The real shape of the problem
Mandarin pronunciation has a reputation it only half deserves. The tones get all the fear, but tones are a listening skill with a dedicated trainer on this site, and most learners crack them faster than expected. The quieter problem is the consonants and vowels, because pinyin - the romanisation everything is taught through - reuses familiar letters for unfamiliar sounds. The letter x is not English x. The letter q is not English q. The letter c is "ts". Read pinyin as if it were English and you are wrong on a third of the alphabet before tones even enter into it.
The fix is to learn what perhaps a dozen pinyin spellings actually do with the mouth, and the lab above animates the ones that need seeing. For the full letter-by-letter table, the pinyin guide is the reference page.
The two families: zh/ch/sh against j/q/x
This is the heart of Mandarin pronunciation, and the pair of sound families English speakers most reliably merge.
zh, ch, sh (and r) are retroflex. The tongue tip curls up and back, well behind the ridge where English sh lives, and the sound comes out darker and duller. "Zhōngguó" (China) opens with one. Watch the side view in the lab: the visible curl of the tongue tip is the entire instruction.
j, q, x are palatal. The tongue tip sits down behind the lower teeth and the BODY of the tongue presses up towards the hard palate. They are brighter, hissier sounds, always followed by i or ü. "Xièxie" (thank you) is two of them.
To an English ear both families sound like versions of "sh", "ch" and "j". To a Mandarin ear they are as different as s and sh are to yours: "shī" and "xī" are different words, not accents of the same word. The ear drill above tests exactly this pair until it stops being guesswork - and because the two families never share a following vowel, your mouth gets a free hint: retroflexes go with a, e, o, u; palatals go with i and ü.
The ü vowel
Say "ee", hold your tongue still, round your lips: that is ü, the same sound as French u. "Nǚ" (woman) and "lǜ" (green) need it. The catch is spelling: after j, q, x and y, pinyin writes the dots away - "qu" is really q + ü, and "ju", "xu", "yu" likewise. After n and l the dots are kept precisely because "nu" and "nǚ" are both real syllables. The lab's slider morphs /u/ into /y/ so you can watch that only the tongue position separates "bù" (not) from the vowel in "qù" (go).
The letters pinyin quietly redefines
- c is "ts" as in "cats": "cài" (vegetable) starts /tsʰ/.
- z is "dz": "zài" (at) starts /ts/.
- r at the start of a syllable is a buzzy retroflex, somewhere between English r and the s in "measure": "rè" (hot).
- e alone is not "eh" but a mid, throaty vowel: "hē" (drink) rhymes with nothing in English; the nearest is the vowel of "duh" said with a straight face.
- i after retroflexes and after z/c/s is not "ee": "shì" and "zì" end in a buzzed continuation of the consonant, not a vowel you know. This is the single most common pinyin misreading.
- -ian is "yen", not "ee-an": "diǎn" (o'clock).
Aspiration: the b/p, d/t, g/k trick
Mandarin b, d, g are not English b, d, g - they are unaspirated p, t, k (no puff of air, no voicing). The real contrast in pairs like "bà"/"pà" is the puff: p, t, k, q, ch, c all carry a strong burst of breath, their partners carry none. Hold a hand in front of your mouth: English "pie" puffs, English "spy" does not - the p of "spy" is the Mandarin b. Your mouth already makes every one of these sounds; only the sorting is new.
Where the tones fit
Every syllable carries one of four tones (plus a neutral one), and tone changes meaning: mā, má, mǎ, mà are four different words. This page deliberately does not duplicate tone teaching - the tone trainer is a listening drill built on real recordings, and tones are best learned by ear, not by diagram. Do the sounds here, do the tones there, and meet both inside real words in the Mandarin course.
Frequently asked questions
Is Mandarin pronunciation hard for English speakers?
The individual sounds are mostly easy - no rolled r, no clusters like "strengths". The hard parts are the retroflex/palatal split, reading pinyin honestly instead of as English, and tones. All three are pattern-recognition problems that yield to drills, which is exactly what the lab above and the tone trainer are.
What is the difference between x and sh in Mandarin?
Place of articulation. For sh the tongue tip curls up and back (retroflex); for x the tip stays down and the tongue body arches to the hard palate (palatal). x is the brighter, thinner hiss. "Shī" (teacher's shī, as in lǎoshī) and "xī" (west) differ by nothing else.
How is "xièxie" (thank you) actually pronounced?
Not "zee-eh-zee-eh" and not "she-she". The x is the palatal hiss above, the ie is "ye": roughly "sye-sye" with the first syllable on a falling tone and the second unstressed and neutral.
Do I say Mandarin or Chinese?
Mandarin is the standard spoken language this site teaches (pǔtōnghuà); Chinese is the family it belongs to, which also contains Cantonese, Shanghainese and many others. For pronunciation purposes the distinction matters: this page describes Mandarin, and other Chinese languages sound very different.