Concordances¶

When you run a query (see Queries), the results are displayed in a concordance view. A concordance is a structured table of matches extracted from your corpus, along with metadata and (for acoustic queries) measurement data. Concordance views provide tools for browsing, playing, filtering, editing, and manipulating query results.

Concordances are saved as .phon-conc files in XML format. They are listed in the project tree under a Concordances folder and can be reopened at any time.

Structure of the concordance view¶

Set operations¶

Concordances support set operations, which allow you to combine results from different queries:

Union (A ∪ B): combine two concordances, keeping all matches from both.
Intersection (A ∩ B): keep only the matches that appear in both concordances.
Complement (A ∖ B): keep only the matches from the first concordance that do not appear in the second.

Each of these operations opens a dialog where you select the other concordance. The result is a new concordance.

Merge¶

The Merge button performs a horizontal merge: it adds columns from another concordance or dataset to the current concordance, matching rows by position. This is useful, for instance, to combine formant measurements with pitch data for the same set of tokens.

Filtering¶

The Filter button shows or hides the filter bar below the toolbar. The filter bar lets you define one or more filter rules to restrict which rows are displayed. Each rule consists of a column, an operator, and a value. The available operators depend on the column type:

For text columns: equals, not equals, contains, does not contain, matches regex, and is one of (select from the unique values in the column).
For numeric columns: =, ≠, <, ≤, >, ≥.

Multiple filter rules are combined with AND logic: a row must satisfy all rules to be displayed. When a filter is active, the count label at the top of the table shows both the number of visible rows and the total number of rows. The Clear filters button removes all filter rules.

The Subset button creates a new concordance containing only the currently visible (filtered) rows. This is useful for extracting a subset of your data based on specific criteria.

Metric columns¶

The Metric column button opens a dialog that lets you compute a distance metric on a numeric column, optionally grouped by a categorical variable. The available metrics are:

z-score: (x − mean) / standard deviation.
Modified z-score: based on the median and median absolute deviation (MAD), more robust to outliers.
Absolute z-score and absolute modified z-score: unsigned versions of the above.
Mahalanobis distance: multivariate distance from the group centroid.

The computed metric is added as a new column. You can optionally create an automatic filter rule (e.g. keep only rows where the absolute z-score is less than 3) to identify and exclude outliers.

Vowel normalization¶

The Normalize vowels button opens a dialog for applying speaker-intrinsic vowel normalization to formant columns. Four methods are available: Lobanov, Nearey 1 (per-formant), Nearey 2 (uniform), and Watt & Fabricius. See Vowel normalization for a detailed description of each method and usage instructions.

Analyze¶

The Analyze button opens an analysis view for the concordance, where you can fit statistical models to your data.

Display settings¶

The Display settings menu controls which column groups are visible:

Match info: file name, layer index, start time, and end time.
Context: left and right context (for text concordances using the KWIC model).
Metadata: file description and properties.
Long format (acoustic queries with n-point data): toggle between wide format (one row per match, with columns F1_1, F1_2, … for each measurement point) and long format (one row per measurement point, which is often more convenient for statistical analysis). This option is only available for queries that used n-point averaging (not midpoint).
Open matches in split view: when checked, clicking View in annotation opens the annotation beside the concordance; otherwise it opens in a new tab.
Highlight targets: show the target column(s) in bold red to make them easier to identify.

Scales menu¶

For acoustic concordances (formant, pitch, or intensity queries), a Scales menu appears in the toolbar. It lets you toggle the display of values in alternative perceptual scales:

Formant concordances: toggle ERB (equivalent rectangular bandwidth) and/or Bark columns, computed on the fly from the raw Hertz values.
Pitch concordances: toggle semitone and/or ERB-rate columns.
Merged columns: if you have added acoustic columns from another concordance via Merge, separate scale toggles appear for the merged data.

