PhraseAnal is a macOS utility for creating timing metadata from an audio file and outputting it to an .epta xml file. This metadata contains timing segments used as sync points by another application. For example, the start of a syllable or letter sound.
Clone this repository, open PhraseAnal.xcodeproj in Xcode, build and run.
Drag an audio file into the default window, or choose File → Import Audio. The window shows a graphic representation of the audio waveform. Zoom in using alt + command and clicking. Zoom out with shift + alt + command and clicking. Listen to the audio by pressing space or clicking on the play arrow. A red line moves showing the current play position.
Use the segment bar below the waveform to manually define segments which represent areas of interest in the sound file. For example, consider the English audio file fc_let_kick.m4a which is a recording of the word kick split into phonemes: k, i, and ck, with a gap between each. In this simple case, the waveform shows three areas with significant amplitude, with gaps of silence between them. Define a segment for each of the three areas.
In the bottom-right of the PhraseAnal window is a text box. To automatically mark the segments in the 'kick' example, enter the phonemes delimited by spaces: 'k i ck'. Choose Edit → Autosplit. PhraseAnal will attept to automatically create three segments in the segment bar, each having a green start arrow and a red ending area. Drag the areas and the arrows to manually fine-tune each segment.
Play a segment by clicking on it. Edit by right clicking.
Once all segments have been marked, choose File → Save or Save As to save the .etpa xml file. It will contain the start and end time of each segment in seconds. The .etpa file can be edited again by opening it with PhraseAnal.