Playlistation LogoPlaylistation
Login

Filters

Speechiness

Needs the Plus or Pro plan. The Speechiness filter keeps the tracks whose speechiness score sits inside a range you set.

What it matches

Speechiness is a score from 0% to 100% that detects spoken words in a track. The closer to 100%, the more of the recording is speech. Spotify's guidance, which the app repeats in the dialog, gives three bands. Above 66%, the track is probably all spoken words: a talk show, an audiobook, poetry. Between 33% and 66%, it mixes music and speech, in sections or layered, which is where most rap sits. Below 33%, it is music with little or no speech.

The filter compares each track's score with your range and keeps the track when the score is inside it, both ends included.

It separates rapped from sung vocals, which genre labels cannot do: an artist tagged "hip hop" also sings.

Where the data comes from

The score is part of Spotify's audio analysis of the recording. Playlistation reads it from Spotify the first time a track is checked and stores it for later updates.

Spotify has no analysis for every track. When it has none, the track has no speechiness score and the filter treats it as not matching: the track stays out of the playlist whatever range you set.

Operators and values

The filter has one operator, Between, so the dialog shows no operator card. You set a minimum and a maximum on a slider with two handles, from 0% to 100%, in steps of 5, so the 33% and 66% marks from Spotify's guidance become 35% and 65% in practice. A new Speechiness filter starts at 25% to 75%.

The Speechiness filter dialog with its two-handle slider
Left is sung, right is spoken.

The sentence collapses at the ends of the scale:

  • Minimum at 0: "The track speechiness is at most 15%".
  • Maximum at 100: "The track speechiness is at least 65%".
  • Anything else: "The track speechiness is between 25% and 75%".

A slider left at 0% to 100% shows the warning "With the current settings, all of the tracks will match this filter".

Examples

  • The track speechiness is at least 35% and The track artist is tagged with one of hip hop, rap. Rap that is actually rapped.
  • The track speechiness is at most 15%. Sung or instrumental music only, nothing that sounds like someone talking.
  • The track speechiness is at least 65%. Spoken-word and poetry tracks in your library, if you have any.
  • The track speechiness is at most 10% and The track instrumentalness is at least 50%. A study playlist with almost nothing that has a voice in it.

Things to know

  • Melodic rap and sung hooks pull a rap track below 33%. If a rap playlist looks thin, lower the minimum to 20% and check the borderline tracks in the track list at the bottom of the page: expanding a track shows "The Speechiness score is 28%".
  • Podcasts and audiobooks would score above 66%, but they are not in Liked Songs, so the filter only ever sees music tracks.
  • On the Free plan the filter is locked in the filter picker. See Plans and billing.
Previous
Liveness
Next
Library stats