Welcome Guest ( Log In | Register )


Important

The forums will be closing permanently the weekend of March 15th. Please see the notice in the announcements forum for details.

 
New Feature (suggestion)
« Next Oldest | Next Newest » Track this topic | Email this topic | Print this topic
jonny
Posted: Jan 7 2003, 02:12 PM


Unregistered









Hi


ATM programs like GKnot or my DivX5Enc are able to predict final size (and make a compressibility test) thanks to the log file generated by the codec in the first pass.
This have 3 limitations:
1) If the output format of the log file change i must change my code to reflect this.
2) In this way is only possible to predict final size for 100% quality based encode.
3) The format of the log file is different for every codec.



I was playing with the source and i've added a command line option to VDub: "/n"
This option is very simple, if you specify it, 2 things happen:

1) For every encode a log file is written.
This log file contain info about the encoded frames (basically a bool that indicate keyframes and the size in bytes of every encoded frame)
Something like this:

1 74393
0 4344
0 19923
0 5454
0 3122
...


2) Frames are actually encoded but not written to the final Avi.


Thanks to this log file, external programs will be able to predict final size for every quality based encode & 1-pass encodes with fixed quantizer (using AviSynth scripts with the SelectRangeEvery or SelectEvery).
But more, the prediction will be indipendent by the codec used.


The modifications needed in the source are very little, the advantages for programmers are really big, we can even predict size for credit parts encoded at fixed quant.


I'm asking if it would be possible to add this option in the official VDub (i think it's not nice to release my mod version adding only this option).

Best regards
jonny
 
  Top
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:
0 replies since Jan 7 2003, 02:12 PM Track this topic | Email this topic | Print this topic

<< Back to VirtualDub Development Forum