Roll Eyes
Lips Sealed
page  <- 12345678910 -> <- 1 .. 8 .. 10 ->
List results:
Search options:
Use \ before commas in usernames
Edit history:
LotBlind: 2016-04-23 10:37:10 am
I've been looking at the best way to get the study key (and you seem to have asked about that): you can get it from outside the room and there's no need to move the clock or anything. The fastest route for doing this (after grabbing the talisman) is very much fraught with danger in the shape of arrows in the back of your head: you temporarily enter the corridor from the W and have to run the hell through to R7 immediately to avoid a bad bad fate. Smiley This and the other E2 alternatives are presented here.

Door quick turn: alright I'm seeing it now. It's just that the door I was testing it on was different. In fact it works best with doors that start opening instantly (no delay during which you take steps backwards), but can also be done with the other kind. It might be useful if there was some trigger somewhere behind a door that you wanted to trigger and then run back, e.g. you can activate the zombies in E3R12 and run right back outside. This "defuses" the trap but doesn't unlock the door to R13.

Testing: Knowing those little things helps us put them in proportion and to get them out of the way. We might find some hidden small variations in length and learn how to manipulate them. BTW you could also count how long the delay before being allowed to select stuff in the inventory is so we know how much time each fast scroll saves. This would give me an idea of whether I need to be trying to manipulate those Tongue

I'm seeing more places where if you zoom in enough, you can make out little corners between doors and walls that you really can get caught inside with really good positioning. Any door that's either jutting out, or is locked should work for this. The others will just open when you collide with them inside the doorway. The only new one I've found isn't particularly useful, but if you want, check out the version history in the google sheets. There's lots of doors that look like they jut out, but I think it might be some kind of rounding error in the viewer. I can't ever get properly caught in them past the 44 degree mark, and so clipping is not feasible.

tigrou: a small thing - why are the coordinates given in the order X Z Y? Is that some kind of convention?

EDIT: Well hang on a minute... I did manage to find a pretty swift new clip. It's the one from E3R13 (but it starts and ends in R1). Should be a faster way to get to R11 without having the key.
Edit history:
tigrou: 2016-04-23 12:45:40 pm
tigrou: 2016-04-23 12:45:12 pm
tigrou: 2016-04-23 12:32:50 pm
tigrou: 2016-04-23 12:31:12 pm
tigrou: 2016-04-23 12:29:27 pm
tigrou: 2016-04-23 12:29:05 pm
tigrou: 2016-04-23 12:28:14 pm
tigrou: 2016-04-23 12:20:07 pm
tigrou: 2016-04-23 12:19:48 pm
tigrou: 2016-04-23 12:10:48 pm
@LotBlind : the room viewer use same axis convention as in Unity : Y axis is the height. It's like that because it's easier. I'm not sure which convention AITD use (eg : Z axis for height). It's not really important.

No feedback about new UI  ? I mention it in my last post but nobody saw it apparently.
Shut up and go back to another earth
Oh dang ! How rude, I tested it and I didn't say anything.

So the interface is nice, having an option menu feels really clean and professional. I wonder if it's not a good thing to keep keys command in the same time though ? So the player interacts with the viewer as he prefers. About all the options I'm sorry but I can't really tell, there's a lot of features added recently that I didn't use, as my use of the viewer is a lot more basic than LotBlind. :s All I can say is that everything is clearer (but I think it's already the case since a few updates): Walls fading away but still visible when you're in another room, precise angles and position, all informations about every hitbox & trigger... But once again I didn't really try all the possibilities, I guess LotBlind's opinion'll be more interesting than mine ^^'

Oh by the way, there's a guy who posted me this on Youtube, he made it by himself. I found it awesome it is my new wallpaper:

Wow, path for study key is really unsafe =o what about the one in your edit ? I managed to reach R11 when Tigrou showed us the viewer and that I tried it everywhere but it was a long and complicated path who sure wasn't interesting to use (besides impossible without the map).
The alternative path by study room efficiency really depends on how fast we can reach R11 (with or without the key), as all the time saved is always balanced by other necessary detours (Saber, key...). How is this new clip ? Can you still grab the lightfire ? How fast ? I will try more the complicated OOB to reach E6 but if I don't find a thing to be faster without loosing myself, it'll be a certitude that study path can't be the good one (at least for a SS).
Also, the "possible faster way to clip out of the boss room" does not seem to be efficient. Even if you train to clip a little faster than in your video, you still are something between 10 and 15 sec behind the classical path.

Did I forgot an alternative ? You found so many paths those last weeks I couldn't tell x)

Door quick turn: I think I only used it in my first run of the game (11:54) when I entered in E2R7 so my trajectory was directly pointing toward the fake book.

Now, about all tests I'm gonna put all the times I took. But remember, and it's also one of the reasons why I think it's not THAT useful, a lot of things could make vary those times. I could be faster or slower, maybe the game lagged a little, maybe the size of the room with all loaded items in it has an influence, we can't really tell. But at least I found some numbers who seem consistent.

Changing action (from menu opening to the comeback to the game)
-Action 1: 0,923
-Action 2: 1,001
-Action 3: 1,234
-Action 4: 1,309
-Action 5: 1,368

What is the longest here is the time you have to wait to come back to the game once the action is choosen, this time may vary, I had an action 4 in 0,868. Also we talked about a long time ago, sometimes the game kinda saves some moves you're doing during gameplay and loads it on the menu, the consequence is pressing down one time'll make you go down several lines in a row. It happens in the 3 episodes of the game, I can sure tell that it depends of your gameplay because when I speedrunned the game, it happened always at the exact same place( I just had to learn it), but if I used a save it erased those phantom moves. Anyway, just to say that these can also alternate results as you go down a lot faster.

