Thanks for your previous reply regarding mjpeg.
I am looking forward to your mjpeg sample.
But anyway: I am still confused about the basic principle. I try to decode a continuous flow of jpeg images. Each of these images is a complete jpeg (it includes all jpeg table definitions), but the DQT (Quant. tables) and DHT (Huffman tables) segments of the jpegs are always the same. Your ijl manual says: "Copying the persisted tables to JPEG_PROPERTIES is typically much faster than appending a table stream to the front of each JPEG data stream and forcing the decoder to process and format the tables at every call." So I tried to save and resuse the formatted tables as described in your flashpix sample documentation. But since every jpeg already includes the jpeg tables, how can I tell the decoder not to use them? I also tried to exclude the tables from the jpeg image and to supply the persisted version of the formatted tables, but this does not work for me (I get a one color bitmap as the result). What's wrong? How should I decode this kind of jpeg stream? Thanks
Alex
