What You’ll Learn

  • How to measure audio statistics with the astats filter
  • How to interpret RMS, peak, DC offset, and dynamic range values
  • How to collect per-channel statistics
  • When to choose astats over volumedetect
  • How to build an audio quality check workflow from these measurements

Tested with: FFmpeg 6.1
Platform: Windows / macOS / Linux

What is the astats Filter?

astats is a filter that prints a wide range of statistics about an audio stream to the console. In a single pass it reports RMS, peak level, DC offset, dynamic range, and clipping indicators — enough for a full first-look diagnostic.

Basic Usage

ffmpeg -i input.mp4 -af "astats" -f null /dev/null

The report is written to stderr, just like FFmpeg’s regular log output.

Audio file statistics

ffmpeg -i input.mp3 -af "astats" -f null /dev/null

Reading the Output

A typical run produces output like this:

[Parsed_astats_0 @ ...] Channel: 1
[Parsed_astats_0 @ ...]   DC offset: 0.000002
[Parsed_astats_0 @ ...]   Min level: -0.999969
[Parsed_astats_0 @ ...]   Max level: 0.999939
[Parsed_astats_0 @ ...]   Peak level dB: -0.000263
[Parsed_astats_0 @ ...]   RMS level dB: -17.834118
[Parsed_astats_0 @ ...]   RMS peak dB: -12.456
[Parsed_astats_0 @ ...]   RMS trough dB: -40.123
[Parsed_astats_0 @ ...]   Crest factor: 32.789
[Parsed_astats_0 @ ...]   Flat factor: 0.000000
[Parsed_astats_0 @ ...]   Peak count: 0
[Parsed_astats_0 @ ...]   Noise floor dB: -90.234
[Parsed_astats_0 @ ...]   Dynamic range: 72.400
...
[Parsed_astats_0 @ ...] Overall
FieldMeaning
DC offsetDC offset (closer to 0 is better)
Peak level dBHighest instantaneous peak (dBFS)
RMS level dBOverall average RMS level (dBFS)
RMS peak dBMaximum RMS level
RMS trough dBMinimum RMS level
Crest factorCrest factor (peak-to-RMS ratio)
Flat factorShare of consecutive identical samples (useful for clipping detection)
Peak countNumber of clipped samples
Dynamic rangeDynamic range (dB)

Filtering Which Measurements Are Reported

Use measure_perchannel and measure_overall to pick which fields to print.

Show RMS and peak level only

ffmpeg -i input.mp4 -af "astats=measure_perchannel=RMS_level+Peak_level" -f null /dev/null

Overall statistics only (skip per-channel)

ffmpeg -i input.mp4 -af "astats=measure_perchannel=0" -f null /dev/null

Emitting Results as Metadata

With metadata=1, per-frame statistics are attached as frame metadata, which can then be printed with the ametadata filter:

ffmpeg -i input.mp4 -af "astats=metadata=1,ametadata=print:file=-" -f null /dev/null 2>&1 | head -50

Segment-by-Segment Statistics (reset Parameter)

Use the reset parameter to reset and emit statistics every N samples:

ffmpeg -i input.mp4 -af "astats=reset=44100" -f null /dev/null

At 44.1 kHz, 44,100 samples is roughly one second, so this prints a fresh snapshot every second. Useful when you want to see how loudness changes over time.

astats vs. volumedetect

FilterPrimary useOutput
astatsDetailed audio quality diagnosisRMS, peak, DC offset, dynamic range, and more
volumedetectQuick volume sanity checkmean_volume, max_volume, histogram

Use volumedetect when you just need a fast look at loudness, and astats when you need a thorough quality check:

ffmpeg -i input.mp4 -af "volumedetect" -f null /dev/null

Detecting Clipping

If Peak count is greater than 0 or Flat factor is high, the audio is likely clipping:

ffmpeg -i input.mp4 -af "astats=measure_perchannel=Peak_count+Flat_factor" -f null /dev/null

When clipping is detected, lower the level with the volume filter or fix the gain at the source.

Checking and Fixing DC Offset

A large DC offset can be removed with the highpass filter:

ffmpeg -i input.mp4 -af "highpass=f=10" output.mp4

Confirm the DC offset with astats before applying the fix.

Audio Quality Check Workflow

  1. Run astats to inspect RMS, peak, DC offset, and clipping
  2. If clipping is present, reduce the level with the volume filter
  3. If DC offset is large, remove it with the highpass filter
  4. If RMS is uneven across the material, normalize with the loudnorm filter

Frequently Asked Questions

What is the difference between astats and ebur128?

astats reports peak/RMS/DC/crest factor frame-by-frame, while ebur128 measures broadcast-standard loudness (LUFS). Use astats for waveform diagnostics and ebur128 for delivery loudness targets.

How do I get only the peak level?

Filter the metadata output: ffprobe -f lavfi -i "amovie=in.wav,astats=metadata=1:reset=0" -show_entries frame_tags=lavfi.astats.Overall.Peak_level -of csv=p=0.

Can astats detect silence?

Indirectly — when RMS drops below your threshold (e.g. -60 dB) the segment is silent. For a true silence detector use the silencedetect filter instead.

Why does astats report different numbers per channel?

It analyses each channel independently. A stereo file with content only on the left channel will show -inf dB peak on the right. Add metadata=mode=single to get an aggregated value.

Can I run astats on a video file?

Yes — FFmpeg auto-extracts the audio stream. For a multi-track video, select the stream explicitly with -map 0:a:0.

Tested with ffmpeg 6.1 / Ubuntu 24.04 (GitHub Actions runner)
Primary source: ffmpeg.org/ffmpeg-filters.html#astats