Skip to content

Transcribe post-2020 Cantopop into the corpus #1

Description

@candpixie

The corpus behind every number in the README stops at 2020. Nothing about how Cantonese tone and melody interact in the last five years is measured.

This is the single highest-value contribution and needs no coding, only a good ear and patience.

What to do

  1. Pick a Cantopop song released 2021 or later.
  2. Comment here with the title so nobody duplicates it.
  3. cp corpus/TEMPLATE.krn corpus/X0001.krn and transcribe it. Format is in corpus/README.md.
  4. python scripts/check_krn.py corpus/ until it's clean.
  5. Open a PR with the single .krn file.

What matters most

Independent and non-radio artists are the biggest blind spot. The corpus is
dominated by mainstream radio singles where 林夕 and 黃偉文 wrote 60% of the
lyrics, so Gareth.T, Kiri T, serrini, per se, Yellow and similar are worth more
per song than another mainstream ballad.

One song is a real contribution. There is no minimum.

Transcriptions are CC BY 4.0 so they pool with the upstream corpus. Transcribe
the melody yourself, do not paste in commercial lyric sheets or sheet music
scans.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    corpusTranscription work, the highest-value contributiongood first issueGood for newcomershelp wantedExtra attention is needed

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions