Real speakers. Curious ears.

Find the voice.
Then train the ear.

What recordings are you working with? Start with an archive example or your own permitted audio. Compare what you hear, choose sound targets, and turn a short phrase or scene into focused practice.

Teacher video · 4:01 →Student video · 5:07 →
Joy of Phonetics: an open mouth projects the project name and sound marks.
—IDEA samples
—GMU indexed samples
—GMU languages

IDEA and GMU are discovery examples, not the limits of the practice workflow. You can also load a local reference in Audio Lab. Counts describe indexed records, not verified playable audio. External recordings require internet access.

Loading archives…
Video guides · prototype walkthroughs

Listen first.
Notice more.

Watch the teacher or student guide, then try the workflow yourself. The optional text lesson below connects Discover, Audio Lab, and Coach. No account is needed.

Draft text lesson by Sergey Ulyanov. Louis Colaianni advises the project's phonetics and pedagogical design; this does not make the lesson a published or author-approved course, an excerpt from his books, or evidence of measured learning gains.

These independent prototype walkthroughs use synthetic male narration and the example “Discover the joy of phonetics.” They show the controls, not an accent model or automatic assessment. Notes save in this browser; audio is not included in session exports. Videos need an internet connection. Review with Louis is still pending.
4:01 · English · step by step

For speech teachers

Turn one permitted recording into a teaching task: load a reference, choose targets, edit a drill, and save a session.

Open video separately ↗ · English captions · Watch on YouTube ↗

Read the teacher video transcript

Let's turn a recording into a practice exercise. We'll load one file, create a drill for one line, and save the result.

On the Discover page, click Start with my own recording. This opens Audio Lab. You can also reach it from the top menu.

Find the section called Your recording. Under Choose local reference audio, click the file chooser and select the audio file on your computer.

Its name and a player should now appear. The file stays in this tab; it is not uploaded.

Press the triangular Play button. This synthetic clip demonstrates the workflow, not an accent model.

Discover the joy of phonetics.

Now click Use this reference in Coach.

Coach Workspace opens with the filename in the recording reference field. Leave Archive sample empty; we're using our own file.

In Short phrase or scene line, type Discover the joy of phonetics.

Under Teaching context, give the session a name.

In Teaching task, write Invite a colleague to explore the resource.

In Sound target one, write Explore the J sound at the start of joy.

In target two, write Move the main emphasis from JOY to PHONETICS.

In target three, write Try a rising ending, then a falling ending. One target is enough for a first session; the other two are optional.

Click Build drill. The draft appears in the third panel, Drill.

This copies your phrase and targets into a starting structure. It has not analysed your recording.

Click inside the drill text and add your own teaching directions: Listen once. Practise one feature at a time. Then say the whole line to your imagined colleague.

Have the student try one choice, then return to the whole line. Choose what serves the invitation; these are not universal accent rules.

Scroll down to Observations and next changes. Write what was actually heard and one question for the next lesson.

Keep the teacher's comments in Expert feedback, and only with permission. This demonstration does not invent student observations or expert comments.

Return to the Drill panel and click Save session.

Check for Session saved in this browser. This saves notes locally in this browser, not the audio file.

Your session should also appear below under Saved sessions. Use Open to return to it later in this browser.

At the top of Coach Workspace, click Export session. Your browser downloads a JSON backup of the session notes.

Keep this backup and the original recording separately. Audio is not included in the backup or a share link.

After a reload, choose the audio file again in Audio Lab. Review your notes before sharing anything.

You now have a recording reference, a clear practice task, and a saved backup. The teacher and student make the performance choices; the prototype does not give an automatic accent score.

5:07 · English · step by step

For acting students

Practise one line between lessons: listen, explore one sound target, save observations, and export a backup.

Open video separately ↗ · English captions · Watch on YouTube ↗

Read the student video transcript

Let's practise one line between lessons. We'll choose a recording, make a short exercise, and save our notes. You can pause this video at any step.

On the Discover page, click Start with my own recording. Audio Lab opens. You can also choose Audio Lab from the top menu.

Under Your recording, find Choose local reference audio. Click the file chooser and select a recording you have permission to use.

Check that the filename and an audio player appear. The file stays in this tab; it is not uploaded.

Press the triangular Play button and listen once. Our short example is synthetic: a workflow demonstration, not an accent model.

Discover the joy of phonetics.

Click Use this reference in Coach. Coach Workspace opens with the filename in the recording reference field. Leave Archive sample empty.

In Short phrase or scene line, type Discover the joy of phonetics.

In Teaching context, name your practice: Invitation practice, between lessons. This is a session title, not an account name.

The next field is called Teaching task. Use it for your intention: Invite a colleague to explore the resource.

In Sound target one, write Compare emphasis on JOY and PHONETICS. We are exploring one performance choice, not trying to copy an entire accent.

Leave targets two and three blank for this first exercise. You can add another focus later with your teacher.

Click Build drill. Your phrase and target appear in the third panel, Drill. This is a starting structure, not an analysis of your voice.

Click inside the drill text and add these steps: Listen once. Say the line with emphasis on JOY. Then say it with emphasis on PHONETICS.

Now say the whole line to an imagined colleague. Which emphasis better serves your invitation? Pause the video and try both. There is no automatic right answer.

You do not need to record yourself to finish. If you want to explore the optional recorder, click Listen or record a practice take.

In Audio Lab, scroll down to Record locally. This video shows the controls only; it does not turn on your microphone or claim a recorded take.

Choose Coach in the top menu to return to your exercise. Your draft is still there.

Scroll to Observations and next changes. Write one thing you actually heard and one question for your teacher. Our screen shows an example question, not a real learner result.

Leave Expert feedback blank unless your teacher has given comments and permission to keep them. Skip the optional ratings if they are not useful.

Return to the Drill panel and click Save session. Check for Session saved in this browser.

Scroll to Saved sessions. Your named entry should appear. Click Open to return to that saved version in this browser.

For ordinary exercise text, click Copy drill, then paste it into your own notes. It copies the drill, not your audio or all the session fields.

For a structured backup, go to the top of Coach Workspace and click Export session. Your browser downloads a JSON file containing the session notes.

This is a data file, not a PDF or an audio recording. Keep the original audio separately. Coach has no import button for this single-session file.

Saved sessions belong to this browser, not an online account. Do not rely on them surviving cleared browser data or appearing on another device. Keep your exported files.

After reloading, choose your audio file again in Audio Lab. Review your notes before sharing them with anyone.

Bring your observation and question to the next lesson. Keep your own voice, and choose what helps your performance.

For speech teachers · lesson preparation and homework
  1. Ask: what recording or scene is the student working with? Choose an archive example, or load permitted local reference audio in Audio Lab. Arbitrary archive imports are not automatic.
  2. Name up to three sound changes for that task. In Coach, enter the recording reference, a short phrase, the teaching context, and up to three sound targets. Build and edit a short drill. An indexed sample is optional for a manual task.
  3. Listen and practise in the lesson. Keep your feedback separate from the student's observations. Export the session as homework; audio is not included in a session link, so share recordings separately only with permission.

For acting students · independent practice between lessons
  1. Choose a role, scene or communication task. Use a teacher's reference, your own permitted audio, or an archive example. Load local audio in Audio Lab; browser format support may vary.
  2. Listen to one short phrase. Describe one feature in this recording, rather than generalizing about a whole country. In Coach, keep the recording reference and scene line, name up to three sound targets, and build or edit a drill.
  3. Optionally record a local practice take, note what you hear, and write one question for your teacher. Export a backup and bring your observations to the next lesson. Self-ratings are subjective; the app does not automatically grade your accent.

Text lesson · 01

Hear a person, not a category.

A country or first-language label is a starting point for finding a recording—not a rule for how everyone in that group speaks.

  1. Open Discover. Choose a GMU sample and read its source information.
  2. Add it to Audio Lab. Listen once without trying to imitate.
  3. Listen again. Choose one short phrase and describe what you hear: a vowel, an /r/, a rhythm, or a pause.

“In this recording, I notice…” is more useful than “People from this place always…”

Use GMU’s shared passage for the comparison in step 02. IDEA samples open on their original website.

