meguIV: The Official Akiba-Online DVD Encoder (v1.0.1.1)

[youmeus]

Hey Vit. When using the frame synthesize option along with progressive post-processing option in MeguVit 1.0 beta3 on a progressive source such as Xvid AVI file, I see heavy blur, almost distortion (a few pixels wide at least), near edges in interpolated frames when there is motion, far more than I did in the MeguVit versions prior to 1.0. In the earlier versions, manually using InterpolateFPSx2 BEFORE everything else in the script kept the edges of the interpolated frames nice and sharp. I would assume that it's because you then interpolate first and THEN repair (if you have a repair script after the interpolation script) ALL frames including the interpolated ones am I right?
Does the new MeguVit v1.0 beta3 only repair original frames? Does it interpolate after repairing or vice versa? or shouldn't it matter?
Am I right on this issue or am I misunderstanding something?
I am hoping you can shed some light on this so that I can understand why I am getting more blur near edges in interpolated frames than in earlier versions of MeguVit.


Anyway I will try to manually edit the script in MeguVit v1.0 beta3 and put the Interpolation script at the very beginning, although I have played around much with this version yet.

 

[Vitreous]

It's probably much simpler than that.

At the same time as working on MeguIVit I have been developing the next QuickTGMC. I have been experimenting with different techniques for fps interpolation because I am going to put that functionality into QuickTGMC. The method in MeguIVit 1.0 beta3 was an experimental version that didn't quite work. InterpolateFPSx2 is better. So just try replacing the QFPSx2() part of the script with InterpolateFPSx2().

I've not experimented too much with putting interpolation first. Can't do that when working with interlaced material, which is why it comes later in MeguIVit. But for progressive material + repair.... It's something for me to look at...

 

[youmeus]

It's probably much simpler than that.

At the same time as working on MeguIVit I have been developing the next QuickTGMC. I have been experimenting with different techniques for fps interpolation because I am going to put that functionality into QuickTGMC. The method in MeguIVit 1.0 beta3 was an experimental version that didn't quite work. InterpolateFPSx2 is better. So just try replacing the QFPSx2() part of the script with InterpolateFPSx2().

I've not experimented too much with putting interpolation first. Can't do that when working with interlaced material, which is why it comes later in MeguIVit. But for progressive material + repair.... It's something for me to look at...

Thanks for the quick reply mate.

I followed your advice and put in InterpolateFPSx2 instead and that certainly did a much better job at interpolating

 

[youmeus]

A completely different thing, but is it possible to just do a video encode to HuffYUV without the encoder proceeding to do an x264 encode and mux afterwards?
As I am doing it now, I write a script loading for example a WMV file that I want to just process to a lossless HuffYUV AVI file, so I load the script and check the pre-rendering option then click "Enqueue". The encoder pre-renders but then proceeds to x264 encode and mux after the pre-render. Can you skip the x264 encode and mux making the encoder stop after the pre-render? It's a detail I know...

 

[Vitreous]

Yes, you can do that. After the main first pass video encode starts, go to the MeGUI main window and hit the "Queue" tab. Scroll to the end. You can see the tasks yet to be processed are shown with a mode of "Waiting". Select these tasks and press "Delete".

Out of curiosity, what do you want just lossless for...?

 

[youmeus]

Yes, you can do that. After the main first pass video encode starts, go to the MeGUI main window and hit the "Queue" tab. Scroll to the end. You can see the tasks yet to be processed are shown with a mode of "Waiting". Select these tasks and press "Delete".

Out of curiosity, what do you want just lossless for...?

Excellent!

Well as you may remember from earlier posts I am on Mac OS X, except for when I encode. I playback and organize all my videos in iTunes. Sadly iTunes and QuickTime is doing a really poor job at decoding the x264 content generated by Megui so I need an additional encoding step to optimize the Megui content for iTunes, for smooth playback, fast forward/rewind and skimming and so on. So I run the lossless AVI's generated by Megui through a Mac OS X commercial encoder which does degrade the image quality some, but nothing too noticeable. This particular commercial encoder application creates mov files that decode ultra smoothly in iTunes. A really shitty thing about it though is that it changes the gamma slightly but what the hell...
It probably seems like madness to you? hehe

 

[youmeus]

When using AssumeFPS how do you actually use the audio_sync=true argument? In my script:

LoadPlugin("C:\meguIV\MeGUI\tools\ffms\ffms2.dll")
FFVideoSource("C:\Users\Public\Videos\Source\nextweb-con.wmv").AssumeFPS(30, true)

I have the argument set to true, and load it up in MeguVit. I set the audio source but then what? Do you need to point to the audio in the script all together instead of setting it in Audio Input?

 

[Vitreous]

I'm not sure what exactly what you mean. Maybe you're trying to do this:

...
v = FFVideoSource( videofilename )
a = FFAudioSource( audiofilename )
AudioDub( v, a )
...

