latents - the LATENT you want to inspect.
• frame_rate - default 10.77. This isn't cosmetic: it's the actual conversion factor between latent frames and seconds, so it needs to match whatever rate your ACE-Step pipeline is actually using or every duration this node reports will be wrong.

Outputs: batch_size, length_frames, duration_seconds, and info (a readable summary string). duration_seconds is the one most people actually look at.

How to install it

Via ComfyUI Manager, search RyanOnTheInside. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/ryanontheinside/ComfyUI_RyanOnTheInside

then restart. No model downloads - it reads whatever latent your pipeline already produced.

Common issues & troubleshooting

duration_seconds looks wrong even though the latent seems fine. This is almost always a frame_rate mismatch, not a broken latent - if you're working with a different ACE-Step generation than this node's default targets, the reported duration will be off by whatever factor separates the two rates. Compare against ACEStepAnalyzeLatent if you're unsure which one actually matches your pipeline.

Not sure which inspector node to use. If your latent came from the 1.5-era Lego/hybrid guiders in this pack, try ACEStepAnalyzeLatent first since it doesn't require you to supply a frame rate manually. Reach for this node when you know you're working with the earlier ACE-Step format, or when you specifically need batch_size.

Batch size larger than expected. If you generated multiple variations in one pass, batch_size reflects that - not a bug, just confirmation of how many clips are actually stacked in the latent you're inspecting.