Text Alignment

The Text Alignment job aligns a provided transcript of neumed text or lyrics on a page to the text that appears on the page itself by using optical character recognition (OCR) and the Needleman-Wunsch algorithm.

Input Ports

• Text Layer - The layer of text in the source image.
• Transcript - A transcript of the text on the page.
• OCR Model - An OCRopus OCR model.

Output Ports

• Text Alignment JSON - A JSON file containing text alignment information.

Miyao Staff Finding

This job performs the Miyao staff finding algorithm. It finds the staves in an image of staves. Note that this is not the job called “Miyao Staff Finder” as that job has different settings and produces a different output.

Settings

• Number of lines - The number of lines in a staff. The default is 4. Zero automatically detects the number of lines.
• Interpolate - Interpolate found line points so all lines have the same number of points. This must be set to true.

Input Ports

• Image containing staves - The staff line layer.

Output Ports

• JSOMR - A JSON file containing staff and staff line information.

Heuristic Pitch Finding

This job finds the pitch of musical element connected components. This requires the output from the Miyao Staff Finding job to work properly.

Settings

• Discard Size - The size of elements to discard. Default is 12.

Input Ports

• JSOMR of staves and page properties - The output of the Miyao Staff Finding Job.
• Classified Connected Components - The classified connected components from a classifier job.

Output Ports

• JSOMR of glyphs, staves, and page properties - A combined JSON file with data on musical elements with pitches, staves, and the page as a whole.