I had a lot of time on my hands yesterday, from getting home from my vacation as well as being sick (I should be in bed right now), so I did a lot of playing around with encoding my videos again. I’ll write more about that later, but one thing that has has proved interesting while finding another possible standard to apply for my library, is how there are a few things I was so adamant about having in the beginning that don’t really matter at all after some practical application.
I can only think of a couple of examples off the top of my head, but it’s funny to see how stuff I just whined and cranked about being a “must have” back in the day, after many hours of actually accessing my library, my assumptions on what I “needed” were wrong.
Perhaps the most important realization I made, though, was that different categories require completely different approaches. When it comes to a multimedia library, you can’t apply the same standard to a movie and to a TV episode. I’m still mulling over in my head if that’s true with live-action and animated shows.
Video metadata: titles and chapters
Way back when, I was adamant about getting chapter information in my metadata, so I could quickly jump to points where I needed to in my videos. This actually makes sense for movies, where there will be something like a dozen or more chapters that jump between obvious scenes, but for TV shows, it doesn’t make so much sense, where there might be four or more on an hour long show, and usually nothing on a half-hour show.
But, setting chapters was probably one of the main reasons for me choosing Matroska in the first place over all the other container formats, simply because it was the only one (aside from OGM) that supported it. OGM was nice, though, but I vaguely recall that there were reasons for not using it. I can’t really think of any right now, but I imagine that since development has stopped, and mkvmerge had more options and supported more codecs and is actively developed, that probably made the choice simple.
The reality is that I never use the chapters. The solution that eventually worked it’s way into the scene was just fixing mplayer-resume to work correctly. Add to that the fact that I can resume playback from a series of a TV show (a playlist of media files), and I’m happy to be able to pick up anytime where I left off, versus flicking back and forth between chapters to find it myself. The problem I worked so hard to avoid never came up. Interesting.
One problem I struggled with a long time was that I wanted to have subtitles, mostly “in case I ever want them.”
I went to a lot of work figuring out how to get VobSub formatted subtitles out of a DVD and into a Matroska file. That, in itself, was rewarding. For one, I needed to learn how to find one possible solution, and this one proved very simple.
The reality, again, though, is that I never use them and in fact they are more annoying than helpful. Only annoying because MPlayer is forcing them to be on, when I don’t want them to, which is a bug, but still.
The “problem” in this case is still there, and I think that as far as long-term solutions and archiving goes, it’s a good idea to keep them, but it’s again telling that I had another must-have feature which never got used.
This is one I’m currently dealing with fixing, since I’m looking at other containers aside from Matroska. Part of me hates to let go of my vehement fanboyism, but a large part of me knows that it’s good to be objective, even if the software isn’t up to speed just yet. Does Matroska have the best support on Linux *right now* to do everything I need to? Absolutely. Should I use it for everything, everywhere, all the time? Not really. I’m going off on a tangent again, though.
I’ve always maintained that the audio quality should remain in the original AC3 (Dolby Digital) format, simply because I might someday (another archival/possibility requirement) want to someday pipe all the audio through my receiver.
Again, the reality completely disagrees with the presumed history. Turning on my receiver, switching my audio devices around, and using the remote to control two different volumes is a real hassle. I never do it. Also, this is another time where the categorical differences come into play. Movies, you want to keep at the highest quality, including sound, because the channels are almost always going to be in surround. Of all the TV shows I have, though, Star Trek are the only ones that are not either stereo or mono.
So, my probable solution is to save a small amount of space and eventually use MP3 or *gasp* AAC in the next storage standard I decide on for TV shows. I’ll admit that the size difference is generally marginal, about 20 MB of space on a 30-minute show … but added up it starts to make some difference. Not big enough that I’d switch my entire library, but if I do end up converting the video mostly because I want to save space, then I might as well do the same for audio if I’m never going to use AC3 for its nice feature. Either way, if I’m switching to MP4, then I have to, no matter what, as the format won’t support Dolby or DTS. Plus, there should be receivers out there that support AAC. I think.
This is the one that got me started thinking about the whole thing, which is actually quite funny considering how much time I spent on the problem, and the incredibly simple solution.
One of my main, longstanding issues with re-encoding my DVDs from MPEG-2 to MPEG-4 has been A/V synchronization. Having the audio off by just a few milliseconds drives me insane, being the observant freak that I am. Which is also one reason I won’t reencode anyway, because I can notice slight video artifacts, and it’s simply more mentally relaxing knowing I didn’t make any changes to the original stream. Ah, the wonders of obsession.
I spent a lot of time playing with encoding options, frame rate settings, etc. to find that perfect mix of configuration harmony where everything would be perfectly in sync so I wouldn’t go insane trying to watch Looney Tunes. I never found that secret combination, but I did learn a lot in the process of what was throwing it off and why. Plus I learned early on that there’s never a magic bullet for re-encoding that will work on every source material.
However, I still had some audio/video sync issues on some TV shows, even with minimal transformation on my end (from DVD straight to MKV, MPEG-2 with AC3). The solution was really simple: I just mapped the Channel Up and Channel Down buttons on my MPlayer configuration to change the A/V sync by 100 milliseconds in either direction. That way, when I can visibly detect the difference, I’ll just tap a button on my remote and all is happiness. Much simpler.
One last problem, or set of problems, I ran into, was with mythvideo itself. I’ve documented those pretty extensively over the course of my blogging. Everything I ran into, though, my proposed solution was to pretty much just sit around and wait for someone to fix it. I never would have imagined years ago that I’d eventually just open an editor, spend a few hours trying to hack on the source code, and actually manage to fix them myself.
Granted, they were all really, really, really small issues … nothing major like adding a TV tuner to the supported hardware list, but the fact that I managed to hack it out still impresses me, no matter how tiny the changes. That’s just cool.
Interesting stuff, I think. This is one of those posts that I look over and think, “my heck, everyone is going to think I’m this maniac nutball who obsesses night and day over a stupid collection of cartoons on a computer” and I start to wonder if I should hit that publish button. The reality, though, is that any subject that I get really fascinated, interested and involved in, I look at it from a whole lot of different levels. And if finding that perfect solution is just a matter of time, research and dedication, then I’m definitely up for the task. It has, of course, taken me a matter of years to learn all that I have, so consider this the side notes to everything I’ve figured out so far. Fun stuff.