Move in inventory (from inventory opening to frame where target is pointed):

-item 1: 0,134
-item 2: 0,234
-item 3: 0,468
-item 4: 0,634
-item 5: 0,768

I found weird to fall often on same two last numbers but maybe it has something with Sony Vega reading 30fps videos by cicles ? I don't know. Important thing: the phantom moves described up above obviously works with item lists too. Theorically you can combiny those results with "changing action" times, as the exact same action is following (enter, then choose the good action and enter again). I did not time a lot of items because at this point we know that the inventory won't be that full. Even with Study Path including the key, you'll have 6 items just before using the key (in 5th position) and the saber (in 3rd position).

Door opening (from the moment where character stops running to his recover):

With recoil: 3,470 - 3,504  - 3,539
Without recoil: 2,202 - 2,269 - 2,436

2 things: obviously, doors that you open towards you are longer to open because the character make some extra steps. You also can add some extra time when you're triggering a new room during the opening, wich leads to the next list... Except that, times don't vary that much. especially considering that I can have some extra frames timed in the process; I tried my best but it's complicated to be sure of wich frame is the beginning of the animation/ the one where you control the character again.

Open/search animation (from animation first frame to item window appaerance):

I took 4 chrono and found 3 times 1,034. I took it as a good sign x) If my feeling is true, I think that the animation of character using an item (mainly the hook at this point) is the same so the time should be identical.

Point of view changing:

0,300 - 0,133 - 0,300 - 0,234 - 0,334 - 0,267 - 0,200

0,367 - 0,400 - 0,434 - 0,601

Not so much regularity here despite the fact that it's always around 0,250. I suspect the fact the more it has to show items, the longer it'll be. But i can't be sure of that affirmation. The 2nd line only contains times from camera changing INCLUDING the load of a new room, and as we can guess it is a little longer.

I did not time the window from item pick-up as it mostly depends of the player hability in a really short period of time. I think my time wont be representative of anything. If I had tools I could find a theorical best time but even though I'm not sure it would be useful for not-assisted-speedrun.

By the way tigrou, nice avatar ! as Chibi would say: "Hi ! Nice mustache" =p
tigrou: I forgot there had been another update! Sorry.

Is it now intentional that you can set both "follow player" and "follow room" on? Because it's so late in making this run it won't be a big deal, but otherwise it wouldn't have hurt to have the key commands also left in. I understand it's more work to maintain both.

The camera views have the opposite names to what they're supposed to.

I don't think "areas" is the right name for the camera views. How about "camera sectors" or just "cameras" should be fine.

For some reason, only hitting the right shift key works for skipping 10 models, the left one doesn't.

You could include a warning that if you switch to the model viewer, it resets the room viewer completely.

So in school, I was taught that to draw a positive angle, you start from the X-axis on the right side of the Y-axis and draw it counterclockwise, so N is 90 degrees, NW is 135 etc.. I don't know if that's any kind of international standard. Looks like you made it start from N and go clockwise from there. If you know this is standard, then keep it.

Looks like the option to "follow" a particular collider or actor is gone. The one where you hit the RMB to get it to show the info on the lower right. Is it still possible to do that? It isn't mentioned in the readme either.

BTW, even if you make all these changes, none of them will affect the speedrun and so you might as well wait for more feedback just in case.

NHG: the "edit" path: just assume I've put every clip I've mentioned in the clips table, and you can see it if you hit "show edit history". You can't really get the lighter without first clipping into R13 to get the lighter, then opening the door to get back into R1, closing the W door and starting the clip again. I.e. if you got matches earlier, not worth getting the lighter there.

I recommend not trying to piece it together yet NHG. Let's also just assume nothing is "random", because it isn't. Could I ask you to put all your results in a table somewhere and keep updating it? That would help keep things organized. I trust your ability to optimize (to a sufficient extent) all the different routes and am ready to leave most of all things for you to test, because you've already done multiple well-optimized runs.

So if you got an action 4 in 0.868, that makes it .45 seconds faster than normal. Did that all come from a faster menu closing time? Did you run all the tests in the same area with the same things happening? You should always minimize variables as I'm sure you know. Still if it depends on which area you're in, we could maybe save time by changing actions in a different area. Another thing is you might wanna record all your testing, even if you never end up looking at the videos again, and keep them around until the runs are out, just in case something weird happens.

As I understand from your notes, if you got 1 "extra move" in the inventory, you could reach item 2 in 0.134 seconds, i.e. 1/10 second faster? How about did you ever get the quick inventory glitch during testing? If I needed item 5 and I could quick scroll down instantly, how much time would I save? All you have to find out is how long the interval is between two items during faster, continuous scrolling, and times 4.

Yeah, the times you're getting aren't quite exact because there might be duplicate frames if you record at 30fps but it runs at 25fps. Why don't you just use native DosBox recording? Sorry if you've already told me... It's fine though, it's close enough for a non-TAS.

Yes, the time when an open/search animation is complete (and the script proceeds) should always be the same. That was one of tigrou's findings but I'm not surprised if you missed it between everything else. We didn't know it was 1.034 ofc.

Wait, you should be mentioning where exactly you tested the camera view changing/room changing. Could you run some tests between the same two rooms/views and make sure those are always the same. Still, I think a running average and the variance is enough as I said, but I'd like to know where the biggest and shortest load times are. If you wanted more exact knowledge (feel free not to), you COULD try to manipulate the variables one at a time like this: take some room/view change somewhere. Look at how many things (how many objects, triggers, even colliders) were loaded into memory and how many got wiped. Then the easiest thing to manipulate is objects: make sure some of the objects, or all of them preferably, are ones that you can move, either by picking them up, or by pushing them so they no longer get loaded into memory during the camera change. Now you can remove one object at a time and see how much faster that makes the transition. Then you might know how long it takes for the game to load in one object. If there's slight variations, you should tabulate everything and compare how "fast" each object is to how simple its model is. Now you might also know WHY there's variations.

