That's clever.
On the other hand, if you take not wobbling, but instead a 30 Hz strobing or blinking (after taking damage, for instance; many, if not most, games blink the character sprite at 30 Hz), this workaround will combine "sprite shown" and "sprite not shown" states in a ghostly translucent form, which will not look as bad as completely discarding every other frame, but still nothing close to how it's actually supposed to look. A workaround for this that is used on SDA's low quality encodes, for instance, will be decreasing the frame rate to 1/3 (20 fps), therefore successfully showing the alternate sprite states, but I bet that's going to make it look all the more jittery.
Any ideas on how to deal with that?
I think doing that manually would be more beneficial in the end since lower frame rate requires less bits to preserve image quality, which in its turn means the video will look better and will be less damaged by YT reencoding. It won't matter for short videos, but for longer ones (say, over 5 minutes), I believe the difference will be easily noticeable.
Well, MEncoder comes with a manual file, which explains most encoding parameters and their use for the included codecs, like subpel refinement quality, reference frames, motion estimation type and range, etc. You don't have to understand how any particular feature or parameter works, what you need to understand is:
1) when to use it,
2) what drawbacks it will have, and
3) what is the syntax to insert in the commandline.
That's pretty much it.
I found that H.264 was very slightly better (nothing new or surprising here, though — H.264 has been a superior codec from the start), and WMV was very slightly worse. The difference was pretty insignificant, though. In fact, a better way to increase quality for YouTube videos is to maximize the encoding quality parameters (not bitrate!) and adjust the filesize so that it's barely below the limit (which is 100 MB). Since YouTube will reencode it anyway with a lower bitrate, bandwidth is not much of a concern.
Just so that people who aren't very good with changelogs know, this version includes reset recording, input display bug fix used for watching the published Super Metroid 100% run, and sound channels toggle patch by Upthorn, among numerous other fixes and additions.
In other words, this is the single universal Snes9x 1.43 version to use for all purposes any of the previous versions were created.
It doesn't make a lot of sense in the context of the discussion; GBx games (and a lot of GBA games, too), especially non-platformers, don't usually track subpixel movement, so "how else do you explain" is… not really a way to explain. If you don't have a solid evidence (which would be one of the memory addresses mentioned above) to base your claim on, how can you be sure it is the cause of your problem? That way I could also blame you for the lack of proper heating in my room, and following your logic you will be the one required to vindicate yourself.
I thought it was common sense to attribute the cause of a problem to some factor only after you've successfully confirmed it. So I take it you have the addresses for subpixel coordinates and/or speed then, which let you know it actually was subpixels, and not something else, that made you lose frames? I'm fine with being proven wrong.
Well, that was slightly better. Two questions:
1. While grenading the tank in the beginning, you run up closer to it from time to time. Is this done because you can't have more than one grenade on screen at any time? Otherwise you'd likely save some time by spamming it from a spot that doesn't require you to leave it until the tank is destroyed.
2. In the hallway, you use the rest of your grenades in a place with little terminator density, and jump over them when it gets more dense. Any specific reason for that?
Nah, it was his regular performance. However, a good point is that he sings suspiciously flawless without any visible strain usually noticeable with even very good singers, and there wasn't a single record of his voice made in an unprepared environment, which led a lot of us Russians at that time to believe the unusual extension of his voice register was made using artificial means. It's been so much time, though, that no-one really cares about him anymore.
Haha, good one.
The game is called "Castlevania", which doesn't make much sense to begin with. What does the word even mean? Like, the "castle" part obviously means a castle and "vania" is a… typical russian name, isn't it?
It's about a family that raises a person trained specifically to deal with frighteningly regular incarnations of Dracula, who appears basically anywhere he wants with a castle chockful of inane demons and other creatures that would appear… well… wrong to any normal person, not to mention that the castle itself is basically yet another living being that doesn't conform to any sane expectations. Something's definitely wrong with that!
Imagine that since the beginning of your life, you're raised in a firm belief that one night in your life, there will be a castle waiting for you with an army of monsters you wouldn't want to see, ever, and you will have to go there and kill the bad guy who is worse than the Death himself. Moreover, Death serves him as one of the guys you need to kill on your way. Kill Death. Wow, that must suck! It's just as if a seal of impending doom has been cast upon your and all your descendants' fate, and you spend your entire life getting prepared for this night that basically means the end of your existence, in the best supposed case.
Would you look brave and happy going there, would you look like a typical hero? No, I think you'd look dreary, at best. Especially since all you have with you is a whip and some equally useless shit. But hey, it's a whip, why not at least have some fun with it and use it as a transporting device? Why not strike down some candleholders? You know, if you strike a candleholder, a heart may come out of it — a big one if you're lucky. Makes sense.
Oh, and in these stables, there's a flying horsehead demon, and under the ceiling, there's a bunch of possessed algae patches. Would a minor thing like that surprise you if you were there? I think not. Add some annoyingly depressive music to it, and it's basically become the most "wrong" game ever.
I mean, good job on picking on a game that comes from a series that was originally composed of all kinds of wrong to begin with, but no, there's still no point in doing that. :)
Also, good run voting yes.
One very important thing to remember when making a speedrun-friendly game is to make multiple routes without making any of them obviously faster than the other. The same is true for character's abilities: several of them should provide decent traveling speed in various conditions, but one of them shouldn't be used for that in absolute majority of the time. Sonic Advance games are a good example of that. The competition and optimization happen only when no ideal path is known, and one thing can be sacrificed to do the other.
Well, that Torizo skip was 2 frames slower than the one I did in NBMB WIP, and in total you lost about 70+ frames up to the diagonal room in Crateria, but nevertheless, it's of good quality.
I must commend you for the early Super Missiles room: you were the third person (after Saturn and myself) to independently discover the no-scroll trick that lets you avoid all the enemies beside the first waver. I was very surprised you didn't take the missile tank there, though, since it's apparently much faster to take than the second missile tank in the treasure room with the r-tank. What was the reason for that? Also, I suggest redoing the room you stopped the smv at, since there's a considerably faster strategy for it (at least 10 frames faster).
Yeah, I think Fano reported that in the m2k2 TAS thread some time ago. It happens if you let go of jump at a certain point, but I have no idea what logic makes Samus touch the door trigger in this case. IIRC, it was even a bit slower due to longer fall in the next room or something like that, though.
I would so give this vote to Baxter for his incredibly hard (and successfully completed) projects, but adelikat takes the edge due to being far most versatile (TASed all consoles with nearly all categories, did both speedruns and playarounds) and most overall productive (20 runs in one year = a lot). JXQ could have picked up the competition if he didn't take a break this long, with his TASing speed he could've cranked out another 2-3 runs, I'd say. :)
Cardboard, for sure. He might have had less of a single-person effort quantitively compared to xipo, but the guy told about himself by not passing on hard to optimize games, making many first-generation runs (not all of them got accepted, though), doing them quick, and doing them mighty well.
No. Nooo. >_<
[EDIT]
After much contemplation, the vote goes to adelikat. His runs were all of great quality and were arguably most entertaining of the bunch (because Randil tends to pick too many niche games :P). Also, taking the lowest rated movie on the site and improving it was a step too important to forget.
Well, this was really tough, but I think Upthorn will get the vote this time.
Also, I'd suggest removing nitsuja and placing other good people (JXQ and adelikat, for example) on the list.