American English

American English is a set of dialects of the English language used mostly in the United States. Approximately two-thirds of native speakers of English live in the United States.
English is the most common language in the United States. Though the U.S. federal government has no official language, English is the common language used by the federal government and is considered the de facto language of the United States because of its widespread use. English has been given official status by 28 of the 50 state governments.
The use of English in the United States was a result of British colonization. The first wave of English-speaking settlers arrived in North America in the 17th century. Since then, American English has been influenced by the languages of the Native American population, the languages of European and non-European colonists, immigrants and neighbors, and the languages of slaves from West Africa

Phonology

Compared to English as spoken in England, North American English is more homogeneous. Some distinctive accents can be found on the East Coast (for example, in eastern New England and New York City), partly because these areas were in close contact with England and imitated prestigious varieties of British English at a time when these were undergoing changes. In addition, many speech communities on the East Coast have existed in their present locations for centuries, while the interior of the country was settled by people from all regions of the existing United States and developed a far more generic linguistic pattern.

Most North American speech is rhotic, as English was in most places in the 17th century. Rhoticity was further supported by Hiberno-English, West Country English and Scottish English as well as the fact most regions of England at this time also had rhotic accents. In most varieties of North American English, the sound corresponding to the letter r is a retroflex [ɻ] or alveolar approximant [ɹ] rather than a trill or a tap. The loss of syllable-final r in North America is confined mostly to the accents of eastern New England, New York City and surrounding areas and the coastal portions of the South, and African American Vernacular English. In rural tidewater Virginia and eastern New England, 'r' is non-rhotic in accented (such as "bird", "work", "first", "birthday") as well as unaccented syllables, although this is declining among the younger generation of speakers. Dropping of syllable-final r sometimes happens in natively rhotic dialects if r is located in unaccented syllables or words and the next syllable or word begins in a consonant. In England, the lost r was often changed into [ə] (schwa), giving rise to a new class of falling diphthongs.Furthermore, the er sound of fur or butter, is realized in AmE as a monophthongal r-colored vowel (stressed [ɝ] or unstressed [ɚ] as represented in the IPA). This does not happen in the non-rhotic varieties of North American speech.
Some other English changes in which most North American dialects do not participate:

The shift of /æ/ to /ɑ/ (the so-called "broad A") before /f/, /s/, /θ/, /ð/, /z/, /v/ alone or preceded by a homorganic nasal. This is the difference between the British Received Pronunciation and American pronunciation of bath and dance. In the United States, only eastern New England speakers took up this modification, although even there it is becoming increasingly rare.
The realization of intervocalic /t/ as a glottal stop [ʔ] (as in [bɒʔəl] for bottle). This change is not universal for British English and is not considered a feature of Received Pronunciation. This is not a property of most North American dialects. Newfoundland English is a notable exception.

On the other hand, North American English has undergone some sound changes not found in other varieties of English speech:

The merger of /ɑ/ and /ɒ/, making father and bother rhyme.This change is nearly universal in North American English, occurring almost everywhere except for parts of eastern New England, hence the Boston accent.
The merger of /ɑ/ and /ɔ/. This is the so-called cot–caught merger, where cot and caught are homophones. This change has occurred in eastern New England, in Pittsburgh and surrounding areas, and from the Great Plains westward.

For speakers who do not merge caught and cot: The replacement of the cot vowel with the caught vowel before voiceless fricatives (as in cloth, off [which is found in some old-fashioned varieties of RP]), as well as before /ŋ/ (as in strong, long), usually in gone, often in on, and irregularly before /ɡ/ (log, hog, dog, fog [which is not found in British English at all]).
The replacement of the lot vowel with the strut vowel in most utterances of the words was, of, from, what and in many utterances of the words everybody, nobody, somebody, anybody; the word because has either /ʌ/ or /ɔ/; want has normally /ɔ/ or /ɑ/, sometimes /ʌ/.

Vowel merger before intervocalic /ɹ/. Which vowels are affected varies between dialects, but the Mary-marry-merry, nearer-mirror, and hurry-furry mergers are all widespread. Another such change is the laxing of /e/, /i/ and /u/ to /ɛ/, /ɪ/ and /ʊ/ before /ɹ/, causing pronunciations like [pɛɹ], [pɪɹ] and [pjʊɹ] for pair, peer and pure. The resulting sound [ʊɹ] is often further reduced to [ɝ], especially after palatals, so that cure, pure, mature and sure rhyme with fir.
Dropping of /j/ is more extensive than in RP. In most North American accents, /j/ is dropped after all alveolar and interdental consonant, so that new, duke, Tuesday, resume are pronounced /nu/, /duk/, /tuzdeɪ/, /ɹɪzum/.
æ-tensing in environments that vary widely from accent to accent; for example, for many speakers, /æ/ is approximately realized as [eə] before nasal consonants. In some accents, particularly those from Baltimore, Philadelphia, and New York City, [æ] and [eə] contrast sometimes, as in Yes, I can [kæn] vs. tin can [keən].
The flapping of intervocalic /t/ and /d/ to alveolar tap [ɾ] before unstressed vowels (as in butter, party) and syllabic /l/ (bottle), as well as at the end of a word or morpheme before any vowel (what else, whatever). Thus, for most speakers, pairs such as ladder/latter, metal/medal, and coating/coding are pronounced the same. For many speakers, this merger is incomplete and does not occur after /aɪ/; these speakers tend to pronounce writer with [əɪ] and rider with [aɪ]. This is a form of Canadian raising but, unlike more extreme forms of that process, does not affect /aʊ/. In some areas and idiolects, a phonemic distinction between what elsewhere become homophones through this process is maintained by vowel lengthening in the vowel preceding the formerly voiced consonant, e.g., [læ:·ɾɹ̩] for "ladder" as opposed to [læ·ɾɹ̩] for "latter".
T-glottalization is common when /t/ is in the final position of a syllable or word (get, fretful: [ɡɛʔ], [ˈfɹɛʔfəl]), though this is always superseded by the aforementioned rules of flapping
Both intervocalic /nt/ and /n/ may be realized as [n] or [ɾ̃], making winter and winner homophones. Most areas in which /nt/ is reduced to /n/, it is accompanied further by nasalization of simple post-vocalic /n/, so that V/nt/ and V/n/ remain phonemically distinct. In such cases, the preceding vowel becomes nasalized, and is followed in cases where the former /nt/ was present, by a distinct /n/. This stop-absorption by the preceding nasal /n/ does not occur when the second syllable is stressed, as in entail.
The pin-pen merger, by which [ɛ] is raised to [ɪ] before nasal consonants, making pairs like pen/pin homophonous. This merger originated in Southern American English but is now also sometimes found in parts of the Midwest and West as well, especially in people with roots in the mountainous areas of the Southeastern United States.

Some mergers found in most varieties of both American and British English include:

The merger of the vowels /ɔ/ and /o/ before 'r', making pairs like horse/hoarse, corps/core, for/four, morning/mourning, etc. homophones.
The wine–whine merger making pairs like wine/whine, wet/whet, Wales/whales, wear/where, etc. homophones, in most cases eliminating /ʍ/, the voiceless labiovelar fricative. Many older varieties of southern and western American English still keep these distinct, but the merger appears to be spreading.