Hi,
I almost have ov519 cameras working in PDP now, but sucking really really badly at C (P-languages spoiled me) I don't understand several things yet and so I desperatly need some help/explanations. I'm soooo close to have the cam working, it's just a little bit of pushing in the right directions needed, I guess.
I even have difficulties finding the right quesitons to ask.
Well, for one: I don't really understand the role of the "newimage" pointer in pdp_v4l_bang(): It is set two times, once as:
/* get the address of the "other" frame */ newimage = x->x_videobuf + x->x_vmbuf.offsets[x->x_last_frame];
then later is is set again without ever being used inbetween:
newimage = x->x_videobuf + x->x_vmbuf.offsets[x->x_frame ^ 0x1];
However if I comment out the first assignement, PDP doesn't work right anymore, so I suppose, it is used somehow in between. Could someone explain to me, where, how, and why?
Second Question:
To decompress video data delivered in JPEG, am I right in guessing, that I have to change the content of newimage's data after it is set the second time in pdp_v4l_bang? (BTW: I currently get the same results visually, whereever I change newimage.)
At least I found out, that newimage carries the compressed data. Sending "newimage" into the jpeg decompression as input to the "source manager" of jpeglib at least gives "something". However I think, I don't really get where to put and where to get the image data, probably because I don't really understand the double buffering correctly, ah, no.
Well, the latest code gives a full picture, mostly even without JPEG decompression errors, however the images is jumping left and right and also is in "vertical splitscreen" mode, that is, it is split in half and two half-pictures are laid out next to each other. The frontier is moving. I would like to do a screenshot, but I cannot get "import" or any other tool to grab the framebuffer-screen.
Ciao