New FreeEnc 0.44

[manolito]

Anyone tried this new version yet? Lots of bug fixes, but at least for 1 pass VBR the new rate control does not work.

Just did a 1 pass VBR encode using the "Incredible" template (only changed the matrix from BVCD to Bach1). Requested average bitrate was 5.269, max BR 8.000, min BR 300.

The size of the resulting encode was just 2.82 GB (instead of 4.7 GB). BitrateViewer reported peak BR as 8.682, average was 3.280.

I think for now I am going back to version 0.31. My 1 pass VBR encodes with this version always came out at exactly the right size, and the bitrate peaks were not much worse than with the latest version.

Cheers
manolito

[guada 2]

hank you manolito, :)

I didn't test it yet.

Some news coming from the site of Wmesquita:
- Fixed -1 switch and added -3,-4 and -5 for more passes.
- Fixed MinBitrate bug (it should fix "Could not open avcodec MPEG2" in some cases...)
- 1 Pass working again.

Downloading:
http://www.vmesquita.com/en/download.htm

To soon.

[Inc]

Do look at my Template if the MIN Quantizer is set to 1 cause as default libavcodec based mpeg2 encoders do use 2. And thats the Point. At such a high avg Bitrate and therefore Q Level the Encoder gets to a min Limit of 2 where your encoding would like to even go under 2 which causes less Quantisation and therefore larger Filesizes.
BTW. In CQ or Q Encoding Mode the AVG Bitrate is not the thing to count with as you do deal using min/max and a Quantizer Value.

[freelock7]

The size of the resulting encode was just 2.82 GB (instead of 4.7 GB).

True. Prediction is not correct. It is based on the P.Cheat libavcodec he left behind unfortunately. So there is no hope to develop it anymore. Unless...

[video_magic]

Thankyou for this, I look forward to trying it, I appreciate your work on this and making it available :)

I had been using Peter Cheats' modified Mplex 2.0 and it seemed to be better than the normal one, so this uses his modified libavcodec or some of the code from it?

Is that guy coming back anyone know? I believe he had lost a computer to some damage :( that was the last I read, but I had noticed he did something with Quenc and rate control which is a shame if that has been aborted, although I know other people have also worked on overcoming that problem but I guess more heads solving a problem the better (don't know if it works like that with coding though :p ).