Hi, I'm sure that I'm going to regret getting into this, but... I'm capturing video for later playback to interlaced PAL-refresh rate output. My capture rate is 50Hz fields, 25Hz frames. My playback rate is the same. I'm using V4L2. Because my input and output formats are interlaced, for best quality I believe it's important for my system to keep the fields straight through capture, into the stored video files and into playback. But - is this right? Perhaps it would be OK to just gather the fields into frames at capture time (wihout any deinterlacing filters) and then on playback just make sure that each frame is put to the frame buffer at the vsync at the start of an top/odd field so that the top and bottom fields are displayed in sequence as the video card paints each field. So the field-by-field resulting stream looks the same even though the stored video file doesn't track the fields specifically. But if I do that then the interlaced frames will not compress so well (in mpeg, anyway) due to the combing effect in the assembled frame... And they won't look so go on a non-interlaced display. So - is there any advice out there for me? As far as I know, only mpeg2 format specifically handles interlaced video? Right now I use mp1e for capture - and that's MPEG1 only. My playback is mplayer, and that can't currently play "true" interlaced mpeg2. Thanks for advice or pointers, Steve