Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
That was really interessting!
I'm wondering how many generations it would take to overcome a human TASer, probably too much to fit in our life span? Does computer architecture on hardware basis that is like a neural network exist? Maybe that would make it more efficient, since you don't need to simulate the neural network then.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
To be fair, the movie rules needs better clarification on co-authoring others when input is copied. Depending on the situation and the game, co-authorship makes sense or not. For a simple game where a section consists only of pressing right and no astethic choices can be made, it makes no sense to add co-authorship when the input is copied. However when the game is rather complex and allows astethic choices, co-authorship will be needed, except the TASer makes own asthetic choises by heavily modifing the original input.
Of course when the TASer thinks others deserve co-authorship, even no input from them was used directly, it welcome to do so.
The problem with that submission is that it is a categorie no one did before, so an user who is not expierenced with the game has no obsoleted version to compare to. Also the lack of information on submission texts makes users sceptical about them, in this case it has no meaningful explanations, then users come to the conclusion the TASer is experienced.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
The movie file (bk2) only consists of input data.
If you wan't to use cheats in a movie playback, you need to load them as well in the cheat dialog.
Also why the heck do you need an infinite lives cheat?
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
Here's a WIP of E1M1 and E1M2: User movie #23321172006617535
Nothing interesting, so don't expect anything entertaining.
Link to videoLink to video
Sorry for the camera shaking, but there's a trick with it to built up speed. Rotating and pressing the c-left or c-right on the right angle will change the speed direction and value.
In the end of E1M2 a damage boost can probably be added.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
Baxter wrote:
You could have just watched the second quest TAS...
Either way, the HUD is a transposed section of a part of the screen below it. Not sure exactly how it was transposed.
Well, it's not a loss for the submitted run, and for the v2 version it may or may not be a loss.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
Do you save greenzones in your project?
If yes, just scroll to the frame where you left off and continue from there. You don't need to playback your whole movie.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
pirohiko wrote:
By the way, the memory address that a file name is stored is $0638 and $0640 and $0648.
You can write it freely there more than 16 characters. ($#00-64)
If you avoid $0618 and run a file name, is it shortened?
Well, only 16 characters. The file name for the used save slot needs to be "ZELDA". But 16 characters should be enough to write a routine that writes the glitch items directly.
The problems are first to get to $0638 without crashing and second writing the instructions.
I look later into it...
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
Nahoc wrote:
Do you even know why people don't bother to change to bizhawk?
TAS input is more clunky, savestates are slower, memory addresses are different and most games aren't more stable than on mupen (glitches still occur on both emulators).
I don't know about the Virtual Pad since I don't use it, but I can confirm that savestates not slow, even in TAStudio (actual savestates, not greenzone). Maybe you need to set the compression to 0. The memory addresses are different, because MHS uses the memory addresses from the mupen64 process on your computer, while BizHawk uses memory addresses from the emulated game. Not sure how messy the addresses are in the mupen64 process, but if you found an address in BizHawk you could theoretically calcualte all other addresses using the difference between addresses.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
The glitched TAS will not obsolete the current 2nd quest TAS.
In the glitched run none of the dungeons are beaten, if game end glitch works out, the 2nd dungeon is visited to get the flute. In case game end glitch doesn't work, the last dungeon is beaten by warping to the room with Zelda.
Also I don't even know why the published runs are tagged with heavy glitch abuse, the screen wrapping can be done by real time runners.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
For those who are interested:
The glitch items happen to become available in the routine at $EC6E. It overwrites the values of a large chunk in the RAM including items available and the item index at $0656.
The item index needs to be greater than 8 to use the glitch items, preferably $66, because that is the index for the warp bomb. The flag wheter this item is possesed or not is at $06BD for the warp sword at $06CF. All flags for glitch items need to be true, else the item index jumps back to the nomal items if it hits a false flag to some extend, I think if there are to much zeros between to addresses it jumps back.
It might be possible to write a routine into the enemies positions, but manipulating 11 enemies is a horrible task, plus other stuff the to be at good condition too.
Editor, Experienced Forum User, Published Author, Skilled player
(1507)
Joined: 7/9/2010
Posts: 1317
Resyncing runs isn't just adjusting lag, the RNG can also cycle on lag frames and thus has a different outcome on an emulator which lags longer. So what want you do? Waiting until the random values are the same, which leads to a suboptimal run.
For every TASer who wants to start a run whould pick the recommended emulator from the site. If someone was already working on a run for a long time, before the new emulator existed, he/she can finish and submit the run with the old emulator.
