It's always the last type you check. by pepper at 2:23 PM EDT on June 25, 2014
Farcry Vengence for Wii. I'm honestly dumbfounded by the fact that they used standard ogg files in the sound dat archive. It just blows my mind since I've never seen a wii game use something other than a sequenced format or adpcm variants. I'm even more dumbfounded I didn't spot the fully intact headers in the dat.
Anyone else have any stories of overlooking the smallest of details and coming back to the format years later and feeling like an idiot?
i was looking for the music for it and overlooked the file they were in, it took me 3 years off and on to find them, until i found them in 1 day in a dam asf file, i felt like an idiot when i found them
I've rewritten ima_rejigger2 a few times, having had to work out a) that it's IMA and b) that it's rearranged a particular way from scratch each time. Just tend to forget it over the years, and then I get the code written and I realize "oh crap".
I don`t see how usage of Vorbis surprises you that much, I`ve ripped very very few WII titles all by myself but even in such a modest resume as mine you`ll find at least two video games which developers not only used OGG for encoding music but also left tracks laying wide open for anyone to pick them up. Of course those 2 are pretty small VGs by developers like Team6 but still it`s a fact that WII is able to handle OGGs just fine and once again it`s usually not console itself that matters the most but game engine and most importantly developers` will to implement whatever codec they desire to use.
PS. Oh and additionally it`s very likely that some 3rd game made by the same team as one of those 2 I ripped personally uses OGG Vorbis for music as well.
PPS. To be slightly more on topic: I can recall saving some XWBs which simply wouldn`t convert with ToWAV but when I remembered them and continued my assault a few years later getting music out of them was no big deal at all as I`ve found out about not only VGMStream but also UnXWB and even way more useful tools. I hope this brings something relevant to the topic.
Of course it's always the last type you check, why would you keep checking after the last type?
Searches in general don't continue once you've found a match, unless you're searching for everything, and usually the universe to search is infinite.
There's another situation where you would keep searching after you found a match: If your search is being carried out in parallel.
You may for instance be distributing the workload among several copies of yourself (think multicore), one of those might continue working on its own chunk after "you" have found the solution. The multiple agents could avoid this by conferring with each other after every check, but the communication overhead here becomes prohibitive, and unless each check is the same duration there will be a lot of idle time while the faster checks wait for the slower checks.
Also consider a "batch" approach: Perhaps you are collecting many samples, but the testing procedure is sufficiently expensive that you must combine these samples into one to test them together. You may not even know which of the samples collected was "it" if the whole batch matches, so it can easily be the case that the last sample collected wasn't the one containing what you were looking for. I guess this is only helpful in the case of detecting the presence of something. Particularly if something is so rare that you don't expect to find it at all, this may be a precursor to a more fine-grained search.
Or how about a search for "isolation": You are searching for something, but you don't know it is the target until you later notice its absence. The example I'm thinking of is on a road trip: You are looking for a gas station that is isolated from another by at least distance d. The only way you can know that a gas station meets that requirement is by having passed it and not encountered another one within d. So it is quite specifically not in the last place you look.
A more general sense of the isolation search is that you are looking for a "best" match. You may have a large set of options to look at, but you may not know what works best until you have exhausted the options. I think it is just as often the case that the first thing you tried has no competition as it is the last thing.
This "best" search is very interesting, in real situations we apply heuristics to try something more likely to match first, and we may give up when we get something "good enough" if we don't have another candidate lined up that we expect will be substantially better. Ideally we'd even be able to collect information from a failed attempt to better inform what goes next.