how can you improve your spanish listening comprehension?
listen to less, more often, with the words in front of you at first and then without them. spanish is hard to follow by ear for reasons that have little to do with how well you read it. its words run into each other where the page keeps them apart, and it sounds noticeably different from one country to the next. what improves comprehension is knowing words well enough to recognise them inside that stream, and hearing the same stretch of speech often enough that the stream stops being a blur.
why the page is easier than the ear
spanish spelling is kind to readers. it is close to how the language sounds, so a learner who reads well can often say a word they have never heard. that same learner can be lost in a conversation, because writing puts spaces between words and speech does not.
the clearest case is what spanish grammar calls sinalefa. the real academia española defines it as the joining into a single syllable of two or more neighbouring vowels that belong to different words. a word that ends in a vowel and a word that begins with one can meet in the middle and be said as one beat.
take a line from driftspeak's spanish volume built around volver: ¿quién es? no espero a nadie ("who is it? i'm not expecting anyone."). on the page, espero and a are two words. in ordinary speech the o and the a can fold into one syllable, so what reaches you is a sound shape the page never showed you.
rhythm adds to it. spanish is usually described as syllable-timed, each syllable taking roughly its own share of time, though a stressed syllable can last up to half as long again as an unstressed one. the small unstressed words that hold a sentence together, the a in that line among them, are short, and once they are joined to their neighbours they are easy to lose. they are worth listening for on purpose, because they carry who did what to whom.
how many words you need to know
a 2013 study by hilde van zeeland and norbert schmitt measured how much of a spoken text listeners need to know to follow it. with informal spoken narratives, most native and non-native listeners understood the facts adequately when they knew 90 percent of the words, though the non-native listeners varied a great deal at that point. at 95 percent the non-native listeners did relatively well and varied far less. the authors put that at between 2,000 and 3,000 word families, against the 6,000 to 7,000 often quoted from reading research.
the study was of english, with forty non-native listeners, and it measured whether people caught facts in a story, not whether they could follow a quick conversation. it did not test spanish, so take the figures as a sense of scale rather than a target. the catch is the word "know". for listening, a word you know only by sight is barely known at all. you need it in its spoken form, at speed, joined to its neighbours.
one spanish, many sounds
spanish is the first language of people in many countries, and its sounds change from one to the next. most speakers in northern and central spain pronounce the letter z, and c before e or i, differently from s. speakers in latin america, the canary islands and parts of southern spain do not make that distinction. in many varieties an s at the end of a syllable is breathed as a soft h or dropped altogether. none of this is a mistake; each is simply how spanish sounds there.
for a listener it means two things. first, settle into one variety for a while, so that your ear has a single pattern to learn. then widen on purpose, because the spanish you meet will not stay in one country.
a repeat that does the work
choose a short piece of speech, a minute or two from a film, a podcast or an interview, and give it several days. listen once without any text and note what you caught. listen again with a transcript or same-language subtitles, and find the places where the sound and the page part company. then take the text away and listen until those places stop catching you. only after that, move on to the next minute.
it has a limit, and it is worth knowing. a recording you have heard ten times becomes easy because you have heard it ten times. the honest test is new speech of the same kind: a different scene, the same speaker or the same accent. if more of that comes through than it did a month ago, the practice is carrying over.
a red kitchen, one line at a time
driftspeak is subtitles for your ears. vol. 13 is built around pedro almodóvar's volver, thirteen tracks of short spanish sentences from the everyday life of the film's world: catching up at a graveside, a knock at the door, something to eat, buying by the kilo. each sentence is said in spanish and then in english. it gives you the minute of speech, with its meaning beside it, that the repeat above asks you to find.
to hear whether espero and a stay two words or become one sound, listen to a sample of vol. 13, whose first three tracks have free samples: the second is one line, ¿quién es? no espero a nadie ("who is it? i'm not expecting anyone."), said in spanish and then in english.