Wikidata:Property proposal/audio transcription 2

audio transcription edit

Originally proposed at Wikidata:Property proposal/Commons

Done: audio transcription (P9533) (Talk and documentation)

Description	transcription of the word/text being spoken in this file
Data type	Monolingual text
Template parameter	"transcription" in commons:Template:Pronunciation file and commons:Template:Lingua Libre record
Domain	short spoken audio files, predominantly files from commons:Category:Pronunciation
Example 1	File:De-Katze.ogg, File:De-Katze2.ogg, File:LL-Q188 (deu)-Sebastian_Wallroth-Katze.wav → "Katze"@de
Example 2	File:De-at-Katze.ogg, File:LL-Q188 (deu)-Natschoba-die Katze.wav → "die Katze"@de
Example 3	File:Fr-chat.ogg → "un chat"@fr
Example 4	File:LL-Q150 (fra)-Aemines6-chat.wav, File:LL-Q150 (fra)-Benoit Rochon-chat.wav, File:LL-Q150 (fra)-DSwissK-chat.wav → "chat"@fr
See also	IPA transcription (P898), media legend (P2096), Timed Text

Motivation edit

(Someone else's proposal from a few years ago: Wikidata:Property proposal/audio transcription)

There are hundreds of thousands of pronunciation files on Commons. It can be tricky to find out whether there is a file for what you are looking for because there are various naming schemes for them and the filenames are also not always an exact match for the text being spoken.

Having the text as part of the structured data would allow us to use the data in queries, e.g.:

To find files for a specific word
To find words we have pronunciations of which don't exist as lexeme forms
To find inconsistencies such as a form representation not matching the text of the linked audio file

The data could also be used in other ways:

To generate a short description of the file, e.g. "Pronunciation of (text) in (language)"
To display the text being spoken when using the file, e.g. like on wikt:fr:chat#Prononciation.

This would not be the same as media legend (P2096). That property would be expected to contain more of a description, not just the words being spoken.

This is only intended for short words and phrases - there is a limit on the length of monolingual text statements anyway - and I suggest using Timed Text for transcripts of longer audio files.

- Nikki (talk) 21:14, 25 March 2021 (UTC)[reply]

Discussion edit

@NMaia, ديفيد عادل وهبة خليل 2, ArthurPSmith, Pigsonthewing: Pinging you since you contributed to the previous proposal. - Nikki (talk) 21:19, 25 March 2021 (UTC)[reply]
Support Don't ping me NMaia (talk) 22:05, 25 March 2021 (UTC)[reply]
Support --Lucas Werkmeister (talk) 18:23, 26 March 2021 (UTC)[reply]

@Nikki, Lucas Werkmeister: Done Now audio transcription (P9533). --Lymantria (talk) 17:06, 9 May 2021 (UTC)[reply]