After this you could take two rooms that are otherwise as similar as possible, but one of them has considerably more triggers. Etc... and that way you can keep making more and more deductions and one day you can tell pretty much exactly why a certain transition is longer or shorter than another one. Again, I emphasize, if you don't find this in the least interesting, don't wear yourself out, but this is precisely how I'd work on a game if I really wanted to understand it.

Just some more tips to make your testing even more efficient: it's good that you remembered to mention which frames you started and ended your test on. Those should be the first frame when you're no longer running etc. until the first frame when you're running again. Ofc it's also good to know what the delays are like inside the menu or whatever it is. So e.g. for opening the inventory and selecting an item, I'd do it like this:

first frame after you've hit enter (i.e. stopped moving) to first frame when you can see the whole inventory


first frame after you can see the whole inventory to first frame on the 1st item


2nd item... etc.

Then first frame after you've hit enter to select the item to first frame of showing the normal view again.

To get more exact times for all of these you can just buffer hitting enter. I.e. just move the selection up and down one more time while buffering the enter in-between, so you know you've selected the item on the first frame the game let you. If you can't buffer it, there's nothing you can do but try to be quick.

If you work in this way, the major advantage is you can see exactly where the variations happen, and where they DON'T happen. And also we then know not just what the variations for selecting different items are, but also how long the whole "inventory detour" takes vs. skipping using the item if possible, which I don't know if you can see in the data you've presented so far. Then you can just forget about everywhere where you saw no variations and focus on where there was variation and try to figure out what's causing it. I did something very similar for another game earlier - Typing of the Dead - and it was exactly like that. For some things I was using excel with running averages... anyway just some tips. That's how I find all the stuff I find Tongue
Shut up and go back to another earth
Edit path: Don't know why I still don't have the reflex to watch the section you made for the game despite the fact that I perfectly know that you conscientiously note everything you find x) I'll watch it but I can already tell that if using this path is better than picking study key, it's probably better to take matches.

About areas there are many. Some from the last speedrun, some that I recorded for the occasion and of course in many different places, and hum... Well I'm kinda maniac guy, I don't like having tons of files everywhere so I... deleted them after timing it >< Sorry about that.

The phantom moves in the inventory are not instantaneous, well somehow they are as I presume they ask the minimum time for consecutive moves, but it's not equal to 0. Reaching item 2 wont be equal to 0,134 but it'll be less than 0,234.
I never had the trick during tests because at this moment I only was interacting with the menu, and I'm sure the cause is something during gameplay. I have absolutely no idea what, besides the shortest were my new reccords, the less it happened so I kinda ignored the thing more and more (especially considering that most of the time it was more a discomfort than a help).  I'll try to see if I can find a clue about that but I can't promise anything. It could be great if we find how to manipulate this thing and use it to save some inputs in the menu.

I think the last time I used native recorder (not in this computer though), I stopped because the game was going slower and slower with time. The weird thing was once I stopped recording and launched the video, everything ran at the good speed, wich was perfect to cheat, so obviously I took another way to record (or maybe this effect happened only when I summoned a color ghost during the play ? I don't quite remember).

Oddly enough I remember than the short time was during camera transition in E2R10 (the one pointing to the angle of the room) but as for the others... I took very different places on purpose as I had the feeling it'll have an influence but I did not really note the places x/ All your process explanation is really relevant I have to say but I really don't know if I'll have the patience to try all of that x) I really like all what I learned about how the game works thanks to you, Tigrou and his brilliant viewer and explanations, but try to understand the game in a "micro-procedural" proportion... I think I won' t have fun working on it :s I would find motivation to do that if it was a TAS because it's necessary to optimize the thing, for a speedrun it'll ask a lot of time and effort that I won't enjoy >< I really like to search and test new shortcuts, tricks and glitches, compare their time, try to optimize them etc until I can draw a "good path", but those micro tests... I don't feel like I'm speedrunning :/

The thing is I really want to help you as you helped me so much finding a tons of tricks, that's why I timed all of this in the first place. But it goes too far for me xD When I think about how much time it took me to collect all of these informations (wich are yet not so many ) I can't imagine the time it'll take to do ALL of your new suggestions t_t

Anyway, in any case I'm gonna finish tests about different paths: see if it's possible to be efficient with complicated OOB in E5, compare all possibilities for study path (key+lighter, no key+matches, wich mactches/oil etc) and I'll give you my conclusion. I'll considerate the possibility to do the library without oil just to see what it leads.
Edit history:
LotBlind: 2016-04-27 12:48:48 pm
LotBlind: 2016-04-27 12:48:14 pm
Use the native recorder, just stop and restart recording now and then and the slow-down should not be so big. Turn off music if it helps.

I'll be focusing on drawing more maps, I just got a new tool for doing so after the old one expired.

Alternatives for E3:

A) OOB straight into E4.
B) OOB into E4 but get the matches first.

C) OOB into E5 through R11 without key by doing the first option in this map.
D) OOB into E5 through R11 without key by doing the second option in the same map.
E) OOB into E5 through R11 with key by simply clipping from R1 through the R13 door from the W side of the W door half (presumed faster than through middle).
F) Get into E5 with key through R12, like this map shows. It mistakenly says "consider getting matches" when the lighter is also available.
G) The same but with the OOB method (B in the same map) to avoid getting the soup and opening 2 doors.

