PDA

View Full Version : Help ripping Ocean's Eleven--Avisynth issue.


Dark Shikari
5th June 2007, 03:46
My script:

Part1=DirectShowSource("I:\VIDEO_TS\VTS_01_1.VOB")
Part2=DirectShowSource("I:\VIDEO_TS\VTS_01_2.VOB")
Part3=DirectShowSource("I:\VIDEO_TS\VTS_01_3.VOB")
Part4=DirectShowSource("I:\VIDEO_TS\VTS_01_4.VOB")
Part5=DirectShowSource("I:\VIDEO_TS\VTS_01_5.VOB")
Part1+Part2+Part3+Part4+Part5
degrainmedian(mode=3).fft3dfilter(bw=16, bh=16, ow=8, oh=8, bt=4, sigma=2.5, sigma2=3, sigma3=3.5, sigma4=3)
ConvertToRGB24()
Crop(3,79,-3,-81)
Lanczos4Resize(976,416)
ConvertToYV12()

It works extremely well, and x264 is encoding the video as I post this. However, the audio is messed up when it comes out of Avisynth. To be exact, the audio from the first VOB file is fine, the second is sped up by a factor of two and then the rest of the time until the next VOB is filled up with silence, and the same for all the rest of the VOBs.

The only other way I've found to do this is to extract the five ac3 files, convert+downsample to wav, and then put them one after another--but this seems to produce jumps in the sound, as if they don't exactly line up perfectly.

Is there any way I can fix this issue within AviSynth?

chipzoller
5th June 2007, 04:16
If I may ask, why are you using directshowsource() to load each VOB file? Why not create a d2v project file which will then demux the AC3 tracks you wish. I had no problem encoding my copy of Ocean's 11 using this standard method.

ChiDragon
5th June 2007, 04:16
Why not use DGMPGDec, and load up all the VOBs together in DGIndex?

Dark Shikari
5th June 2007, 04:22
Why not use DGMPGDec, and load up all the VOBs together in DGIndex?

I guess I'm new to the world of DVD ripping and as a result tried to apply my tried-and-true video encoding methods to a situation where there exist better programs :p

I'll try that. Will my video still match up with the audio produced by this project? I would rather not run the incredibly time consuming encode again if possible.

chipzoller
5th June 2007, 04:27
I wouldn't count on it. From now on, use DGIndex to make your d2v file and use MPEG2Source() to call the d2v file...shouldn't have any problems.

Dark Shikari
5th June 2007, 05:30
I wouldn't count on it. From now on, use DGIndex to make your d2v file and use MPEG2Source() to call the d2v file...shouldn't have any problems.
Well I restarted the encode and dropped the number of reference frames to 4 to save time this time around.

The denoiser filter seems to be a huge CPU hog but I guess that's expected.

The output quality is amazing; its far, far better than the original MPEG-2 source IMO.

foxyshadis
5th June 2007, 07:55
ConvertToRGB24()
Crop(3,79,-3,-81)
Lanczos4Resize(976,416)
ConvertToYV12()

????

Lanczos4Resize(976,416,3,79,-3,-81)

Saves you a color conversion. fft3d is a killer on a slower system, but you can always remove it for the first pass to save some time (but raise degrainmedian's strength), it won't affect the first pass much. Consider adding deblocking (cpu=4) in mpeg2source, as well.

Dark Shikari
5th June 2007, 12:10
ConvertToRGB24()
Crop(3,79,-3,-81)
Lanczos4Resize(976,416)
ConvertToYV12()

????

Lanczos4Resize(976,416,3,79,-3,-81)

Saves you a color conversion. fft3d is a killer on a slower system, but you can always remove it for the first pass to save some time (but raise degrainmedian's strength), it won't affect the first pass much. Consider adding deblocking (cpu=4) in mpeg2source, as well.
I'm using crf mode anyways so no first pass.

And its hardly a slow system--its a 2Ghz Core 2 Duo. But I'm used to slow encodes.