Using AnthemScore to Transcribe Music

Installation

For installation instructions and help, click here.

Transcribing Files

Open a song by clicking File > Open and selecting an audio file and save location. Most common audio formats are supported, though the exact list varies with operating system. In the open file dialog you have the option to process only part of a song or skip finding notes and just generate a spectrogram display. You can also increase the frequency resolution and/or decrease the time step if you want a higher resolution image. A higher resolution image will take longer to generate, but doesn't affect the automatic note detection.

You can add additional songs to the transcription queue while the song is being processed by repeating the above steps. If you add songs one at a time, the open file dialog will pop up and you can adjust the settings. If you add multiple files at once, they will be transcribed using the default settings and saved to the "Default Save Folder" listed above the file queue in the settings panel. To cancel a song, click the red X next to it.

Be patient!
Be patient!

The progress bar at the bottom of the screen shows you the job status. It can take several minutes to 1 minute of music. To reduce the wait time, you can process the song in parts (such as 0 to 30 seconds, then 30 to 60 seconds, etc.) or use a computer with more processor cores.

Editing Notes

Notes (black) and potential notes (white). Right clicking will add or remove a note.
Notes (black) and potential notes (white). Right clicking will add or remove a note.

You can edit notes in AnthemScore by right clicking in the spectrogram display. Confirmed notes are displayed as a black with a glowing border. Potential/suggested notes are shown in white. You can see more or fewer potential notes by adjusting the threshold slider in the side panel. Notes you add will be snapped to the nearest automatically detected start time. Currently, you can only add or remove notes. To edit the note type or duration you must save a sheet music file and edit in a separate music notation program.

Viewing Sheet Music

Once processing is complete, the sheet music will be automatically opened with MuseScore, if installed. In MuseScore, you can edit the file and save it as a PDF, MIDI file, or in another format. If you installed MuseScore and the sheet music does not open automatically, use Tools > Set path to MuseScore and select the MuseScore application (.exe on Windows). Then, click the green sheet music button . This button will save any edits you have made to the notes and open the sheet music in MuseScore. The musicXML file is automatically saved in the default save folder listed in the settings panel, unless you specify a different path when opening the audio. You can also save the sheet music to a new location by clicking File > Save As. MusicXML is a common sheet music format supported by Finale and many other music notation programs.

Save Options

To see the options for saving files from AnthemScore, click File > Save as.

Format

Staffs

The staffs options allow you to move all notes to the treble or bass clef without changing their pitch.

Pitch Adjust

The pitch adjust options will shift the pitch of all notes up or down by a fixed number of semitones and/or octaves.

Custom Note Range

The custom note range lets you set the lowest and highest notes you want to see in your sheet music. When there is a note outside of this range, it will either be moved within the range by changing the octave or removed from the sheet music, depending on the option you select. If you play an instrument other than the piano, you may want to set this to the lowest and highest notes you can play on your instrument.

Save as Default

By checking save as default, you can apply these settings every time sheet music is generated.

CPU Load and Speed

By default, AnthemScore will use all available processor cores to speed up processing. You can reduce the number of worker threads spawned in File > Preferences. It's best to only have one instance of the program running at a time. Multiple instances of AnthemScore will slow each other down and may overwrite shared files.

Command Line Interface

You can call the program from the command line to process songs in the background (no GUI), and automatically save the musicXML file or spectrogram data. Use the -h or --help flags to see a list of options. Example basic use:

AnthemScore audio.mp3 -a -x output.xml

Understanding Music

The Spectrogram

A spectrogram of the song will be displayed in the main window. A spectrogram is a color plot of the energy at different frequencies over time. By default, the horizontal axis is time and the vertical axis is frequency (log scale), but this can be changed in the settings panel on the left. The color shows the amplitude. Dark blue/black indicates low amplitude and red indicates high amplitude. If the "note lines" box is checked, horizontal lines are drawn at the frequencies of the 88 piano keys. The bold white lines are the lines of the treble and bass staffs (G, B, D, F, A) and (E, G, B, D, F).

You can move the mouse over a line to see the name of the note and left click to listen to it. Select View > Mouse harmonic lines to see the the harmonic frequencies. Those harmonic lines are the locations that you would expect to see a high amplitude if there was a note at the mouse cursor. You can view individual audio channels by selecting the channel from the drop down menu on the settings panel.

What are Harmonics?

Harmonics are equally spaced. They appear to grow closer together here because frequency is on a logarithmic scale.
Harmonics are equally spaced. They appear to grow closer together here because frequency is on a logarithmic scale.

When a note is played on an instrument, the air vibrates at multiple frequencies, called the harmonics of the note. The pitch of the note is the frequency of its first harmonic. For example, when you play C4, or middle C, on the piano, the piano string vibrates at the frequency of C4 (261.63 Hz). But it also vibrates at multiples of that frequency: 523.26 Hz (C5), 784.89 Hz (G5), 1046.52 Hz (C6), etc. You only hear a single pitch because your brain recognizes that the frequencies are multiples of 261.63 Hz and groups them together. Usually the 1st harmonic is the strongest and each successive harmonic is weaker, but it can vary. For example, the clarinet has strong odd harmonics (1, 3, 5, 7, ...) and weak even harmonics (2, 4, 6, 8, ...).

The relative amplitude of the different harmonics is partly what gives different instruments their characteristic sound, or timbre. Sometimes the 1st harmonic may be missing entirely and the only way to identify the note is to look at the spacing between harmonics. It's also possible, and quite common, for two notes to played at once where one note falls on a harmonic of another note (for example C4 and C5, which are an octave apart). When this happens, the higher note's harmonics will be hidden by the lower note, unless they stand out from having stronger amplitudes.

Spotting Drums

Drums usually occur at lower frequencies and can be identified as features that are narrow in time and very wide in frequency. Percussion instruments often produce frequencies that are are not harmonic (not multiples of a fundamental frequency).

Repeating bass drum at low frequencies and drumstick clicks at high frequencies. Both are narrow in time and wide in frequency.
Repeating bass drum at low frequencies and drumstick clicks at high frequencies. Both are narrow in time and wide in frequency.

See a few examples of what music looks like.