Slashdot Mirror

Can you share what tools / scripts / arguments you use for this?

I've not found any way to do it without a massive headache.

I use Windows and AnyDVD, which removes all protections from both DVD and BluRay, but if you only use DVD, pretty much anything will work to get the content into the right format. I would recommend something like DVD Shrink to quickly grab only the main movie and the soundtracks and subtitles that interest you.

After that, I prefer semi-manual means instead of an all-in-one system, so I use DG Index to feed into AVISynth. You don't need anything fancy in AVISynth, but you will want to use something like SmartDecimate to de-interlace the movie back to the original 24fps. Apply other filters as you see fit.

Then, to convert to H.264, something like x264 --preset slower --tune film --crf 18 --output "MovieName.264" "MovieName.avs" will do the job fine, giving you filesizes averaging about 700MB/hour. Eventually, you can learn more about x264 and do some fancier things, but all they will buy you is a smaller filesize for the same visual quality.

Then, rename the MovieName.264 file to MovieName.h264 and use mkvmerge or the GUI front-end to combine the original audio and subtitles from the DVD with your newly encoded video. This will get you movies you can watch. If you want to add things like chapters, or modify the original audio, you'll have to use a few more tools, but those are completely optional steps.

Thanks for the positive comment; I guess all the time I've spent idling on various IRC channels (which, with the exception of #mplayer and #mplayerdev, are mostly anime-related) has to count for something.

I'm involved with a couple of fansub groups, as well, and that's where most of the knowledge of subtitle formats comes from; most groups are moving towards soft-sub (ASS) MKVs with x264 video for their primary releases, and the quality of stuff out there is very good, with subtitles that are generally better than the licensed DVDs, particularly in terms of subtitles (you can't do much with DVD subs, unfortunately).

It also doesn't hurt that I'm geeky enough to hack at subtitles with text editors (instead of tools like Aegisub. IMO, it's a sign of how awesome vim is that there's SSA syntax highlighting for use with it. Plus, with tools like mkvtoolnix, it's easy to extract tracks and remux MKVs, which is great if you're obsessive enough to want to correct subtitles, or to merge a video file and external subs.

If nothing else, hopefully a few people picked up something worthwhile from the post; one of my friends saw it and was thanking me for the MPlayer config; I guess he didn't realize that he could set MPlayer to auto-play specific tracks. That has always been one of my favorite features.

Still, despite all of the VLC-bashing in this story, I have a lot of positive things to say about it, it definitely has some value as a player, even if I seldom use it. Mostly, I recommend it to less technical friends who aren't going to use it to watch soft-subbed video, since it can generally play everything and the kitchen sink, without having to monkey with codecs. And the streaming features are nice, should you need that sort of thing.

Thanks again.

You'll want MkvToolNix. It's gui's called mmg (and is part of the package).

Create an avi with video and audio in dvd::rip, have it extract the vobsub file. Then use mmg to merge the avi and the vobsub file (mmg can split the files too, so don't do it in dvd::rip).

If you're ready to invest more time you might look at how to create a x264 video (if you got the horsepower for playback that is) because it's much better than even XviD.

Ffdshow is an easy replacement, as it's based on a Linux project, ffmpeg.

The extent of my encoding comes from recording TV or VHS tapes and throwing a few filters at them for cleanup and encoding. So I've never had a chance to become familiar with what features avisynth has. Some people over at Doom9's Linux forum have mentioned getting Avisynth running in wine.

Tmpgenc should be a pretty easy replacement. Transcode seems to be a favorite for mpeg2 encoding. There was a problem when I first started using Linux, which slowed mpeg encoding to a snails pace for a while, and I've never found a chance to get back to see how it works when fully functional. A lot of peoplel swear by it though, and it's feature list is pretty impressive. Avidemux, as well as quite a few other programs also provide mpeg1 encoding, I think through ffmpeg.

Everything in VirtualDub/VirtualDubMod except video capture should work fine under wine and that's what I used for quite a while before Avidemux appeared. The design and functionality are very similar to virtualdub, and it also includes encoding to mpeg1.

One of the two best media players for Linux, Mplayer, should be noted to have better support for matroska, in cvs, than is available for the format under windows. And these tools can be used for creation and editing of matroska files.

For capture from a tv card, I use nuppelvideo, and then use avidemux to edit and convert to a smaller format.

