Skip to main content
Audio Analyze

Audio Analyze · screenshot coming soon

​
What this node does

Audio Analyze runs an FFmpeg measurement pass over a track and returns a JSON report as text. It does not change the audio. Depending on mode it uses ebur128 (broadcast loudness), volumedetect (peak/mean dB), astats (detailed statistics), or silencedetect (silence segments). In loudness mode it also compares the result against common platform delivery targets. It takes a COMFYTV_AUDIO snapshot (or a COMFYTV_VIDEO, whose audio track is used) and outputs COMFYTV_TEXT. It has a ▶ Run. A source is required. To feed native ComfyUI AUDIO in, insert a Bridge — see bridges.md.

​
When to use it

  • Check a mix’s integrated loudness (LUFS) and true peak before delivery to YouTube, Spotify, etc.
  • Find the peak and mean volume to plan gain/normalization.
  • Locate silent gaps in a recording (for trimming or splitting).

​
Parameters

​
mode

What to measure. Options:
  • loudness (default) — EBU R128 integrated LUFS, LRA, thresholds, and true peak; adds a platforms compliance breakdown. Requires a clip of at least ~400 ms.
  • volume — mean and max volume in dB, plus sample count.
  • stats — overall statistics: DC offset, peak/RMS levels, flat factor, peak count.
  • silence — a list of silence segments (start/end/duration) plus a count.

​
silence_noise_db

Silence detection threshold in dB, -100 to 0 (default -60). Anything quieter than this counts as silence. Used only in silence mode.

​
silence_duration

Minimum length (seconds) a quiet stretch must last to count as silence, 0.01–60 (default 2.0). Used only in silence mode.

​
Outputs

​
Tips

  • loudness needs a clip long enough (≈400 ms minimum) or ebur128 produces no summary and the node errors.
  • The platforms field in loudness mode tells you whether the track meets typical streaming targets at a glance.
  • This node is read-only — pair it with Audio Loudness or Audio Dynamics to actually correct the levels it reports.
  • Audio Loudness — normalize to a target LUFS after measuring here.
  • Audio Visualize — see the waveform or spectrum instead of numbers.
  • Audio Dynamics — compress/limit to control peaks this node reveals.