N° 048

June 3

La Phrase du Jour

A single French sentence, scored sound by sound.

Je ne comprends pas

Honest and useful — no shame in it.

Tap to record. Read the phrase out loud.

Nouveau demain à minuit

Notre méthode · Our methodology

How the test actually works.

01 · The model

A scorer trained to do the opposite of dictation.

Most speech recognition models are designed to ignore accent. Their job is to understand anyone, regardless of region, age, or first language. Ours does the opposite. We trained a deep neural network from scratch to do one thing well: hear your French the way an average Parisian would, sound by sound. It's the same family of model behind modern voice assistants and voice cloning, pointed at the opposite job. The network listens phoneme by phoneme, in context, and tells you which sounds an actual Parisian would let pass and which ones they'd flag.

02 · The training data

Native Parisian speech, not generic French.

The network was trained exclusively on native Parisian speech — not generic French, not Wikipedia commons, not Quebec, not Marseille. Parisian. Specifically. The reason is calibration: "correct French" isn't one accent, and our users are aiming for one in particular. Training on a regional mix would teach the network to forgive features Parisians actually flag. Training on Parisian only means the score corresponds to what a Parisian's ear would do in real life.

03 · The linguistic review

Every coaching tip, reviewed by a native.

Every coaching tip — the mouth-mechanics descriptions you see when you miss a sound — was reviewed by a native Parisian linguist on our team before the test went live. The criteria: would a Parisian recognize this description? Would following it actually produce the right sound? Tips that didn't pass got rewritten. The model is technical; the coaching prose is reviewed by an actual native speaker who can hear when an instruction is right.

04 · Why it's free

Confidence in the engine.

We put the model on the open web, no signup, because we have to be confident in it. If your score doesn't match what you'd hear in Paris tomorrow, the test is broken — and we'd hear about it within a day. The full coaching layer (daily training, mouth-mechanics walkthroughs, progress over time) lives in the Parlez app on iOS, but the scoring engine you're using right here is the same one that powers it.

Questions fréquentes · French accent FAQ

About the French accent.

Why is the French accent so hard for English speakers?
The honest answer is geometry. French uses three sounds your mouth has never been asked to make in English — front rounded vowels like the /y/ in "tu" (lips like "oo," tongue like "ee," at the same time), and the three nasal vowels in "un," "an," "on" (the vowel itself goes through your nose; there's no consonant at the end). English has neither feature. On top of that, French rhythm is syllable-timed where English is stress-timed, which means every syllable in French gets roughly equal weight. Get those four things wrong simultaneously and a Parisian clocks you instantly. Get them right and you sound French — even if your grammar is a mess. Take the test and you'll hear which ones you're missing.
What's the hardest French sound to pronounce?
Most learners think it's the uvular /ʁ/ in "Paris" — the gargly R — because it's the most obviously different. Our resident native Parisian disagrees: she says it's the front rounded /y/ (the vowel in "tu," "vu," "rue"). English speakers default to either the /u/ in "too" (lips right, tongue wrong) or the /i/ in "see" (tongue right, lips wrong), and the difference between "tu" and "tout" can change what you're saying. Close behind are the three nasal vowels — /ɛ̃/ in "pain," /ɑ̃/ in "France," /ɔ̃/ in "bon" — because English doesn't have nasal vowels at all. The French /ʁ/ is third on the difficulty list, not first. Try a phrase and find out which one trips you up.
Can adults learn to sound native in French?
Near-native, yes. The "critical period" idea — that you can't fully acquire a new sound system after puberty — is much weaker than pop-science articles suggest. Recent research is pretty clear: adults can reach a level where untrained native ears can't reliably tell, especially for languages they have prolonged exposure to. What adults can't do well is acquire it the way kids do — by ambient exposure alone. They need explicit, specific feedback on which sounds they're missing and how to physically produce them. That's the actual bottleneck, not biology. The reason most adults plateau is they never get that feedback at the phoneme level. Build the habit of getting it daily and the ceiling moves a lot higher than you'd expect. Start with a phrase and see.
How long does it take to lose an English accent in French?
It depends on what "lose" means. Reaching the point where Parisians stop switching to English on you takes most learners a few months of focused work — not years. Reaching the point where they can't tell where you're from until you mention it takes longer, usually six months to two years of deliberate practice on the specific sounds you miss. The "it takes a decade" myth comes from people who never got pointed feedback — they practiced for years and improved slowly because they were guessing what to fix. Targeted feedback compresses the timeline dramatically. Our team's native Parisian has watched students go from obviously American to plausibly Belgian inside a year. Test where you are now and we'll show you what's left.
Do French people care if you have an accent?
Short answer: yes, and it's complicated. French speakers — Parisians especially — are unusually attentive to accents because their education system drills "le bon français" hard. They notice immediately. The good news: most appreciate effort and warm up the moment they hear you trying. The bad news: they'll often switch to English to be "helpful," which is the worst thing for a learner. Our native Parisian's blunt take: she'd rather hear someone struggle in French than default to English — switching is well-meant but it stings. The fix isn't to demand French; it's to get good enough that switching feels unnecessary. That's a few months of work on the specific sounds Parisians flag, not years. See how close you already are.
What's the difference between Parisian and other French accents?
Parisian French is what news anchors, dubbed films, and most teachers default to — it's the standard. Quebec French uses different vowels (long /a/ in "pâte" diphthongizes), preserves older pronunciations Parisians have dropped (you'll hear "moé" for "moi" — a kept /wɛ/ Paris shifted to /wa/), and has its own slang. Marseille and Provençal accents are sing-song with brighter vowels, a velar tail on nasals — "demain" lands closer to "demaing" — and a tendency to pronounce the silent schwa Parisians drop. Swiss French is the closest to Parisian, with mostly prosodic differences and its own number system ("septante," "nonante"). Belgian French sits a hair further out, with subtler nasal differences. To French ears these aren't subtle — Parisians can place a French speaker by region within a sentence or two. Our model is trained specifically on Parisian, so it'll flag Quebec or Marseille features as "wrong" even though they're perfectly correct French. That's deliberate.
How do I practice French pronunciation alone?
Three things in order. First, shadowing: play a short clip of native French audio, pause after each phrase, repeat it out loud immediately. Do this with podcasts, news, films — anything natively spoken. Second, record yourself. You'll notice things in playback you can't hear in real time. Third — and this is where most self-study breaks down — you need feedback that goes beyond "that sounds about right." You can't reliably hear the gap between your /y/ and a Parisian's /y/ from a recording; the difference is too small for an untrained ear. That's the whole reason we built this — the daily test gives you a per-phoneme score so the gap is named, not guessed at. Two minutes a day beats two hours a week.