Currently: A file downloaded from the camera: dennis_at_dennis testing $ ./a.out 0xffd8 - Found StartOfImage at 0x0 0xffe1 - Found ApplicationSegment1 (EXIF) at 0x2 0xffdb - Found DefineQuantizationTable at 0x61a9 0xffc4 - Found DefineHuffmanTable at 0x622f 0xffc4 - Found DefineHuffmanTable at 0x6250 0xffc4 - Found DefineHuffmanTable at 0x6271 0xffc4 - Found DefineHuffmanTable at 0x6328 0xffc0 - Found StartOfFrame at 0x63df 0xffda - Found StartOfScan at 0x63f2 0xffd9 - Found EndOfImage at 0x22226 (Notes: The ApplicationSegment1 contains the thumbnail, that is why the real image does not start until 0xffdb at offset 0x61a9. The actual filesize is 0x22800 bytes) A "real" JPEG: dennis_at_dennis testing $ ./a.out 0xffd8 - Found StartOfImage at 0x0 0xffe0 - Found ApplicationSegment0 (JFIF) at 0x2 0xffdb - Found DefineQuantizationTable at 0x14 0xffdb - Found DefineQuantizationTable at 0x59 0xffc0 - Found StartOfFrame at 0x9e 0xffc4 - Found DefineHuffmanTable at 0xb1 0xffc4 - Found DefineHuffmanTable at 0xd2 0xffc4 - Found DefineHuffmanTable at 0x189 0xffc4 - Found DefineHuffmanTable at 0x1aa 0xffda - Found StartOfScan at 0x261 0xffd9 - Found EndOfImage at 0x3b7eb (Notes: no thumbnail, image starts (0xffdb) already at offset 0x59. The actual filesize is 0x3B7ED bytes. In other words, this file ENDS with the EndOfImage marker as expected, while the Exif file has a few hundred bytes extra) See page: http://web.usxchange.net/elmo/jpeg.htm for good stuff about JPEG.Received on Mon Sep 09 2002 - 11:49:35 CEST
This archive was generated by hypermail 2.2.0 : Tue Oct 20 2009 - 17:33:19 CEST