In
phonetics, a broad transcription is a
transcription of a language into
phonetic symbols, with enough information to distinguish between different
phonemes (basic sounds of the language), but without showing the precise shade of sound.
A narrow transcription uses extra symbols and diacritics to make much finer distinctions, and would enable you to pronounce it accurately, with less of a foreign accent. The difference between broad and narrow transcriptions is a continuum, and somewhat subjective.
Phonetic transcriptions are enclosed in square brackets. The usual set of symbols used is the IPA. On the Web we have to use an ASCII version of the IPA, either SAMPA or IPA/ASCII.
In English, the words bid and bead have quite similar vowels. The one in bead is longer, and is pronounced with the tongue a little bit higher in the mouth. Either one of those facts is enough to distinguish the two vowels. So we could mark the difference in length and say the two words are [bid] versus [bi:d], where the colon indicates vowel length; or we could use different symbols for the two vowels and write [bId] and [bid], without needing to mark the fact about English that [I] is always short and [i] is (in a stressed syllable anyway) always long.
A narrower transcription, however, would indicate both facts together by writing [bId] and [bi:d]. This level of marking is fairly common. Even narrower transcriptions, less commonly used, could note other facts such as a pitch difference, a difference in the length of the closing [d], and a difference in the position of the tongue root between the two vowels.