Just to drop a few off: if you have the key, using method E is faster than method F, and F only has the advantage of giving you the oil, but taking the oil in E1 and then method E here beats that easily. Neither require actual OOBs. G requires an OOB, and only saves time in getting to ignore the soup really, so it's probably even slower than F. If you don't have the key, option D is probably faster than C with no downsides. This leaves us with A, B, D, and E.

I noticed if you push an object far away when you're OOB, then go to the save menu or inventory, it'll disappear. In fact what's happening is the game is refreshing the RAM and doing some checks again, including which objects are supposed to be loaded in and which are not. To see this happening you have to at least push it into another room (through a room trigger).

Here's a new way to clip through objects (read "doors"). So when you push a lightweight thing so it's a little ways inside a door (i.e. where the door would be but there's a different room loaded into memory), making sure the thing you're pushing is in a room that makes it be loaded in when you approach the actual locked door later, you can clip through the door by running into the lightweight thing at an angle of about 45 degrees. Evidence. This only works if the object is no more than a little ways in, and you can't use it for going both directions for that reason, I don't think. You could place the chair so it's mostly through the door inside R11, OOB through yourself using another method, and use the chair to clip through on your way back. This avoids the key detour.

You can push light objects right through doors, and presumably other thin objects, but this means they go all the way through. EDIT: Looks like you don't have to make the chairs be in the middle of the door, it's enough to just have it besides the door. The angle that works is now more like 5 degrees towards the chair that's next to you (i.e. almost straight towards the door). If you push one chair through the E3R11 door and go get another one for the OOB, that might be faster than doing any other OOB for reaching the room yourself.

Potential use - get OOB from R1->R13->R1 (alternative D from before). The game loads the chair in R12 into memory and you can go retrieve it when OOB in R1. Push it through R13 into R11 and continue the OOB into R11 yourself. Now you have a means of exiting via R11 without the key (and you only needed to get one chair to do all of it). The only thing better than that is if you could use the same chair both ways, avoiding a tiny OOB detour to reach R11.

For some reason leaving R11 seems even more difficult than getting in: it keeps pushing the chair through instead of ejecting you through. I think this might have something to do with the colliders around the chair, and you might need to leave it towards the side of the door instead of straight at the middle. On the other hand if it's around the middle but the other half of the door prevents it from moving, it might work equally well. I'll do some more tests later, but the route I laid out above will certainly be better than taking the key detour. Hope it's also better than going via E4.
Edit history:
tigrou: 2016-05-01 01:47:13 pm
tigrou: 2016-05-01 01:45:21 pm
tigrou: 2016-04-29 04:18:55 am
Hello. It is not speedrun related but here is some interesting art pictures I found about the game. I think it came from the (cancelled) AITD remake.
It's a lot more gloomy than the original.

EDIT : That jump from underground (found by LotBlind) will always make me smile
I couldn't give a damn about remakes... Tongue that's the only style they even know of these days.

Looks like you can also jump through the cellar exit (instead of running), and this makes you skip the whole crawling animation. LULZ

You could theoretically use the bird in E5 to get some kind of entrance OOB from any point between two rooms. The two ways to achieve this would be to find a way to juke it so it's rotated the wrong way and can't instantly attack you (doable, by juking it around the entrances), or simply by killing it and using the corpse before it dies. I almost managed to make the worm kill it without it harming me, then used the corpse to get a little ways OOB, but then it got me. In any case, because we know of "the long OOB" + the required setup time (the bird starts far away from any openings), this won't ever see use I'm fairly confident.

Another way to do this would be to run to where the cthonian normally is from study, juke it to start moving but by careful tactics you can manage to slip back into E5R1 (might require a potion and lots of health), then use its side to clip OOB easily. Speaking of using the worm to clip OOB, dangit. Can't find any faster ways of doing it, like running into it head-on trying to clip towards the E.

Interestingly, the Deep Ones don't take melee damage seemingly :/

You can jump into the chasm in E5R0 (i.e. use the jump command), and before you fall down, drop some item. If it went right, you'll be suspended in the air and can run around any part of the room as if the bridge was there and covered all of it, which might save time for an in-bounds run or something. Source: dicking around Wink

EDIT: Oh... I guess you can also do stuff like this with it. Oh well. I think if you quickly run to where Emily is in the clip, that's probably the new fastest way into the boss room: even when the stone door is in place, you just squeeze between it and the wall. You'll have to find a good spot for landing (or cancel fall damage somehow) because you can't interact with the sarcophagus unless you can touch it. Ofc you can use the same trick to "navigate" the pirate maze quicker, skip the jumping puzzle... whatever you please really. Tongue It'll stick through a save/load but not when moving between zones.

I'll post some route suggestions later. I was hoping E3 would be best done via study but with these two new things... not looking great for that.
tigrou: I wanted to check what the scripts looked like for E3R12 (dining room). Where in the port are they in? Do I have to have an interpreter?

A tiny little change in the viewer logic: whenever I'm stuck somewhere (during a clip or something), and I temporarily move the camera away from the PC, it won't snap back to where she's at because she can't move, so to correct this you'd have to make it check for her animation changing instead to know when to snap.

Note: Saving the game (or loading) will reset the smoke if you need to. I was also somehow able to glitch it not to happen after some saving and loading in R13 - it just stopped appearing. It's gone in all the different save files I have around that area. That might be really good for a segmented run... if only.

