German spelling tells you the pronunciation, which is a genuine gift and also the reason people stop working on it in week one. A learner reads a paragraph aloud on the first evening and is roughly right, concludes that pronunciation is dealt with, and moves on with two vowels missing, both varieties of ch wrong and a rule about final consonants never applied.
The list of things actually left to do is short and it does not shrink on its own. Two vowels to build, two consonants to separate, and one rule to start applying.
Six things to put in your mouth
- ch after a, o, u is the Bach sound, made at the back of the throat, as in Buch and auch.
- ch after i and e is softer, made at the front, as in ich and Milch. Two different sounds under one spelling, chosen by the vowel before them.
- ei is eye and ie is ee. Wein is vine, Wien is veen. Two letters in two orders, and reversing them changes the word.
- ä, ö, ü. The last two are made by saying eh and ee with the lips rounded, and they are the sounds English speakers substitute away fastest.
- the final g in Tag is a k. Voiced consonants harden at the end of a word, so Tag ends in k and Hund ends in t.
- r after a vowel is a vowel. Uhr and Bier end in a sound closer to a soft ah than to any English r.
What the spelling gives you, and what it does not
The spelling settles vowel length, which is phonemic here: Stadt and Staat are different words and the difference is length. A double consonant after a vowel means the vowel is short; an h after it, or a doubled vowel, means it is long. That covers almost every word and it is the one rule to learn before any sound.
What the spelling does not give you is a pronunciation you already own. The rounded front vowels are the two to build: ue is the ee of see with the lips rounded, and oe is the e of bed with the lips rounded, and both are made by holding a vowel you have and moving only the lips. The ch is two sounds rather than one, and which one you get is decided mechanically by the vowel in front: after i, e, ae, oe, ue and the consonants it is the light one at the front of the mouth, and after a, o and u it is the rough one at the back. Nobody has to choose. The word chooses.
The one rule to start applying is final devoicing. A b, d or g at the end of a syllable is pronounced p, t or k, so Rad and Rat are identical, Tag ends in a k sound, and halb ends in a p. English does not do this and an English speaker will keep the voicing without noticing, in every one of those words, for years.
The language rewards you too early
This is the stall, and it is caused by the regularity rather than by any difficulty. In most languages a beginner reading aloud is obviously wrong and knows it, so the work continues. In German a beginner reading aloud is broadly right, gets told so, and stops. The remaining errors are the ones the spelling cannot warn you about, and every one of them survives because nothing in daily use exposes it.
The specific casualties are predictable. The rounded vowels get replaced by the nearest sound you own, so schon and schoen merge, Mutter and Muetter merge, and the pairs almost never occur in the same sentence, which means context repairs the loss every time and the merge is never noticed. The test that exposes it takes a minute: dictate a pair of them to a phone and see which word appears on the screen. A recogniser has no interest in being polite, which is the whole reason it is useful, and it is the same principle described in the stalls every learner hits, and the way through each: get a measurement from something that is not you.
What the page is telling you
The spelling cues, sorted by what they settle
Almost every pronunciation question in German is answered by something visible in the word itself, which is not true of English or French.
A written German word
- Vowel lengthPhonemic, so Stadt and Staat are different words
- Double consonant after it: short
- An h after it: long
- A doubled vowel: long
- Which chDecided by the vowel in front, never by choice
- After i, e, ae, oe, ue: the light one
- After a, o, u: the rough one at the back
- At the start of a borrowed word: usually k
- The end of a syllableWhere the voicing rule applies
- b becomes p, as in halb
- d becomes t, so Rad and Rat are one sound
- g becomes k, as in Tag
What the tree does not contain is the two rounded vowels, because no spelling cue helps with those: the letters tell you they are there and nothing on the page tells you how to make them.
Two ways to check yourself
The dictation key on a phone keyboard, set to German, is the cheap instrument. Say schon and then schoen and read what appears. It replaced asking a German speaker whether that sounded right, which reliably produces encouragement, because a native speaker understood you from the context and is not going to interrupt an afternoon to discuss a vowel. Free and already installed. Treat a wrong transcription as a real finding and a right one as weak evidence, since recognisers in 2026 are trained on accented speech and are forgiving.
YouGlish is the other one and it answers questions a rule cannot. Type a word and it plays real speakers saying it inside real sentences, one after another, so the two varieties of ch stop being a description and become a sound you have heard forty times. It replaced the single recording attached to a dictionary entry, which is one speaker on one day and cannot show you what happens to the word in a sentence. Free.
The mistake: pronouncing every letter that is written
German spelling is regular, which invites a learner to say what is on the page, and there are a few places where the page is not asking for that. The r after a vowel is not a consonant at all in most German speech: Uhr and Bier end in a vowel-like sound rather than in anything an English speaker would call an r, and pronouncing an American r there is one of the loudest accent markers available.
The other case is the final e in words like Katze and heute, which is not silent as in French and not full as in Spanish, but a weak neutral vowel, and giving it a full value makes the word sound recited. Neither of these is difficult and neither is in the spelling rules a course teaches, which is why they persist in speakers who are otherwise entirely correct. Both are caught immediately by the recording comparison above, and by nothing else.
What sounding right does not settle
Being intelligible is not the same as following what comes back, and the standard form you have learned to produce is not what a large number of German speakers use at home, which is a separate problem taken up in the pages on the dialects. The endings you now pronounce correctly still have to be the right endings, and choosing them depends on gender and case, which is nominative to genitive, and which one a preposition takes. Getting the sentence out at the speed a conversation runs at is building a sentence when the verb is at the end, and where the stress falls on a verb decides whether its prefix detaches, which quietly makes this page part of compounds you can decode, and the ones you have to learn. Getting somebody to be blunt with you about any of it is the arrangement in what a teacher is worth, and what you can do alone.