Issue with XviD_Encraw and encoding to achieve a target size

[onesikgypo]

Hi,

Ive recently began using xvid_encraw as a command line alternative to gknot for encoding my mpeg2 to avi - i have done so with quite good success, however issues arise in attempting to get it to achieve a target size.

Though i use the -size flag int he second pass encode, it is generally not accruate as gordian know usually does. I then decided to compare the mediainfo for the video output(no audio) of the xvid_encraw and the gordian knot one - and apart from the bitrate which si different, the only difference i could see which would affect that was listed under video as "Stream size" and "Bits/(Pixel*frame)"

On the XviD_Encraw it was:

Bits/(Pixel*Frame)               : 0.173
Stream size                      : 16.3 MiB (99%)

Whilst the one done by goridan knot via vdubmod was:

Bits/(Pixel*Frame)               : 0.188
Stream size                      : 17.7 MiB (99%)

I used the same AVS for both encodes.

I anyone could provide any assistance, it would be greatly appreciated.

Thanks.

[midders]

Though i use the -size flag int he second pass encode...

Just a stab in the dark, but I was always told that you should use exactly the same options for the first and second passes, apart from the -pass 1/2 and the name of the output file; have you tried putting the -size flag in the first pass as well?

Slainte

midders

[jagabo]

Stream size, and bits per pixel per frame are simply derivatives of the bitrate:

bits per pixel per frame = bitrate / fps / width / height
stream size = bitrate * running time

So obviously, different bitrates will result in difference in those measurements.

Compare things like number of consecutive B frames, min and max quantizers, GOP size, Motion Search Precision, Adaptive Quantization, etc. Those settings can cause and xvid encoder to miss the requested bitrate.