Another way to do E3: enter R12 through R0, walk into the trigger that closes the doors, but backstep away so you don't get trapped, which is just possible to do (or just turn around and run if that's faster). Now reenter R12. Aggro zombie next to the door you used, run to the free chair and push the free chair through the R13 door. Now dodge the aggroed zombie and push his chair into R0 and R1, then through the door into R13. Clip through yourself. Now you have two chairs in R13 (fastest way I could think of). Push one of them through to R12 and use the other one to clip through yourself. Now you have a means of exiting via study without the key. I was trying to do the same by using a zombie instead of the second chair to get inside R12, but I just can't clip using monsters when it's an object-object clip.

Wait a second tigrou: why do you have 275 hp in that clip? Did you hack that? You made a pretty funny clip out of it! I'll use something like that for the funnies reel, unless it makes it into the runs.
Edit history:
tigrou: 2016-05-12 06:21:36 am
tigrou: 2016-05-04 01:28:16 am
tigrou: 2016-05-04 01:19:10 am
tigrou: 2016-05-04 01:19:02 am
tigrou: 2016-05-03 04:48:02 am
tigrou: 2016-05-03 04:46:59 am
tigrou: 2016-05-03 04:46:49 am
tigrou: 2016-05-03 03:36:37 am
tigrou: 2016-05-03 01:17:03 am
tigrou: 2016-05-03 01:15:58 am
tigrou: 2016-05-02 07:41:14 am
tigrou: 2016-05-02 07:35:38 am
tigrou: 2016-05-02 07:30:34 am
tigrou: 2016-05-02 07:30:22 am
tigrou: 2016-05-02 07:30:03 am
tigrou: 2016-05-02 07:28:59 am
tigrou: 2016-05-02 07:28:51 am
tigrou: 2016-05-02 07:28:14 am
tigrou: 2016-05-02 07:28:08 am
tigrou: 2016-05-02 07:27:53 am
tigrou: 2016-05-02 07:20:26 am
tigrou: 2016-05-02 07:18:17 am
tigrou: 2016-05-02 07:17:25 am
tigrou: 2016-05-02 07:16:46 am
@LotBlind :
For the scripts : you need  a tool that I made some months ago. I will send it to you by PM.
275 hp : yes it's a hack. In lot of my savegames, the PC hp has been set to high values. I did that because I was sick and tired of dying against monsters while trying new tricks. Editing player hp is very easy using a game cheat tool like Cheat Engine

EDIT : if Cheat Engine crash your computer, you might try ArtMoney. I used it in the past with success. It is commercial software but if I remember well they have a free version.
tigrou: might look at that later. I'm sure I can figure them out, I know just about enough Cheesy Sadly, Cheat Engine tends to crash my computer a lot so I avoid using it.

I'm looking at a certain path that I'm thinking of as the default right now: you go straight through E3 into E4, clip against cthonian in E5, do the OOB stuff to get to E6 via maze entrance, then use the new levitation trick for fast boss entrance, fall onto the platform in the N of the room, do the boss and exit normally. When in E5, I think the fastest is to activate levitation, run through R8 into R2, clip through and exit via cellar. I think when you factor in the cellar itself, exiting via study is faster even coming from that side, but that means either you've pushed a chair in place for a clip out of the study, or have done the study key detour. Both will take some time. The study entrance to cellars would save time in E5 because you could just go OOB without waiting for the cthonian to do anything, AND the cthonian will remove the boulders for you. This means the escape sequence will probably look like: levitation - R7 (don't have to fall, fall, climb and climb) - etc. exit through study again if have key or pushed chair in place.

NHG: are you looking at timing stuff? To have a baseline table of times would be useful. If you have a hard time naming everything, just include videos of what you mean with each sequence. I should in hindsight have names all the routes.

You can simultaneously clip through a door (used a chair) and start opening the door from the other side (i.e. without having to turn around to do that). Observed with N door in E3R13.

I sometimes fail at pushing lightweights through doors completely - I think this is whenever there's monsters around and the lag they cause prevents it.
Edit history:
NHG: 2016-05-22 03:06:24 pm
NHG: 2016-05-22 03:05:53 pm
NHG: 2016-05-22 02:47:00 pm
NHG: 2016-05-22 12:47:55 pm
Shut up and go back to another earth
Hi there ! Sorry I made a little 3 weeks trip =p it was nice !

I did not use internet all this time but I had my computer and I tried things, many things concerning possibles paths. I took a look at what you found since then but I'll talk about it later. Basically, I timed all differences between a Chtonian path and a Study path to have an idea of what is the best solution. Don't worry I did not forget anything in the process. I quickly excluded the study+key path as the way to grab it is long as hell even with the viewer, clip through study room was a lot better. Still, Results are kinda rude. Every time there's a difference who saves time compared to the Chtonian path, the extra conditions makes it worse. The only thing who really worths it in study path is the "long OOB" taking place in E5 who saves one hell of a time, enough time to correct all the other loses... Well, in theory. Considering all the other variations, Study path saves more or less 22 seconds (comparing to new chtonian path of course) but this chrono was timed with the help of the viewer. I mostly used it to perform Study clip and long OOB. Without the viewer, I tried my best to execute the same OOB. I found a way, then a faster way, then a faster way again wich I think'll be the best I can do as a non assisted-human. I timed this human path WITH THE VIEWER so I can have an idea of the best time this road can gives. It's a lot longer now, between 20 and 25 seconds wich is exactly what this path can theorcaly saves from the other one. Considering the study OOB is probably slightly longer without the viewer and how much this path is more complicated and riskier to realize anyway, I have no doubt about the fact the Chtonian path is still better.

