On Mon, 2010-05-31 at 09:19 +0200, Jean-Francois Moine wrote: > On Mon, 31 May 2010 00:03:10 +0200 > Ondrej Zary <linux@xxxxxxxxxxxxxxxxxxxx> wrote: > > > > I would try extracting a JPEG header from one of the files captured > > > by the camera in stand alone mode (either a JPEG still or MJPEG > > > file), and put that header together with the image data from the > > > USB capture. It may not look perfect, but hopefully you will get > > > something you recognize. > > > > Just thought about the same thing so I uploaded a video file: > > http://www.rainbow-software.org/linux_files/spca1528/sunp0003.avi > > > > > Attached was Theodore's first attempt of such a procedure with a > > > header extracted from a standalone image file from my Jeilin based > > > camera and USB snoop data from the same camera. It wasn't perfect, > > > but it was recognizable. > > I could not believe it! I already tried the image as JPEG, but I got > just big colored pixels. I changed the 'samples Y' from 21 to 22 and > I got something coherent! Here is the same image as yesterday with > JPEG 411 header, compression quality 80% and insertion of 0x00 after > 0xff. Very nice work! I think I understand the 'samples Y' change. According to ITU T.81, you changed the Vertical sampling factor in the Y component. So, I guess Luma is undersampled vertically for that mode of the camera (the camera only really has a 240 line sensor)? Regards, Andy -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
