Tony: a tool for melody transcription

{{toc}}

h1. Tony Introduction

Tony is a program for computer-aided melody annotation. It has a graphical interface based on the  "SV libraries":/projects/sv and uses the "pYIN":/projects/pyin Vamp plugin to extract pitch track and notes from monophonic audio.

h2. Getting Started

h3. Installation

We provide ready-to-use software builds for Windows and Mac OSX on the "Downloads page":/projects/tony/files. On Windows you can simply download and execute the installer (.msi), and on OSX you download the disk image (.dmg), mount it and drag "Tony" into your Applications folder.

For Linux users with ample compilation experience we also provide the source code, available on the "Downloads page":/projects/tony/files.

h3. User Interface

Below is an annotated screen shot of the Tony user interface. We will refer to these elements in the paragraphs below.

!/attachments/download/1114/Tony_User_Interface.png!

h3. Tutorial 1: From Zero to Annotating Notes in an Audio File

This mini-tutorial walks you through the complete process of using Tony to load an audio file, annotate the notes in it and export the notes to a .csv (comma-separated values) file.

# Start the _Tony_ program by double-clicking (or whatever you routinely do to open programs).

# Open an audio file of your choice. Mind that _Tony_ only deals with single melodies *without* accompaniment, so choose a little bit of accompanied solo singing or the like.

# Upon opening of the file, _Tony_ will automatically analyse the file using "pYin":/projects/pyin pitch and note extraction, and pitches and notes should appear. _If no pitch track and notes occur, please see Analysis Settings and Visualisation, below._

# Get comfortable moving around and playback... a few suggestions:

#* play audio by clicking on the play button in the top toolbar, or simply by pressing the space bar on your keyboard

#* use the Navigate tool to move around in the audio file: simply left-click and drag in the main pane

#* try navigating with the keyboard, too. The left and right cursor keys will move the play head

# Get comfortable with different visualisations and sonifications:

#* try switching (toggling) different visualisations on and off in the bottom toolbar: waveform, pitch track (the black line), note track (the blue boxes), spectrogram (the gray-scale background)

#* try switching on the pitch track sonification (sound representation) by toggling the loudspeaker symbol next to the pitch track symbol in the bottom toolbar -- you should be able to hear a funny sine-like tone at the same pitch as the original audio when sonification is on.

#* turn off pitch track sonification, and turn on note track sonification -- what's the difference?

#* you can switch the original audio on and off too (next to the waveform)

#* what effect does the pan wheel (the one with the green circle) have on sonification? -- you can choose where in the stereo image to sonify the sounds: left, right, or anywhere in between.

# Note correction. Tony will have provided you with a first, dirty note track which you can correct -- *that's what _Tony_ is all about*.

#* Make sure you can see the notes. If not, turn note visualisation on.

h2. Tony User Documentation

h3. Analysis Menu

The "Analysis" menu allows you to re-analyse the audio for pitches and notes and to choose some analysis settings.

* *Analyse Now!* will run "pYin":/projects/pyin to extract pitches and notes and replaces all currently annotated pitches and notes. While you can undo this, you might still want to consider carefully whether to use this option if you have put a lot of work into annotating notes or pitches.

* *Auto-Analyse New Audio*. _Tony_ will, by default, analyse any new audio that you load. If you do not want this behaviour, you can de-select it here.

* *Unbiased Timing (slow)*. By default the "pYin":/projects/pyin analysis is based on ordinary YIN, which has a funky asymmetric window and introduces a small timing bias of up to 5ms. If you need zero time bias, then tick this item. This will be much slower, since this option cannot use the Fast Fourier Transform.

* *Penalise Soft Pitches*: this option is on by default and reduces the impact of very low-amplitude content, effectively suppressing some spurious pitches. If you don't want this, untick.

h3. Sonification

h2. Nerdy matters

[[Development Wiki]]