Actually it would be even nicer if two games where the fire buttons are the same were used because then the sync between the two videos would be even clearer.
I'm just wondering if in auto-scrolling games like this there aren't any places where it's simply impossible to play both at the same time without dying. Just imagine one of the games having a long narrow corridor on top of the screen while the other has one on the bottom of the screen... Would be quite bad luck, though. :P
The h.264 video encoding wasn't that impressive, many random display errors and frame drops, not used to watching a run with that many errors - although that's not a JXQ issue - just an encoding process issue.
I didn't notice any such problems. Are you sure the problem is not in your end?
It is relatively easy for a newcomer to create a TAS within 90% of the current TAS record.
I would like to see a newcomer creating a megaman1 TAS which is even close to 90% of the current record (especially without seeing the current version)... ;)
Also, some runs require such a huge amount of planning that it's basically impossible for a "newcomer" to just start doing it and get only 10% behind. Think of the Zelda TAS. (Again, assuming the newcomer does not watch the current run first...)
But with most other runs you are right, of course.
Coding self-made array classes in C++ is rather tricky to get 100% correct. Dynamic allocation of memory is precisely one of the things which are most diffiult in C++ and it's very easy to shoot oneself in the foot.
Unless you have very good reasons to not to use std::vector ("I think it's inefficient" is not a good reason because you are probably wrong; "I don't know how it works" is not a good reason either), I recommend using that instead of allocating arrays "by hand".
In general, you should avoid 'new' as far as possible. If you can avoid 'new' by eg. using STL containers, you should always prefer doing so unless you have a good reason why none of the STL containers is efficient enough for what you are doing.
Sometimes having to use 'new' is unavoidable. In that case you should always remember good encapsulation, and you should really learn how constructors, destructors, copy constructors and assignment operators should be created for such classes.
The easiest way of avoiding problems with classes which allocate memory is to simply disallow the copy constructor and the assignment operator. This is done by declaring them in the private part of the class (and not implementing them). Often these types of containers can be used without the need for copying or assignment.
If you simply can't avoid having to copy and/or assign these objects, then you'll have to be very meticulous about the copy constructors and assignment operators. I recommend the books "Efficient C++" and "More Efficient C++". They teach these and many other useful things to know.
In the Quake speedrunning community 100% runs, ie. runs which get all kills and secrets, are rather popular. I was thinking if there were any console games where such run would be feasible (supermetroid is not feasible because the monsters respawn, but this seemed like a natural place to post the question...)
I agree. Who the heck would be stupid enough to think that preloading all the links in a page could be a good idea?
The whole idea of having a website organized hierarcically in pages and subpages, forming a tree of linked pages is that you don't have to download *everything*, just the things you are interested in.
This comment had something deep in it: "It's like tasting a bit of the dark side."
Nach wrote:
I don't see mountain climbers quiting because they invented jetpacks. Whatever happened to doing something for the sake of the challenge?
There's one prominent example of this: You can fly comfortably in a plane to the south pole. It's not even very expensive, if you are in eg. Australia. Yet there are people who walk to the south pole. Why would they walk there for months even at the risk of their lives when they could just jump onto a plane and fly there comfortably in a couple of hours?
Why indeed. The point is not getting there. The point is the journey, the achievement, the feat, the incredibly show of strenght and perseverance.
Do people flying to the south pole with a plane somehow lessen the achievement of someone who walked to the south pole?
I have the impression that Fastclick/Valueclick ads are usually quite distracting and make me want to remove them from my view.
I do not wish to offer that experience to the users of this site.
Besides, if I'm not completely wrong, the cookies given by those ads are considered spyware by most spyware-removing programs. Not very recommendable for a serious and honest website.
Coming from Lisp, Ada, and QBasic, I guess my warning about C/C++ is that they don't do much bounds-checking for you.
The great thing about C++ is that you can create data structures which are easy to use and have automatic bounds checking or whatever you want. In fact, such data containers already exist in C++, such as for example std::vector() which, if you index it with its at() function it performs bounds checking. (It also offers unchecked indexing if you are worried about performance.)
In C you could offer "bounds checking" by writing a separate indexing function, but the problem in C is that nothing forces the programmer to use that function, but instead there's a too high urge to index arrays directly.