Can use whatever source filters you want in that script, and the audio source can be a video that contains audio provided the source filter will accept the file type.

 

[youmeus]

I'm not sure what exactly what you mean. Maybe you're trying to do this:

...
v = FFVideoSource( videofilename )
a = FFAudioSource( audiofilename )
AudioDub( v, a )
...

Can use whatever source filters you want in that script, and the audio source can be a video that contains audio provided the source filter will accept the file type.

Thanks. It was exactly what I was trying to do.

Was trying to speed up a clip from 29.97fps to 30fps while keeping the audio in sync. I suppose should look like this then:

LoadPlugin("C:\meguIV\MeGUI\tools\ffms\ffms2.dll")
V = FFVideoSource("C:\Users\Public\Videos\Source\nextweb-con.wmv").AssumeFPS(30, true)
A = FFAudioSource("C:\Users\Public\Videos\Source\nextweb-con.wmv")
AudioDub ( V, A )

 

[Inca]

Having a problem with what I believe may be a bad ISO.

After decrypting, I did a oneclick encode with Vit's 1.0 beta 3. When i came home it had finished but the last job to mux the video and audio had errored out. I noticed another job earlier had errored out which i think was the demux. So i only had the video and no audio.

No biggie. All I need to do is demux separately and then mux, right? Well for some reason it keeps erroring out when I do. I tried doing each VOB file individually and it works until I get to the third VOB, VTS_01_3.VOB. Then it errors out again.

Here's the log:

The error "seeking is severely broken" is telling me that VOB file is the problem. Furthermore, when i play VTS_01_3.VOB mpc tells me its 26 hours long! I tried decrypting that file again and get the same problem.

Is there any other way I can get the audio?

 

[Vitreous]

Have you tried merging the VOBs into one big VOB then extracting the audio from that? VOB parts aren't intended to be worked on individually.

 

[Inca]

Have you tried merging the VOBs into one big VOB then extracting the audio from that? VOB parts aren't intended to be worked on individually.

I have. In fact I have to because I'm still having that indexing problem with every ISO I try to encode. It will only encode the first VOB. So I merge them all with Vobmerge. When I play the merged vob file it plays the audio the whole way through but still errors out when I try to extract. I also tried using adobe audition to extract the audio from video. It doesn't extract anything either.

This is such a great Mayumi video too. It's got the scene where she cleans the house wearing hot pink shorts :exhausted:

Any other options?

 

[Rollyco]

Your DVD probably has broken time maps. Process your titles with Vobblanker, using the default settings plus "Fix always VOB SYNCI/SRI pointers in Titles" checkmarked. For more convenience, I also suggest checkmarking "Use Input Folder." A backup folder will be created with the pre-processed files, in case you need to revert.

 

[Inca]

^Same errors after processing with Vobblanker. :notagain:


Thanks for you guys' help. I've spent all day with this and think I'm just gonna give up!

 

[Vitreous]

I've never seen a *.d2v.bad file, or a *.fix.txt file before.

This is fine, it means that a problem was found and fixed during indexing of the VOB files. Probably a DVD authored out of spec or something? Whatever it is, it's not too uncommon and I've never had any problems when it happens.
____

Inca: There's something screwy with your setup if you still can't index files properly. Some ideas:
- Rename the merged VOB to a .mpg and use DirectShowSource as a source filter, that may get you the audio filters that are working correctly on playback
- Try indexing with DGIndex directly (not via MeGUI)
- Try indexing or processing that audio with a fully updated MeGUI (using the dev servers to update) instead of MeguIVit

 

[youmeus]

The output quality of the x264 encode in an MKV container will be the same regardless of pre-render or not right?

 

[Vitreous]

The output quality of the x264 encode in an MKV container will be the same regardless of pre-render or not right?

You should get an identical output in fact (unless you use the non-deterministic setting of x264, which nobody does). Use pre-render only if it's faster or suits your workflow for some reason.

 

[youmeus]

You should get an identical output in fact (unless you use the non-deterministic setting of x264, which nobody does). Use pre-render only if it's faster or suits your workflow for some reason.

Excellent. I don't have a lot of room so this way it'll actually be realistic for me to rip my blu-rays

 

[Vitreous]

...blu-rays

Ah, HD. You may run out of memory without pre-render - forget your system RAM, Avisynth is a 32-bit app and can only use 2Gb in normal circumstances.
You will almost certainly need to cut down the number of threads in the "Custom Processing" tab. Maybe main threads 4 and sub-threads 1.

 

[youmeus]

...You may run out of memory without pre-render.

Oh, maybe I should consider getting a scratch disc then - because I take it that reducing threads will slow the process considerably. I already have some speed issues as I don't think I can go much faster than 1-1.5fps pre-rendering 1080i content :-( on a 2.2Ghz i7 quad core.
A pre-render of 60mins of 1080i content to 60p would eat up at least 150gb wouldn't it?