... Aaaaand it is now confirmed by your new Chtonian OOB who allows to do the long OOB in a Chtonian path, there's no way for the study path to be faster now. I saw your chair trick but just putting that in place... as Trump says "waste of time !", as grabbing the key who costs many more time than exiting the mansion by study room saves in (I mean you still have to open doors etc...) I can time those two sequences to be sure but I think it's not necessary.

About the jump trick: Really cool stuff who'll be useful in E5R8 and to exit E5, unfortunately it won't help to reach Pregzt room. For unknown reasons, reaching the room by this trick makes impossible to touch Pregzt with the lantern. The game just refuses to kill him, it reflects the lantern as if the talisman wasn't in place.  Maybe it's because you don't enter the room by the good spot ? I don't know... If we don't find a way to avoid this weird effect it won't be possible to use this trick :/

Edit: Now that I think about it, I'm pretty this jump-trick makes the ghost corridor way faster. I'll time that tomorrow

Edit 2: I finally tried it just after the edit x) it is indeed faster. I also noticed that you can't use the jump to exit E5. If you do, character'll bug when you'll use the stairs in E4 and you'll fall under the ground in E3, making impossible to touch red hitboxes. Did not find a way to avoid that so far.
Edit history:
tigrou: 2016-05-23 09:37:24 am
tigrou: 2016-05-23 08:41:39 am
tigrou: 2016-05-23 06:36:50 am
tigrou: 2016-05-23 05:42:30 am
tigrou: 2016-05-23 05:42:21 am
tigrou: 2016-05-23 05:21:56 am
tigrou: 2016-05-23 05:21:43 am
tigrou: 2016-05-23 05:13:37 am
tigrou: 2016-05-23 05:12:45 am
tigrou: 2016-05-23 05:11:52 am
For unknown reasons, reaching the room by this trick makes impossible to touch Pregzt with the lantern. The game just refuses to kill him, it reflects the lantern as if the talisman wasn't in place.

This is something that would be interesting to investigate by looking in the game scripts. There is probably a flag or something like, that isn't set when you use that shortcut.
Could you provide a savegame in front of Pregzt, when it is not possible to kill him with lantern ?
EDIT : when you put talisman in place, do you hear a special sound (the same as when you grab talisman for the first time) ?
Hiya, I'm just in the process of moving house, I might not have the Internet for a couple of days. I'll take a good look after that!
Shut up and go back to another earth
As far as I know, nothing change except the fact that the Lantern falls into water. No sound or anything. Maybe something about the position of the player ? I have this weird feeling that she doesn't fall correctly when she comes back to the ground, the animation seems weird. Anyway I made the save but I don't know how to extract it x)

What about the problem in cellar after the jump transition ? Here again, I have the feeling that the character is not positionned the usual way. When she begins to climb the stairs, she'll fall a little under it. Then she'll come back to a normal height but at the end of the stairs she will fall again. Probably because of that, she'll spawn under the normal height at E3, making her clip through the floor. I made a savegame there too so you can check simply by using the stairs.
Edit history:
tigrou: 2016-05-24 05:37:02 am
tigrou: 2016-05-24 01:09:33 am
Anyway I made the save but I don't know how to extract it x)

Open the AITD folder (where the game exe is). The savegames are the files named SAVE0.IDT, SAVE1.IDT, ... (there is a file for each save slot, starting from 0). If you are unsure which one it is, zip them all. You can share that file using google drive (or if you don't have such thing, try using a free upload file service).
Shut up and go back to another earth
Sorry I took my time !

Here it is:

I put all files but if I'm correct, the two last ones are what you want.
NHG: Do you have a table for the times you got for different routes? If you have one, I'd like to see it (or whatever form they're in) so I can think HARD on the question once more and see if the time gaps can be bridged even in a segmented run. I'm still kicking myself for finding that Cthonian OOB :/

Which way did you use for exiting after having used the study path or worm path (I'm getting tired of writing that long word now Smiley ) - did you notice that when you do the study path, the boulder will be gone, but the worm OOB doesn't have that benefit? If the boulder is gone would that not make exiting via cellar somewhat faster?

What's the "ghost corridor"? Ah that one. So it's faster than going via E5R7 even if the boulder has been removed?

The weird stairs behaviour: I actually ran into that thing earlier as a result of something I did with the Necronomicon in the library. I couldn't tell you exactly what it was. AFAICS when you use the two evil books, neither will actually change your coordinates, not even De Vermiis even though it looks like it might. In any case I have a save file where the strange behaviour can been seen starting from E2. Here it is if someone's interested. You'll see whatever I did has some secondary effects as well. The flags are messed up in such a way that you don't die from going to below 0 hp, you just get zombified, but even then by following a very precise sequence of steps you can cure your condition, or I think you can even move around on negative hp. Because of the stairs stuff I never got to test this if you start from E3-E4 etc. so I sadly can't tell what would happen if I had activated the glitch only later on. In any case getting the Necro and performing the whole ritual (whatever exactly I did) takes time that I can't instantly see how it could save.

BTW I can confirm that even if you exit E5 via study by jumping, it gets you into the glitched state and you'll fall through a staircase eventually. If you go back down into E5, it looks like the normal state is restored. I noticed you just don't walk far enough when you enter floors that you then fall off, you stop moving a bit before and there's no floor underneath so of course you fall.

BTW To avoid the effect, you can simply perform the jump trick (i.e. try to drop something before landing) in E4 or E3.
Edit history:
tigrou: 2016-05-27 02:42:39 pm
tigrou: 2016-05-27 02:35:03 pm
tigrou: 2016-05-27 02:34:47 pm
tigrou: 2016-05-27 01:24:26 pm
tigrou: 2016-05-27 01:24:17 pm
tigrou: 2016-05-27 01:15:54 pm
tigrou: 2016-05-27 01:15:35 pm
tigrou: 2016-05-27 01:07:36 pm
tigrou: 2016-05-27 01:07:26 pm
tigrou: 2016-05-27 01:07:17 pm
@NHG :
I might be wrong but I think the reason why you have been unable to kill pregtz is simply because the lantern is off. I was not able to turn it on because in the save game you gave me, the oil can is empty and matches are wet.
EDIT: Forget about this.

