Skip to content
AudioLab .tools
Sign in

/docs · VoiceLab · Intro

Why filler density matters more than filler count

Counting "ums" is the wrong measurement. Density tells you whether they are a problem.

Every podcast QA tool wants to count your fillers. Most of them stop there. Counting fillers as a raw number is mostly noise: a 5-minute interview and a 90-minute discussion produce very different totals, but the experience of listening to them depends on something else entirely.

That something is density: fillers per minute of actual speech.

The measurement

filler_density = filler_count / speech_seconds * 60

Note the divisor: not total file duration, but speech seconds. A podcast with five-second filler bursts followed by silence is different from one where every second is dense with hesitation.

Rules of thumb from the literature and from listening — VoiceLab does not measure filler density, see below, so these are for your own judgement rather than a number the product returns:

DensityListener experience
< 3 /minSmooth, professional
3–6 /minConversational, fine for talk-shows
6–10 /minNoticeable, may need editing
> 10 /minDistracting: heavy edit pass needed

These are guidelines, not laws. A high-energy improv podcast at 12 fillers/min can be brilliant. A scripted explainer at 5 fillers/min is broken.

Why count alone misleads

Take two podcasts:

  • Podcast A: 60 minutes, 40 fillers. Total: 40 fillers.
  • Podcast B: 10 minutes, 30 fillers. Total: 30 fillers.

By count, Podcast A is “worse.” By density, Podcast A is 0.67/min and Podcast B is 3/min. The latter is a more honest summary of what a listener will feel.

Pause length matters too

A 200ms pause is usually invisible. A 600ms pause is dramatic. A 1500ms pause is a hesitation, often replacing a filler. VoiceLab tracks pause distribution — that part it does measure — because replacing a filler with a long hesitation isn’t actually an improvement. It’s the same problem in a different shape.

How VoiceLab detects fillers without ASR

It doesn’t, and we measured why.

The proxy this article used to describe (short energy bursts standing alone between pauses) is built and shipping. But run it over twelve read-aloud recordings with no hesitations in them and it still reports 0 to 36 bursts per speech minute, averaging 10. This article recommends staying under 6. The false positives alone are above the target.

So we looked at what those bursts actually were, hoping to filter them out. Eighteen of twenty were voiced, at an ordinary speaking pitch of 132–286 Hz, with a flat envelope. Not plosive artefacts: short standalone words. And an “uh” looks acoustically identical to a “yes”.

That is not a tuning problem. There is no threshold that separates “uh” from “yes” without knowing which word it is, and knowing that is speech recognition.

What VoiceLab does ship, and what it is honestly good for:

  • Pause rhythm: count, median, longest, and the split into short (80–400 ms), medium and long. Gaps under 80 ms are excluded because they are consonant closures, not pauses; that alone changed the count on our reference recordings from 18 to 10.
  • Pause clustering: ten-second windows carrying more than four short pauses. This is the second proxy this article always described, and it holds up.
  • Timestamps to listen to: every short sound standing alone between pauses, as a clickable moment. A half-hour recording becomes six places to check. Your ears decide in two seconds what no threshold can.

For an actual filler count you need speech-to-text (Whisper, Deepgram, AWS Transcribe) over the audio, and then the density formula above applies to a real count instead of a proxy.

What to do with the number

If your density is above 6/min and you’re editing the show, three tactics from working editors:

  1. Cut, don’t replace. Empty space between sentences sounds better than a filler in 80% of cases.
  2. Tighten before you cut fillers. Most “filler” perception is actually pace. Tightening overall delivery makes fillers fade into the background.
  3. Leave intentional ones. A “you know what I mean?” with character beats a sterile script. Don’t edit the personality out.