Wikidata:Property proposal/Music mood

music mood

edit

Originally proposed at Wikidata:Property proposal/Creative work

   Done: music mood (P12830) (Talk and documentation)
Descriptionqualifier carrying an emotion (mood) relevant to an audio recording (musical or not)
Representsmusic mood (Q126179530)
Data typeItem
Domainmusic (Q638)
Allowed valuesinstances of audio recording (Q3302947), music track with vocals (Q55850593), music track without lyrics (Q55850643)
Example 1My Generation (Q8353114)energetic (Q126271501)
Example 2Happy Together (Q2733482)happy (Q126207699)
Example 3Love Me Tender (Q1426427)romantic (Q126272565)
Example 4People Have the Power (Q7165881)passionate (Q126596716)
Example 5White Riot (Q967376)angry (Q126274660)
Example 6Love Will Tear Us Apart (Q1426177)gloomy (Q126275248)
Example 7Strange Fruit (Q109664864)sad (Q126275836)
Example 8Suzanne (Q1796940)calm (Q126276136)
Expected completenessalways incomplete (Q21873886)
See alsoemotion (Q9415)
Distinct-values constraintno
Wikidata projectWikiProject Music (Q5830855)

Motivation

edit

The music industry, notably digital listening platforms and audiovisual music supervisors, have long used the notion of Mood to add enriched description to an audio recording (https://www.wikidata.org/wiki/Q3302947), music track with vocals (https://www.wikidata.org/wiki/Q55850593) or music track without lyrics (https://www.wikidata.org/wiki/Q55850643).

It is now possible for music fans to create Mood centric playlists (https://www.moodplayl.ist/). The international DDEX-MEAD standard is involved in trying to better standardize this concept (https://kb.ddex.net/implementing-each-standard/media-enrichment-and-description-(mead)/). The DDEX industry standard body has not yet reached an agreement on whether or not mood allowed value set can be determined by industry stakeholders. I sustain the idea, it would be better served by crowdsourced dynamics, such as the one offered by Wikidata.

While it is possible to add a Main subject (https://www.wikidata.org/wiki/Property:P921) as a predicate-property to an audiovisual object-item in Wikidata, it seems that being able to add a Mood to an audio recording would help the digital music ecosystem to better describe such entity. The one argument against this would be the very subjective nature of mood semantics. Therefore, we have done a literature search in order to identify databases and current usages of the mood concept by some industry stakeholders.

My research has identified several Mood systems initiatives : the Greek physician-philosopher Hippocrates humors, University of California at Berkeley Proceedings of the National Academy of Sciences (Cowen et al., 2020), psychologist Robert Plutchik’s emotion theory, DDEX-MEAD (Media Enrichment and Description) industry consortium standard, simbals mood and audio descriptors, Deezer Flow dynamic mood playlists, Cyanite.ai multi-label score based classifier, SoundCloud’s Musiio tag system and the Musicovery research initiative. Other initiatives include crowdsourced user-generated mood lists or extensive mood lists, like the AllMusic mood exploration tool (https://www.allmusic.com/moods). Our literature search is public and available on Zotero : https://www.zotero.org/groups/2165362/laticce/collections/AIH4HA3A

The AllMusic mood exploration tool (https://www.allmusic.com/moods) allows to retrieve the moods of tracks and their relative weight based on votes by the community. This could permit to source the moods added to Wikidata. See this example for the track 'My Generation" : https://www.allmusic.com/song/my-generation-mt0005694826#moodsThemes

For now, I’m proposing to be synthetic and have narrowed down a potential mood list to eight primary moods shown in the above examples. The community could add more music mood items over time. Youyouca (talk) 09:22, 3 June 2024 (UTC)[reply]

Discussion

edit