Converting multilingual text from binary to text utf8

[Michael Goller]

Hi - I have an odd problem that I am stymied with. My php/mysql system stores multilingual text in a BLOB field. I want to add Full Text search for the system which means I have to get the data out of the blob and into a TEXT field. Problem is, no matter what I do, once the data is in the TEXT, when I retrieve the data and display it in the browser it shows up as question marks instead of characters.

I have tried the following without success:
(1) created a text field in the table and used an update statement to copy the data to it
(2) created a varchar field in the table and used an update statement to copy the data to it
(3) repeated steps (1) and (2) using the CONVERT function in the SQL

A few other notes:
-I output the data from the BLOB to a csv file, then read it and display it from PHP and the characters look fine (chinese, japanese, hebrew etc.)
- I output data to csv, the load it back into the script and update the table with it and still get question marks.
- The header of the php page is set for utf8
- The form through which data is entered in the system is set to accept utf8 charset
-Data saved in the blob comes out without any problem; it's only data that's converted that's a problem

The reason the data is in a BLOB is that there could be a HUGE number of characters stored in that field, and I didn't know about the TEXT data type at the time. Now I need to move the data to TEXT and have 3000 rows of multilingual data already.

Question:

What is the proper way to convert or move my multilingual data from the BLOB to a TEXT field such that the data is preserved properly and accessible as multilingual data in utf8?

-THE BINARY DATA IN BOTH THE BLOB AND TEXT/VARCHAR IS EXACTLY THE SAME AFTER THE UPDATE, so I know the bits are ok. It's got to be some encoding issue that I don't understand.

Any help is appreciated.

Thanks,
Michael