I worked on a project with a MS SQL server 2008 containing data of NVARCHAR type in multiple languages,
including asian characters. It is a known issue, that the PHP MSSQL functions are not able to retrieve
unicode data form NVARCHAR or NTEXT data fields.
I spent some time searching for possible solutions and finaly found a work arround, that provides correct
display of latin and asian fonts from a NVARCHAR field.
Do a SQL query, while you convert the NVARCHAR data first to VARBINARY and then to VARCHAR
CONVERT(VARCHAR(MAX),CONVERT(VARBINARY(MAX),nvarchar_col)) AS x
While you fetch the result set in PHP, use the iconv() function to convert the data to unicode
<?php $x = iconv("UCS-2LE","UTF-8",$row['x']); ?>
Now you can ouput the text to UTF-8 encoded page with the correct characters.
This workarround did run on IIS 6.0 with PHP 5.2.6 running as FastCGI.