In C++ it's not only easy to add bounds checking to data containers, it's also very flexible on what should be done if boundaries are crossed. For example, you could implement it so that the data container grows automatically as needed.
I didn't say C++ is more efficient than C (at least not speedwise). I just said that once you master C++, it can be a very efficient language.
However, when designing and implementing a (big) program C++ can be a much more efficient language than C. You can achieve more with less code, and you can write code which is much more modular, secure and easy to use, making overall developement more efficient and less error-prone.
In a few cases C++ can actually be speedwise more efficient than C too. The GTK+ library is a prime example of this: In this library pointer-cast checking (between GTK+'s widget types) is done at runtime, while with C++ it could be done at compile time. Also many arguments are given as char strings which are parsed and interpreted at runtime, while with C++ it would be easy to make most of those things at compile time. I'm also rather sure that eg. dynamic binding is faster in C++-generated code (using C++'s own tools) than with the GTK+ C code which emulates dynamic binding.
Would C be your first programming language which you would be learning? If your answer is yes, then don't. C is probably one of the worst programming languages to learn as a first language. It teaches you huge amounts of bad habits which are really, really hard to eradicate.
If your answer is no, and you are quite fluent with some other language already, I would still not recommend using C. If anything, I would recommend C++, using a book which approaches it from object-oriented programming point of view.
Personally I love C++ and I hate C. C is good only for two things: To make very small programs for an embedded system (at least assuming that no C++ compiler exists for that system) and to make an IOCCC entry. For those things C is great. For anything else it's really bad.
C++ has tons of quirks too, and there are tons of ways to shoot yourself in the foot. However, once you master it, it can be a wonderful and efficient language. You can't say the same about C.
I have used some big libraries made in C which are very well done, one good example of which is the GTK+ library, but due to the limitations in C they are inherently cumbersome to use.
If things like "object oriented programming" gives you a negative feeling, then perhaps you should learn more programming and less C until you start liking that concept.
I'm so downloading this one!
(Not that I wouldn't download other videos. In fact, I have every single published video in my HD. However, I'm so downloading this right now!)
To those who are interested, I've mirrored the file on the Internet Archive along with three re-encodes (at normal quality, which is comparable to Bisqwit's encode, which is 26 MB, low quality, which is 16 MB, and high quality, which is 134 MB).
May I ask why? I really can't even begin to imagine what the heck would be the advantage in making a "low-quality" version of the video which is 16MB compared to the 32MB of the original. I can't begin to imagine how in the modern world it would be so much advantageous to download just 16MB compared to 32MB. I could understand it if it was 160MB vs. 320MB, but with such low filesizes I just don't get it.
Also, what the heck is it with a whopping 134MB file?!? What the heck for?!? Who would be so crazy to download a 134MB file when the original 32MB file is so much smaller and which image quality is superb? (Of course there will be people downloading that, assuming that the 32MB file probably looks like crap, when it really doesn't. That doesn't make it ok, thought. You are just making people waste their bandwidth and disk space.)
I, too, am also wondering the same as a previous poster: Do these videos have the proper texts explaining the nature of the videos?
Why mirror them anyways? The more people download them from somewhere else, the less seeds the original torrent will have.
I, for one, rather enjoyed the brief 'subtitling' explaining the magnet beam graphics. As it is obvious that the movies are becoming increasingly impossible to follow without behind the scenes explanations, it might help if more subtitling like that was introduced, to explain the glitchier segments of the movie, pointing out the exact locations for some of the more obscure, or hidden tricks.
It has been suggested in the past that TAS makers could create subtitle files explaining those interesting bits of information. There are standard subtitling file formats which are easy to create by hand (ie. they are pure ascii) and which many mediaplayers support.
Whenever one movie gets ahead the others, pause the ahead-going one until the others catch up. During pause, the screen might be dimmed or something to indicate clearly when it's going ahead.
I think that the original idea was more like showing them sequentially, ie. first show the level from the first published movie (or even better, if it's "too" glitched, just create a "regular" unglitched tas of that level), then show the same level from the next published movie and so on.
