TextGrid Batch Inspector

Batch inspection of Praat TextGrid files: tier consistency, diagnostics, CSV exports, and TGA batch duration statistics.

v1.1.4

Upload

Select a directory, several files, or one TextGrid file. Files are processed locally in your browser.

Choose input

Drag and drop TextGrid files here, or use Select directory for folder input.

Directory selection uses browser-mediated access. Absolute local paths are not read or displayed.

Upload summary

Discovered files

Settings

Settings affect diagnostics, nPVI calculation, and TGA batch duration statistics. By default, result tables use the union of all observed interval tiers. After changing settings, re-run analysis.

Reference tier strategy
Empty-labelled intervals

nPVI now uses the confirmed pause/silence labels below, not this setting alone.

Tier-name diagnostics

Variants are reported but not silently merged.

Export privacy

Relative paths may reveal corpus folder structure. They are excluded by default.

Pause/silence labels

These confirmed labels are used by TGA batch duration statistics. The app suggests labels found in the current batch from a built-in candidate list.

Label inventory

Files

One row per selected file after parsing.

Tiers

Corpus-level tier inventory and tier-name consistency indicators.

Tier presence matrix

Rows are TextGrid files. Columns are tier names. “+” means the tier is present in the file; “-” means it is absent.

Diagnostics

Errors, warnings, and informational messages.

TGA batch

Validation-dependent duration statistics for TextGrid tiers. Each reference tier generates three CSV files: segments, pauses, and full trimmed sequences.

Export identifier

Required. The sanitized name and date are used in the ZIP filename, internal directory, and reports.

ZIP preview: Enter an export name.
Validation policy
Validation strictness
Silence and pause handling

TGA batch uses the confirmed pause/silence labels from Settings.

Pause/silence labels used: empty, _, -, #, +, sil, silence, p, #$p, pause, pau, sp

Output preview

Validation preview

TGA output tables

Preview the exact rows that will be included in the per-tier CSV files. Generate TGA output first, then select a tier and sequence type. Download the ZIP from the Downloads tab.

Downloads

Download diagnostic and batch-analysis outputs.

Generate TGA output in the TGA batch tab first.

CSV exports harden user-controlled cells against spreadsheet formula injection.

Help

What this app does

TextGrid Batch Inspector analyses Praat TextGrid files in batches. It extracts IntervalTier durations, checks tier consistency, reports annotation diagnostics, and generates TGA batch duration statistics. nPVI, rPVI, counts, and descriptive statistics are provided by the TGA batch export.

Input modes

You may select a directory, multiple TextGrid files, or one TextGrid file. Directory input shows discovered files first, and only selected TextGrid files are analysed. Folder input is supported through the Select directory button; drag-and-drop is intended for files.

Long-form and short-form TextGrids

The parser supports labelled long-form TextGrids and compact short-form TextGrids. Malformed files are reported in diagnostics instead of stopping the batch.

TGA batch statistics

nPVI, rPVI, counts, and descriptive duration statistics are produced in the TGA batch ZIP export. nPVI is calculated from adjacent durations in the non-silent annotated segment sequence. Initial/final silences are trimmed, confirmed pause/silence labels are removed, and then the mean of absolute pairwise duration differences divided by pairwise mean duration is multiplied by 100.

Reference tiers and observed tiers

The Tiers tab shows the number of interval tiers observed in the selected corpus and the number of reference tiers used for validation and TGA batch export. The default strategy is Union of all interval tiers, so all observed interval tiers are included unless you change the setting.

The Tiers tab also includes a tier presence matrix. TextGrid filenames head the rows, tier names head the columns, and cells contain “+” for present tiers and “-” for absent tiers. This is intended to make tier corrections in Praat easier to identify.

TGA batch

The TGA batch tab validates the current batch and generates duration-statistics output. For each reference tier it produces three CSV files: annotated segments only, internal pauses only, and the full sequence after initial/final silence trimming. It uses the confirmed pause/silence labels from Settings. Every TGA CSV row starts with the TextGrid filename, followed by the tier and sequence type. After generation, the tab shows on-screen previews of the generated CSV rows using tier and sequence selectors. Download the ZIP from the Downloads tab.

Privacy and security

The app runs locally in the browser. It does not upload files, does not use analytics, and does not load external scripts or styles.

Version

TextGrid Batch Inspector 1.1.4. Distributable filename: textgridbatchinspector-1.1.4-tga-output-preview.html.