Multiple entries: Add "Kana reading" field for Japanese names

[szc126]

Most atwiki.jp and dic.nicovideo.jp entries have the pronunciation of the entry title written in hiragana. It's also sometimes given for usernames, UTAU names, etc.

For Japanese-language users of VocaDB, it can make searching easier (instead of having to input the precise kanji/katakana/hiragana used in a name: 唄 or 歌 or うた or ウタ, or inputting romaji), would be good for phonetic sorting, and would be a good addition to entries as well.

It is roughly related to the Romaji field, both being readings, but the Romaji field has potential searching barriers such as restoration of loanwords and word spacing, as opposed to a hiragana string.

[ycanardeau]

Could this be done as part of #10 with the ja-Hira locale? AniDB and MusicBrainz will be useful as a reference.

[szc126]

I like that way of thinking of it.

[blueset]

Here is a dump of automated ja-Hira name notations based on http://vocaloid.eu/vocadb/dump.zip in 2021 Sep 04. Only the titles able to be confirmed automatically are included (with some manual additions).

Hope this helps
vocadb_safe.csv

[ycanardeau]

@blueset Thanks for your help! Note that the dump file isn't always up-to-date and it's manually done. Please let me know if you need the latest dump. By the way, I think it would be nice if you could share your automated script somewhere for future usage.

[blueset]

Thanks for the suggestion. I’ve uploaded the source to https://github.com/blueset/vocaloid-yomigana, along with some data from Vocaloid Wiki (Fandom/Wikia) and 初音ミク wiki (atwiki). Hope this helps.

[szc126]

A related musing (alternative script forms): Do we "need" to restore lyrics that were only provided in romanized form? https://vocadb.net/S/289656 (romanized Russian, Japanese; Vocaloid Lyrics Wiki has produced a "restored" version); https://vocadb.net/S/68831 (romanized Japanese)? Perhaps we need lyrics tags like "ja-Latn + Original [from the producer]" instead of treating these as inferior ("Romanized" lyrics instead of "Original" lyrics). One would expect "Romanized" lyrics to be secondary, but in such cases the "Original" lyrics are secondary.

Also (related even more tenuously: script shenanigans) https://vocadb.net/S/594303 (provided in 2 Hokkien Romanizations and 1 Hokkien "Hangulization": hbl-Latn-TL, hbl-Latn-POJ, hbl-Hang? and the editing UI only allows one "Romanization")

[mn7216]

This would be an additional field in addition to the Non-English, Romanized, and English fields, correct?