I'm not discrediting what you guys do, in fact I admire the hard work and dedication you put into making everything pixel perfect. To me this site is more about hard work to make a finished masterpiece, rather than a show of skill, and I respect that, which is why I thought of posting this here. But I also don't consider the run we did in the same category at all.
It's very nice to get some positive attitude from speedrunners for a change instead of all that "fake cheat" crap. :)
Regular speedruns and tool-assisted speedruns indeed form two quite distinct categories, and both forms are quite enjoyable in their own way. I believe most of the people who enjoy TAS runs (because they often look so awesomely cool) do also enjoy regular speedruns (because they show incredible human skills, besides looking awesomely cool too). I myself am a big fan of the QdQ runs and I always anxiously wait for new runs.
It's too sad that so many people dismiss tool-assisted runs because they have this prejudice that they are "fake" and "cheated" and "too easy to do". All that is complete nonsense. There's nothing fake nor cheated about them (they are genuine and honest tool-assisted speedruns), and the amount of work which goes into making them is often huge (just look at the amount of work which has been put eg. in the latest rockman submission). It's sad that so many people are missing all the enjoyment of TASes because of their prejudices.
As I said, it's nice for a change to hear positive comments from the speedrunning community. Not all of them are prejudiced. :)
And then some people despise tool-assisted speedruns because they believe they are too easy to make...
Btw, a classification idea came to my mind while reading this thread:
First-generation tool-assisted speedruns: This is where all began. The genre was new and mysterious, and everything tool-assisted looked superb. Almost everything made "as fast as possible" using emulator tools was classified as a good TAS. Typically optimizations didn't go to frame-level, it was just enough to perform "superhuman" feats and complete the game faster than any human playing in the regular way could.
Second-generation tool-assisted speedruns: Several emulators were being enhanced with tool-assisted speedrunning in mind, bringing more stability to rerecording as well as powerful tools such as frame advance and frame counter. This allowed for much more precise optimization to squeeze ever last frame out of a TAS. Also more attention was paid to look for glitches and other similar things in games which could be abused.
Third-generation tool-assisted speedruns: The games themselves are disassembled and studied, tools are coded to optimize the TAS (such as robots) and emulators are patched on a per-game basis in order to make the optimization of the TAS of a certain game easier.
I don't think that's a good solution. If you vote "undecided" you are expressing an opinion. Too many "undecided" votes says something about the movie: If many people have concerns about the quality of the movie, then perhaps it should be reviewed more carefully.
OTOH, if many people just vote "I don't care" it just means that the game is uninteresting and the movie doesn't make it any more interesting either (but it's not bad).
Does "meh" mean "I don't care if it's published or not", or "I'm not sure if it should be published or not"?
IOW, is it "no opinion" or "undecided"? There's a significant difference between those two. "No opinion" is like not a vote at all, while "undecided" means that the voter has some concerns about the movie and can't decide if it should be published or not.
Perhaps there should be two different "mehs"?
Just a small comment on the game itself (not this run):
Am I really the only one who is overly bothered by the horrible main character graphic? The legs are way too short and the body way too long. It makes the character look horribly disproportionate and ugly.
And what's with the horrible-looking ugly running animation? It looks just awful.
What I wonder is: Didn't the game makers see this? They must have seen it. Why didn't they just tell the graphics artist to remake the main character graphics? I bet he could have fixed both problems in less than an hour.
You seriously can't compare huge blinking gif ads to the moderate text-only google ads.
and it's not intrusive like forced banner ads are.
Which banner ads? I can't see any banner ads.
Well, I actually can see one. It says "Nesvideos - Tool-assisted superplay movies" and when you click it it goes to the nesvideos main page.
Secondly i won't recommend Google's Advertisement Why? because they pay for Click not Impression. As every Regular visitor know that Site is getting Highter Page Visits daily, so i suggest to use a Advertiser who actually pays for CPM (Cost Per Thousand Impression)
I have no idea what do you mean by "Impression", but I disagree on going to anything else than google ads.
As far as I have seen, google ads are the only ones which are moderate, text-only, non-intrusive and non-annoying ads. Everything else seems to be blinking ad-banners on tops of pages. Those are annoying and distracting and I strongly oppose using them. They also affect negatively the overall look and layout of the pages.
