To be really useful for the vidoesecurity applications plugin needs to implement pan (without decoding all the throuwn-away data), not just zoom as we discussed a year ago :-). At least with a step of one macroblock (16x16 pixels with no decimation), and even with JPED DC coefficients are not independent (difference in scan-line order is encoded).
Of course that will require getting inside the Mplayer to add the additionbal processing, but that is really needed for the multiple cameras applications.
lowres parameter is passed to mplayer as -lavdopts lowres=<value>. From man mplayer:
lowres=<number>[,<w>] Decode at lower resolutions. Low resolution decoding is not supported by all codecs, and it will often result in ugly artifacts. This is not a bug, but a side effect of not decoding at full reso- lution. 0: disabled 1: 1/2 resolution 2: 1/4 resolution 3: 1/8 resolution
I look that with mjpeg it works fine. Decoding with 1/8 resolution needs 1/2 of full decoding time. But 1/16 is not implemented in mplayer. --Sergey Khlutchin 12:34, 20 November 2005 (CST)
Yes, I understand that Mplayer does not have windowing feature that would allow faster decoding just a selected window from the full frame that is streamed. That will require to get into Mplayer itself - both pass the additional parameters for the window selection and actually implement windoiw processing. But I don't think it is too difficult and we definitely need it.