Table features¶

The concordance table supports the following interactions:

Double-click on a row to play the match (if a sound file is bound).
Double-click on a base cell (match text or auxiliary column) to edit it in place. Edits are added to the undo/redo stack and can be reverted with Ctrl+Z.
Right-click on a row to open a context menu with Play match, View in annotation, Open in Praat (if configured), Bookmark, Edit event text, and Delete row(s).
Right-click on a column header to access column-level operations: Sort ascending, Sort descending, Rename column, Recode (for text columns), and Transform (for numeric columns).

Sorting¶

You can sort the concordance by any column by right-clicking on the column header and choosing Sort ascending or Sort descending.

Renaming columns¶

Double-click on a column header (or right-click and choose Rename) to give it a custom alias. This is useful for giving short, descriptive names to acoustic measurement columns (e.g. renaming F1 to F1_Hz).

Recoding columns¶

Right-click on a text column header and choose Recode… to open the recode dialog. This creates a new column where values have been remapped according to a mapping table. For example, you could recode detailed phonetic transcriptions into broader phonological categories.

Transforming columns¶

Right-click on a numeric column header and choose Transform… to open the transform dialog. This creates a new column by applying a mathematical formula to each value. For instance, you can convert Hertz values to Bark using bark(x) or log-transform durations using log(x). See Column transformations for the full list of available functions.

Applying coding protocols¶

A coding protocol is a description of a coding scheme (such as the numeric schwa coding used in the PFC project) which tells Phonometrica how to split each coded value into its constituent fields and, optionally, how to replace raw codes with human-readable labels. Protocols are stored as JSON files. See Plugins for the full specification of the protocol format.

There are two ways to attach a protocol to a concordance column. Both are available on the right-click menu of any text column header:

Apply coding protocol… opens a file picker; pick an existing .json protocol and Phonometrica creates one new column per field in the protocol, to the right of the existing columns. Raw codes are replaced with the labels defined in the protocol; values that do not match any label are kept as-is so that no information is lost.
Build coding protocol… opens the protocol builder (see below). The clicked column supplies sample input for the live preview, and the protocol can be applied directly from the dialog when you’re happy with it. This is the recommended entry point when you don’t already have a .json protocol for your coding scheme.

If some rows do not match the protocol, or contain field values that aren’t defined in the protocol, Phonometrica will display a warning summarising how many rows are affected. Non-matching rows are left blank across all new columns; rows with undefined field values keep the raw code in the affected columns.

Building a coding protocol¶

The protocol builder is an interactive dialog for creating or editing a coding protocol without writing JSON by hand. It can be launched in two ways:

From Plugins > Build coding protocol… (standalone mode). The sample area starts empty; type or paste coded values directly into it to test the protocol as you build it.
From a concordance column’s right-click menu, via Build coding protocol… The first ten values of the column are copied into the sample area automatically, and the dialog’s Apply button becomes available to apply the protocol directly to that column.

The dialog is divided into three panes. The field list on the left lets you add, remove, and reorder fields. The field editor in the middle configures the selected field: its name, the match_all regular expression (which must match every possible value of that field), and a table of match / label pairs that translate raw codes into human-readable labels. The preview on the right shows the sample text split into columns according to the current protocol. Each row is marked ✓ (fully translated), ⚠ (matched the protocol but contains one or more values not defined in the label table), or ✗ (did not match the protocol at all).

Above the preview you can set the protocol’s name, the separator that appears between fields in a coded value (leave empty when fields are concatenated directly, as in 1412), and whether the protocol is case-sensitive.

At the bottom of the dialog, Load… reads an existing protocol into the builder; the menu lists every protocol installed under a plugin, plus a Browse file system… entry for arbitrary paths. Save… writes the current protocol to a .json file via a file dialog. Apply… is enabled only when the dialog was launched from a concordance column and applies the current protocol to that column.

Keyboard shortcuts¶

Playback, stop, and row deletion in concordance views are available via the toolbar, the right-click context menu, and double-click (which plays the match when a sound file is bound). No dedicated key bindings are currently registered for these actions.

Concordances¶