Listen · compare · imitate

Audio Lab

Start with your own permitted recording, or compare up to four archive references from Discover. No archive sample is required for a practice task.

GMU-indexed audio uses external stream URLs; optional detailed metadata comes from a third-party mirror. Check the original source and its terms. IDEA audio opens at IDEA. This app does not re-host archive audio or guarantee playback.
1 · Your recording

What recording are you working with?

Use a permitted teacher reference, rehearsal recording, or audio from your own collection. To work from another database, open its source yourself; automatic import is not supported.

The audio stays in this tab. It is not uploaded, saved in the session, or included in a share link. Browser format support varies; reloads require choosing the file again.

No local reference selected.

Standardized comparison

GMU elicitation passage

Please call Stella. Ask her to bring these things with her from the store: Six spoons of fresh snow peas, five thick slabs of blue cheese, and maybe a snack for her brother Bob. We also need a small plastic snake and a big toy frog for the kids. She can scoop these things into three red bags, and we will go meet her Wednesday at the train station.

Speech Accent Archive, George Mason University · CC BY-NC-SA 4.0.

Your take

Record locally

Nothing uploads
00:00
Actor / coach workflow

Coach Workspace

Document a teaching task, try a source, and keep observations separate from expert feedback.

Expert walkthrough · before you start
  1. Ask permission before keeping comments; do not record the meeting without separate permission.
  2. Let the expert choose one real teaching task and try it. Note where help was needed.
  3. Keep observed actions, expert opinions, and your proposed changes distinct.
  4. Agree on up to three next changes and, separately, whether the expert may be named publicly.

Prototype review, not a learning-outcomes study. Notes stay in this browser; export a backup. A share link includes coaching notes—review them before sharing.

1 · Recording & scene

An archive sample is optional.

Keep enough information to find the source again. Text only—no audio is attached.

2 · Up to three sound targets

What do you want to explore in this phrase? Describe choices you hear, not a rule about a whole population.

Optional phonetic prompts and word banks

3 · Drill

IPA quick insert

Edit this draft to fit your task. Listen to the reference, try the phrase, then note what changed. No automatic accent grading.

Record comments only with permission. Feedback is not an endorsement or confirmed collaboration.

Saved sessions

Reproducible annotation

Research Mode

Local annotation profiles, provenance, review flags, and exports. Profiles do not verify independent people; agreement is not a learner outcome.

Code a sample

Choose a sample.

Local profile agreement

Review queue

Activity log

Why this exists

Recordings are evidence.
A workflow makes them usable.

Joy of Phonetics helps teachers and actors turn recordings into documented listening and practice. IDEA and GMU provide discovery examples; users can also load permitted local reference audio in Audio Lab and create a manual Coach task. It does not automatically import every archive, generate accents, or score whether a voice is “correct.”

IDEA

Roughly 1,800 primary-source accent and dialect samples from around the world, with biographical details, transcriptions, and commentary on many records.

Open IDEA ↗

GMU Speech Accent Archive

Thousands of speakers reading one standardized passage. Consult the archive’s current terms before reuse; access and licensing are separate from indexing.

Open GMU archive ↗

Method

Each speaker is an exemplar, never a proxy for a nation or social group. Observation, source commentary, inference, and coaching hypothesis should remain distinct. A practice task can use a short scene as well as an archive passage.

Project and practitioner advising

Sergey Ulyanov — research framing, product design and AI-assisted implementation.
Louis Colaianni — practitioner adviser on phonetics and pedagogical design. His October 18, 2026 formative review informed clearer source-independent framing and teacher/student guidance. This is an independent prototype, not a university-endorsed program; learner outcomes and classroom adoption remain untested.

Provenance. IDEA audio is not copied or re-hosted here. GMU audio URLs point to the original archive; optional detailed metadata comes from a third-party mirror. Consult the Speech Accent Archive for provenance and terms. Availability is not verified by index counts. Use local recordings only when you have permission; session links do not include audio.

Research question

How can a source-visible recording-to-practice workflow support teachers and self-directed learners without collapsing speaker variation into stereotypes?