rr vs r · trill and tap One difference: how many times the tongue taps. perro taps two or three times, pero once — and that single sound separates "dog" from "but".
Same position, one tap or several
Each pair differs only in how many times the tongue taps. Play the two cells in alternation and listen to how long the buzz lasts.
Four steps to the trill
Single first. The trill is just the tap, sustained.
Confirm the tap
Say American butter and notice the light touch of the tip. That isolated gesture is ɾ. Do not substitute the Mandarin "l" or the English r — neither touches the alveolar ridge.
Use t / d as a springboard
tr and dr are the easiest to start: tres, drama, otro. Said quickly, the tongue tip flickers by itself after the stop. Start there, then drop the t.
Relax and blow
Rest the tip on the ridge, keep it completely relaxed, and blow. The tip is carried by the air, not shaken by you. One flick at first is normal; a couple of weeks is a common timeline.
Put it in minimal pairs
Practise the four pairs here: perro / pero, carro / caro, forro / foro, parra / para. Check yourself against the duration gap on this page — the trill should be clearly longer. Equal length means no trill yet.
Which one are you
The three most common readings.
My rr comes out as "l" or an English r
Most common. Neither of those touches the alveolar ridge, so nothing can trill. Fix the tongue position first (ridge, not curled back), then work on the airflow.
I can tap once but not several times
More than halfway. The trill is not harder pushing — it is keeping the air going. If one tap works, continue the airstream and the second and third follow.
Fine in isolation, gone in sentences
Expected. In phrases rr is often squeezed by a preceding consonant (the l of el perro). Practise the word alone first, then with the article.
Word list
16 words grouped by where the sound sits in the word. Click a word to hear it. These are recorded as separate words, not cut out of the sentences above — a word cut from a sentence carries its neighbours' transitions and always sounds muddier. All three groups use the same speaker, so what changes between them is position, not voice.
Word-medial8
Word-final8
Corpus and acoustic measurements
Click any word in a sentence to play from that word. The pale band over the waveform marks the measured boundary of the target word.
Per-token data
Every number below is measured directly from the audio, not hand-annotated.
| Token | Accent | Sex | Target | Word length | Target rhotic duration | F0 | Pitch range | HNR |
|---|
Where these numbers come from
Every duration here comes from forced alignment over these 12 recordings.
Corpus: 12 sentences, two Peninsular voices. Four are minimal pairs: same position, same vowel, only rr versus r (perro/pero, carro/caro, forro/foro, parra/para), each pair read by the same voice.
Boundaries: MMS_FA forced alignment; every word on this page came back char-anchored. The duration is that of the rhotic itself (r or ɾ), not of the word.
Result: trill 101–131 ms, tap 40–60 ms — a 1.67–3.27× difference. The gap is a consequence of tapping more times, not a target in itself.
Where this gap applies — and where it does not: the 2–4× ratio holds only for these synthetic recordings. In our human read-speech corpus, segments labelled r (trill) have a median of 40 ms with 76% under 60 ms, while those labelled ɾ run 60 ms — the direction even reverses. Two causes: real read speech compresses trills to one or two taps, and the labels come from spelling rules (rr, word-initial r), not from acoustics. ⇒ Do not use duration to count taps; it only shows that the two differ in this audio.
Two things we did not manage, stated plainly: (1) We could not count the taps. Counting local minima of the envelope gave a median of 1 for the trill and 0 for the tap — audio compression smooths those troughs away. So this page reports duration only. (2) "Word-initial r is also a trill" did not hold in this corpus. The r of rata measured 30 ms, the same range as a tap; only the digraph rr reached 120 ms consistently, which means the synthesis engine trills only rr. Hence all four pairs are rr versus r.