Hi. I'm using an IDE (don't tell anyone, but it's Microsoft Visual Studio 6) > so I can watch the qmin and qmax values as they change, and I can see > the size of the output buffer for each setting of the quality and other > VBR values. In the case of the video file that I'm using here, I always > get a buffer size of 11556 bytes, no matter what the global_quality, > qmin, qmax, lmin, lmax, mb_lmin and mb_lmax elements of the codec > context are set to. >
But it is different from the size used when you NOT specify anything else? I mean, the size should be considerably different? > > From this, I'd guess that there's something else that needs to be > changed or set somewhere, or that the encoder has to be closed and > reopened for each quality value change and the data is not treated as > dynamic. So the encoder needs to be closed/opened every time we want another quality level? Btw, if you don't close/open the codec, the PTS needs to be advance by 1 every time as you said, correct? > > > It's almost tempting to tell the customer that they need to use a > hardware JPEG encoder, 'cause it'd be a damned sight cheaper than having > to pay me to do all of these experiments, but a) I did quite lot of > work with early CCube encoders (late 1980s, early 1990s) and they > weren't much fun either, and b) I need the money. > As long as they ready to pay - take it :). Regards. _______________________________________________ libav-user mailing list [email protected] https://lists.mplayerhq.hu/mailman/listinfo/libav-user