I did that (convert it to nicer format that is). Ogg media stream container, mpeg4 video using mencoder+libavcodec (2pass, high quality, 4 motion vectors per macroblock, bitrate 800 kbit), Ogg Vorbis audio (quality 1). Result is a 58MB big file. Obviously there is quality loss (the original was 141MB), and mencoder seems not to encode all frames, making it a bit skippy at times (this also probably explains the small size).

It is good quality IMHO (good enough for me to dump the original), and the audio is in sync (the original file has two audio streams, but mplayer ignores the shorter one, for this encoding I decoded both streams to raw pcm, concatenated the streams and encoded the one stream). If you don't like it, don't use it.

Oh, and I had to split it in two since I don't have enough webspace in one spot for 58 megs. Either play them consecutively or join them with ogmcat(1) found in the ogmtools .

File 1, 39MB
File 2, 19MB

Pardon me, but I always thought the caption was stored as image files and then superomposed onto the picture. It seems like that here, where the process of OCR'ing and re-encoding the subtitles is described. This method is also described by the person who made the pictures available. But I still do not believe that someone can hear a phrase from this movie and translate it into "get your pussy-face to my ass" or whatever. It might be possible if it was translated from Mandarin (Or any other dialect) into English, but wouldn't it be easier to just listen to teh English track? I know there's English-speaking versions of this DVD out there.

In some cases, that might be a good thing if development of the "main" branch is languishing...

I'm eager to see Ogg Theora ready, but I've been kind of disappointed at how "quiet" the developers are. (As of this moment, my CVS grab of Alpha 1 from September is STILL showing no updates since then - I just moments ago did a cvs up ...). The developers rarely reply to "outsiders" on the mailing lists or even post updates to it. Offers of help are at the moment being ignored (again, at least on the public mailing list, which until very recently still had more traffic from "how do I use VP3 with Microsoft(tm) DirectShow(tm)" than from discussion of Ogg Theora development...

We already have "ogm tools", the "blame" for which I put squarely on the desire to have Ogg video (with vorbis audio) and lack of (or just slow, quiet, and/or "offline") development on it from the official developers (xiph.org).

Even this slashdot "article" (at the moment, at least) has this sort of "secretive" feel to it - none of the links on it seem to have anything indicating a release today (or, indeed, any information more recent than September 25th), there have been no announcements (or even much in the way of comments) on the mailing lists, and CVS doesn't appear to be any more recent than September 25th. Personally, I suspect somebody has just noticed Theora Alpha 1 - released two months ago - which was billed as the "first Milestone"?. I DO recall seeing a "press release" for this release wherein Emmett Plant (Xiph.org ceo) mentioned that they had "big plans" for a December 1st release of the next alpha, but requests for comments or hints on what those "big plans" may be on the Theora mailing lists get only silence.

I fairness, I imagine a lot of this comes from the development being so "early" and the core developer(s) not wanting a million people popping in and immediately "wiggling" the path of development around, but some comments on the mailing list from the developers about the current development plans would go a long way to making the development seem open...

Does ANYBODY have any more detailed information about what's going on out there at xiph.org/Ogg Theora? Will there be the mysterious unnannounced appearance of a "theora2" branch in CVS instead of an update of the existing code (much as theora appeared without updates to the 'vp32' branch back in September)?

Incidentally, I DID grab Ogg Theora Alpha 1 when it was released 2 months ago. I ran into a LITTLE difficulty getting it compiled, but in the end got it going...and it kicks butt. Even unoptimized, the performance was "pretty good" (certainly comparable with, say, mjpegtools' mpeg2 encoders) and the video quality looks good - I'd even say the quality is pretty good at the lowest "quality setting"...

So...please excuse my pessimistic post and mild rant - the code (on those occasions that it gets released) being open but the development work seeming closed and secretive has been eating at me, considering how interested I am in the project and am relatively eager to help. (I am not qualified to do much coding, but testing and documenting I can - indeed, I have (egad, I'm about to slashdot my own server, aren't I?...).

1. Moritz's DVD ripping and transcoding with Linux howto
2. Linux SVCD guide, written in French
3. My own Linux Digital Fansubbing Guide (shameless plug) -- intended for anime fansubbers but perfectly serviceable as a ripping guide if you ignore the stuff about subtitles.

I agree with the xine recommendation, have a look here: here it talks about mplayer, but you can surf down to the middle of the page. What you want is avifile -- grab that.

Have a look at section 4 of the Xine Faq about where to put them, and exactly which files your after.

How about this from our friends at Linux Liberation Front, the win32codecs (half-way-down(win32-codecs-1.0-1plf.i586)) all packed up for 'ya.

Enjoy!