The previous comment (from hans) seems to make no sense at all, since it will not change the encoding and possibly result in a "multiencoding" string (that the browser and anything else will be unable to render, of course).
I use a little function to decode the whole header to a specified encoding. It is as follow:
<?php
function mb_list_lowerencodings() { $r=mb_list_encodings();
for ($n=sizeOf($r); $n--; ) { $r[$n]=strtolower($r[$n]); } return $r;
}
function decodeMimeString($mimeStr, $inputCharset='utf-8', $targetCharset='utf-8', $fallbackCharset='iso-8859-1') {
$encodings=mb_list_lowerencodings();
$inputCharset=strtolower($inputCharset);
$targetCharset=strtolower($targetCharset);
$fallbackCharset=strtolower($fallbackCharset);
$decodedStr='';
$mimeStrs=imap_mime_header_decode($mimeStr);
for ($n=sizeOf($mimeStrs), $i=0; $i<$n; $i++) {
$mimeStr=$mimeStrs[$i];
$mimeStr->charset=strtolower($mimeStr->charset);
if (($mimeStr == 'default' && $inputCharset == $targetCharset)
|| $mimStr->charset == $targetCharset) {
$decodedStr.=$mimStr->text;
} else {
$decodedStr.=mb_convert_encoding(
$mimeStr->text, $targetCharset,
(in_array($mimeStr->charset, $encodings) ?
$mimeStr->charset : $fallbackCharset)
)
);
}
} return $decodedStr;
}
?>
Hope it helps.