Trivia : interestingly, the scripts of the CD-ROM version seems to be slightly different from the floppy drive version. Differences are small, they are mostly related to music and sound samples which have been remapped for the CD version (eg: floppy version use midi like files while CD-ROM version use audio tracks)
I tested the boss fight (made sure the lantern 100% was lit) -- it looks like you have to visit R5, then everything works normally. You can do it pretty quickly because you're above the wall, but it's not quite as fast as I hoped, so testing will be needed...

tigrou: Here's a save file where the lantern is lit.
Edit history:
tigrou: 2016-05-28 11:45:31 am
tigrou: 2016-05-28 10:20:54 am
tigrou: 2016-05-28 08:03:10 am
tigrou: 2016-05-28 08:02:23 am
tigrou: 2016-05-28 04:06:31 am
tigrou: 2016-05-27 02:41:37 pm
tigrou: 2016-05-27 02:41:04 pm
tigrou: 2016-05-27 02:35:57 pm
tigrou: 2016-05-27 02:35:48 pm
tigrou: 2016-05-27 02:34:20 pm
tigrou: 2016-05-27 02:32:44 pm
tigrou: 2016-05-27 02:25:21 pm
tigrou: 2016-05-27 02:23:33 pm
tigrou: 2016-05-27 02:21:29 pm
tigrou: 2016-05-27 02:20:06 pm
tigrou: 2016-05-27 01:53:57 pm
tigrou: 2016-05-27 01:52:19 pm
tigrou: 2016-05-27 01:52:01 pm
@LotBlind : good finding. Visiting R5 indeed fix the problem. I think I found the reason :
When you are in E6R6, look near E6R0 center, there is a small actor. This one is responsible for turning on/off the light while you are in the maze (depending if you have lantern or not).
It also set some global variables (those C_VARXX, which i think are necessary to trigger prezgt end).
Here is whole script :
   0: IF PLAYER.ROOM < 6
  12: 	IF VAR178 == 1
  24: 		LIGHT 0
  28: 		SET VAR178 = 0
  36: 		IF C_VAR12 == 0
  48: 			MESSAGE "This room is in the dark!"
  52: 		END
  52: 	END
  52: 	IF "An Oil Lamp".STAGE == 6
  64: 		IF "An Oil Lamp".ROOM < 6
  76: 			IF VAR7 == 1
  88: 				C_VAR13 = 13
  96: 			END
  96: 		END
  96: 		GOTO 142
 100: 	END
 100: 	IF INHAND == "An Oil Lamp"
 110: 		IF VAR7 == 1
 122: 			C_VAR13 = 1
 130: 		END
 130: 		GOTO 142
 134: 	END
 134: 	C_VAR13 = -1
 142: 	GOTO 170
 146: END
 146: IF VAR178 == 0
 158: 	SET VAR178 = 1
 166: 	LIGHT 1
 170: END
 170: ENDLIFE 

So actually being in any room which number is lower than 6 (maze rooms) should do the trick.
EDIT : if you still use room viewer, please take some time to download last version. It contains many small fixes.
Edit history:
NHG: 2016-05-28 05:16:01 am
NHG: 2016-05-28 04:39:28 am
NHG: 2016-05-28 04:38:39 am
NHG: 2016-05-28 04:35:59 am
Shut up and go back to another earth
Okay, I'll time the trick with the detour tomorrow. Before, it saved something like 8 sec in comparaison to the updated maze OOB, we'll see if it still worths it.

Lot': Hope your moving was fine =p I do not have a table, all I kept is some numbers... Of course I took the boulder in consideration xp I took the (new) Chtonian path as starting point and saw that:

saber takes +17sec according toone of my run
Reaching E5 (by study clip MADE WITH VIEWER) takes + 14sec (instead of reaching E5 by Chtonian path)
Without boulder saves -11 sec
Long OOB saves -32 sec WITH VIEWER, but only -10 sec with a good path without it (from start in E5 to E6R0)

As you can see, it may be better but there's an huge difference between using this path with or without the viewer. If you make the operation, it could save 12 sec... Well at least that was before. Using the long OOB with Chtonian clip saves aproximatly 25 fucking sec (without viewer). The boulder advantage is almost entirely cancelled by the ghost path: 46sec between E6R0 and E5 against the same time with a no-boulder path. I did not time the thing with the jump trick though, it'll probably save between 5 and 10 sec. About the exit way I'll check tomorrow how many time it saves to get out through study room compared to E4, but if it does not save 13 sec (wich sounds kinda crazy because if you want to use this way you'll have to add some extra steps) it wont worth it.

Edit: Here's some results. First of all, can you describe Lotblind the way you visit R5 before reaching Pregzt room ? When I'm doing it it's then impossible to load Boss room. So obviously I did not time it yet =p

Just to be sure to not confuse myself with wich time is what I timed some sequences again.

