Hi Michael, On Mon, 2003-10-06 at 08:21, Michael Hanke wrote: > I spent much time to understand the inner working. Moreover, I am not very > sure if I am using LVS the way the authors thought.
Well, LVS has a really tight way of using it. It's (unfortunately) rather easy to crash, too. :-(. Coding LVS (i.e. the connectivity between LVS and mjpegtools, thereby also all functionality) is pretty much terrible. I personally hate unix pipes. They're fine for simple data transfer, like in non-realtime asynchronous protocol-based stuff, but video editing really isn't. Recording is pretty much ok, but the editing features in LVS are just plain bad. I've lib'ified parts of lavplay so I should be able to improve parts here, but have never spent time into actually implementing all this in LVS. Bad me. I'm pretty much in favour of "maintaining" LVS in a way where it's stable, but doesn't get any "really cool" features added. It's rather useless, the framework just isn't good enough. GStreamer allows some of this, but not everything. It'll take quite a while before we've got a really cool editor under Linux like I want it. ;). > Where can I find some information about the scene detection algorithm used in > lavplay? That's actually lav2yuv. The scene detection algorithm is also in lav2yuv.c (mjpegtools). It's simply a "treshold-check for brightness-change" and puts a new scene mark if the change in luminance surpasses the treshold. Could be more elegant, but basically works in some/many cases. :-). Ronald -- Ronald Bultje <[EMAIL PROTECTED]> Linux Video/Multimedia developer ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Mjpeg-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mjpeg-users
