CART is a browser-based tool for completing the Consensus Auditory-Perceptual Evaluation of Voice – Revised (CAPE-Vr). Optionally, it also prepares an acoustic analysis for Praat, imports the results, and produces a combined, printable report. Optional documentation tools help draft clinical text from transparent, clinic-defined rules.
Documentation and reporting aid. It does not diagnose disorders and does not replace clinical judgment. See Disclaimers.
Version 0.1.0 · tested with Praat 7.0.02
- Online: https://maypychan.github.io/CART/
- On your own computer, with no internet needed: download this repository
(Code → Download ZIP), unzip it, and double-click
index.html.
Either way, nothing is installed, nothing is stored in the browser, and nothing you enter is sent anywhere. Ratings, recordings, and results stay on your computer, in files you save yourself. When you use the online link, only the page itself is loaded from GitHub. As with any website, GitHub may keep standard access logs of that page visit.
Chrome or Edge is recommended. For the acoustic analysis, install Praat (tested with 7.0.02). Praat is separate software and is not included.
A step-by-step guide is in docs/QUICK_START.md.
The tool follows the order of a voice evaluation:
- CAPE-Vr form: rate with visual analog scales and descriptive options, as in the published form. Save and reload sessions as files.
- Acoustics (optional): add WAV recordings, export a ready-to-run Praat
package, run it in your own Praat, and import
results.json. - Documentation (optional): an editable Clinical Summary and EMR draft, a CSV export, and rules your clinic defines.
- Report: the filled-in CAPE-Vr form, plus the acoustic results, the methods, and the Clinical Summary and EMR draft, printed or saved as PDF.
A progress bar at the top shows each step as not started, in progress, or complete.
- CAPE-Vr form.
- Scales are 0–100, with direction-only end labels and no severity labels.
- Numbers are hidden while rating by default.
- Further named attribute scales can be added below the blank one.
- The Table 1 instructions are included.
- File-based. Sessions are saved as files you download and load again. Nothing is stored in the browser.
- Praat does the analysis.
- A packaged script and settings file run in stock Praat.
- Every setting is recorded in
results.jsonand summarized in the report's methods paragraph.
- Measures:
- sustained vowels: F0, intensity, CPPS, HNR, jitter, shimmer, and F1–F3;
- sentences, extemporaneous speech, and reading: F0, intensity, and CPPS.
- Recording tools.
- Tasks are suggested from file names.
- Takes are numbered automatically, and a take can be excluded.
- A floating player lets you listen while rating.
- Transparent documentation tools.
- Every generated line shows the value and the rule behind it.
- All rules start empty and are used once filled in; categories and reference values are clinic-defined.
- Each rule statement can be unticked to leave it out of the texts.
- Rule sets can be shared as files.
- The texts can be copied once the session has been saved.
- Exports: the session (JSON), the Praat package (ZIP), the report (print or PDF), and a CSV with one row per evaluation. "Hide names" options are available.
Rate the CAPE-Vr form ─► Save session
│
├─(optional) Add WAVs ─► Make Praat package ─► run in Praat ─► Import results.json
├─(optional) Documentation: rules ─► Clinical Summary / EMR draft ─► Copy / Export CSV
└─► Generate and print report
| Output | File | Contents |
|---|---|---|
| Session | <ID>_capevr_<date>.json |
All ratings, recording metadata (not audio), settings, rules copy, results, documentation texts |
| Praat package | <ID>_praat_package.zip |
Renamed WAVs, cape_settings.tsv, cape_acoustics.praat, README.txt |
| Praat results | results.json (written by Praat) |
Every measure (or null with a reason), the settings, and the versions |
| Report | print / PDF | The CAPE-Vr form filled in, the acoustic tables, the methods, the Clinical Summary and EMR draft (each optional), and attribution |
| CSV | <ID>_capevr_<date>[_names-hidden].csv |
One row: session details, all scores, acoustic values, rule outputs |
| Rules | rules_<name>_<date>.json |
The clinic-defined interpretation rules |
Session files, packages, reports, and CSV files can contain identifying information and voice recordings. Store and share them according to your organization's policies.
index.html the application (open this)
css/app.css styles
js/ the application code, one file per task (see below)
praat/
cape_acoustics.praat the Praat script (also runs on its own in Praat)
cape_settings_template.tsv example settings file for running the script by hand
docs/
QUICK_START.md step-by-step guide
tools/
sync-praat-script.ps1 for developers only
js/ file |
Purpose |
|---|---|
session.js |
Session data, loading checks, saving |
notices.js |
Disclaimer and attribution wording |
praat_script.js |
Generated copy of the Praat script, put into each package |
upload.js |
WAV intake (header details only), tasks, takes, player, pre-export checks |
export.js |
Settings file, ZIP package, CSV |
report.js |
Printable report and acoustic tables |
import.js |
results.json checks and import |
rules.js |
Clinic-defined rules and the rule editor |
summary.js |
Rule statements, Clinical Summary, and EMR draft |
form.js |
CAPE-Vr form, progress bar, and start-up (loaded last) |
There is nothing to build and there are no dependencies.
- Not validated. There is no validation study yet against hand-measured Praat values or other software. Verify results in your own workflow.
- Browsers and platforms: tested in Chrome, with Praat 7.0.02 on Windows. Firefox, Safari, and macOS Praat have not been tested.
- Acoustic results are shown as soon as they are imported. To rate
without seeing them, complete and save the ratings before importing
results.json. - "Hide names" is not de-identification: dates, age, gender, and free text (including the Clinical Summary and EMR draft) remain.
- No norms. Categories and reference values exist only if your clinic defines them.
This software assists with documentation and reporting only.
- It does not diagnose disorders and does not replace clinician judgment.
- Acoustic results depend on recording conditions, equipment, analysis settings, preprocessing, and software versions. Users are responsible for verifying them.
- Generated text is rule-based and must be reviewed and edited before clinical use.
- Any reference values or ranges are clinic-defined. The software provides no normative values and makes no claim of clinical validity.
- It is provided as is, without warranty.
Attribution. CART implements a documentation workflow around the CAPE-Vr (Kempster, Nagle & Solomon, 2025, J Voice, doi:10.1016/j.jvoice.2025.01.022). The CAPE-Vr form content is adapted under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Modifications include electronic presentation of the form, additional documentation and reporting features, and application-specific notes. CART is independent and is not affiliated with or endorsed by the CAPE-Vr or CAPE-V authors, ASHA, The Voice Foundation, or Elsevier. Praat is separate software by P. Boersma and D. Weenink.
- Acoustic measures are computed by Praat, not by this application.
- Generated documentation remains traceable to explicit rules and reported values.
- The application operates entirely locally and does not require network access.
If you use CART in research, please cite:
- Kempster GB, Nagle KF, Solomon NP (2025). Development and rationale for the Consensus Auditory-Perceptual Evaluation of Voice—Revised (CAPE-Vr). J Voice. doi:10.1016/j.jvoice.2025.01.022.
- Boersma P, Weenink D. Praat: doing phonetics by computer (version used).
- Chan MPY, CART (CAPE-Vr Clinical Acoustic Reporting Tool). Version 1.0. GitHub repository, 2026. https://maypychan.github.io/CART/
The report's methods paragraph lists the Praat version and every analysis setting used.