From E3 to E5 by R11 using the key = The time is the same than reaching E5 by E4.
I did not try the chair trick as it seemed to me very long to implement in your description. Also, note that I only talk about the time it takes to reach E3 from E5, I did not take in consideration the time needed to grab the key in E2. So if you compare to the time asked for the same path without key (so you clip to reach R11) you save ~14sec

Chtonian path using the long OOB, from E5 to E6R0 WITHOUT VIEWER = 55sec
Study path using the long OOB, from E5 to E6R0 WITH VIEWER = 53sec
I can't insist enough about the "with viewer thing", the main difficulty here is the transit between E5R1 and E5R2 wich is a lot harder without viewer. I first came with a strategy consisting of triggering cameras to help me but it made the thing a lot longer. So I found another shorter but riskier path based on steps counting just to save some camera transitions. Chtonian OOB begins after this section, wich makes it easier. In fact, considering the fact that I timed this one without the viewer with a difference of 2secs only, I'm pretty sure it's theoretically better. In any case, if we consider a non-viewer run, there's no way for the study OOB to not lose something like 10-15 secs compared to the viewer time (and I'm really nice here, my best score was 21sec lower, I probably can optimize it a little more) I tried a new strategy without the viewer who gave me a nice 1:03. I can reach a 1:00 with training but not less as I'm straight as I can be in my road, wich makes it hardly competitive with Chtonian OOB path (as I'm sure I can optimize it a little too).

Exit by study path (without boulder + study room door already open), from E5R9 to ending = 55sec
Exit by Chtonian path (without boulder) from E5R9 to ending = 1:10
Exit by Chtonian path (ghost corridor) from E5R9 to ending = 1:16
Don't really know why I timed the middle path x) So here we can see that the study exit as itself really worths it. the jump trick cancels the bug that I noted long time ago (when Emily/Carnby automatically re-used stairs once she/he was upstairs half of the time). The problem is, at this point we have an extra 17sec (saber) + an extra ??sec (key/chair trick) to catch up. Let's forget for a moment that this path is a lot more complicated (most because of key path + beginning of long OOB). If I don't forget anything in my calcul, the key grabbing/chair trick will have to take ~5sec to hope saving time compared to Chtonian path. And once again, this calcul is based on the fact that we can accomplish long OOB with a viewer time, wich is nearly impossible I'll let you judge that by yourself.
Wait, how do I know which VAR and C_VAR is what? Is it something I can tell by looking at the first script that mentions it?

Tigrou: I guess you realized you just needed to jump once to get her to fall. Sorry.

Nice! The textured versions of the objects look exactly like the game now. Why wouldn't you make that the default? A tiny little thing someone might wanna do - if you could move the camera up-down it would let you see some details more easily on some tall models.

BTW: if you wanted to, could you easily extend the viewer to show AitD 2 and 3 assets as well, in case someone wants to look at them later?

NHG: Can we have all of that in one table? Here, I've started one for you. Just alter the formatting if you don't think it's appropriate.

How to do the E6R5 detour. I just fell into the trigger so I'm landing inside the wall. Then jump out of it. You don't actually have to do the jump trick again if you landed sufficiently close to the R6 trigger.

I never noticed this before - when you start a new game and you just wanna skip to the start, I always used to mash enter. Instead, if you first select the character so it starts the intro sequence, then immediately hit space, it instantly cancels it and you get straight in-game.

So please confirm this: if you take the E5R11 entrance to E6, the boss fight works as normal? It might have been me entering E6 twice (the boss fired a fireball). I can't tell if the save file I used had already been to E6 or not. Also why, if the script says you have to visit a room with ID < 6, doesn't it work when you're coming from R0?
Shut up and go back to another earth
LotBlind: I sent the autorization to access the table, didn't know it was necessary now =o

Thanks for the video, I'll time that soon to see if it still worths it even with this little detour
Edit history:
tigrou: 2016-05-29 01:40:48 pm
tigrou: 2016-05-29 01:40:40 pm
tigrou: 2016-05-29 01:37:38 pm
tigrou: 2016-05-29 11:23:54 am
tigrou: 2016-05-29 11:01:43 am
@LotBlind :

- I found by myself how to make Emily land on the floor Smiley Funny enough if she fall from too high she will die. It's interesting to see developers have considered this case (which AFAIK can only happen after a glitch).

- I have added your suggestions to the model viewer.

- AITD2 / AITD3 models : I might be wrong but I think format used in these games is exactly the same as in AITD1 (so it should work). Maybe color palette is different. Unfortunately I don't have those games to give it a try.

- About those VARxx : those are variables which can be used only inside the scripts (there is no way to access them inside the game engine to perform some logic). They store game internal state (eg : is kitchen main door open ? is pot of soup in place ? is cellar door locked ?, ... ).

Initially, they had meanful names (eg : KITCHED_DOOR_STATUS, CELLAR_DOOR_LOCKED, ...) but those names had been replaced by arbitrary numbers during compilation (which transform human readable scripts to bytecode, in order to be executed by the CPU). LifeDISA do the oposite. There is no way to get the original names back (unless developers would send us the original scripts).
So we have to guess what they are supposed to do. It is usually not that hard.

- About C_VARxx : they are similar to VARxx variables (they store game internal state), but difference is they can be read/written by the game engine.
For example when you choose the character (Emily or Edward) the game engine will store 0 or 1 inside CVAR8 (so the scripts knows which character you have chosen).

This is for example used in the living room (E2R10)
IF C_VAR8 == 0
  MESSAGE "A Photograph of Jeremy with his niece!"
  MESSAGE "A Photograph of uncle Jeremy with me!"

Another example : when player is dead, a script will wrote 1 to CVAR15 so the game engine knows player is dead and will redirect you to game menu.
To summary, CVARS are used when there is interactions between scripts and game engine.

Here is a table for the most common CVARS you will find in scripts :
8  = CHOOSEN_PLAYER //Emily or Edward ?
12 = KILLED_SORCERER //aka Pregtz