Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
WIP of 3-1 is here. Now 109 frames faster.
Improvements come from attacking the first enemy, which prevents him from shooting to the right; better zipping glitch, including odd-rule after the zip to get speed to 25; and not being hesistant using the shurikens. Frame-rule is very tight this time. Only lag improvement wouldn't save time, but position and lag improvement would, but it looks quite hard in this stage.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
And here's a WIP of World 6.
It's faster by 2:03.85 minutes (7431 frames). I found some technique which let me attack Meta-Knight each 32 frames, instead of 54. He need to be at the corner, then do a dash attack immidiately followed by an arial attack. I tried to lure him to the right corner, but he didn't cooparate.
Also only 1003140 rerecords, don't laugh, i'm amateur taser :(
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
It's heavily improveable. Without putting any effort into it, I did level 1 219 frames faster.
Improvemts come from extending jump lenght by releasing B on very specific frame after jumping; not losing momentum with spin-jumping.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
marzojr wrote:
This is guaranteed because Bresenham's algorithm minimizes the error between the ideal line and the raster approximation it computes, so you will always be checking the points closest to the ideal line.
So I finally added the algorithm, but it does some weird things. First it gives most of the time a (X, Y) coordinate which has a higher movement error according to the RAM variables than the old method, but the error from the ideal angle is lower; and second it often chooses a (X, Y) coordinate that is so bad that the angle is off by 10% or more. File for reference: User movie #30053802212310494
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
jlun2 wrote:
From BizHawk 1.11.6, I'm not sure why, but sometimes TAStudios Clear/Delete/Clone functions using Control+Insert, Delete, etc, don't work. Restarting the emulator fixes it, but I'm not sure why it stops working.
Go to Edit and press any of the functions, then they'll work again.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
Isotarge wrote:
Spot on, this is exactly what it's intended for. I originally wrote it to make a pretty graph of a moonkick in DK64 but refactored the code so that any game could use it.
Why not use a canvas and display the values as a graph for the last 50 frames, Bizhawk has some canvas lua but I'm not sure what the current status on it is.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
Here's a WIP up to 2-3. So far 50 frames faster. The frame rule was beaten due to beating the boss faster, by adding a sword beam before the first bomb hits him. Also 2-1 has been improved and has a better steam section, but there still might be some lag to reduce, but not to much that it doesn't overflow the frame-rule as well as 2-2 where time was saved by doing a boost after getting bombs.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
The RR counting is actually not working as intended. The current method of "increment rerecords when greenzone state is invalidated" is only working for platforms which have greenzone states on every frame, but with n64 it might often not invalidate a greenzone state.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
niamek wrote:
Getting back to lua.. Need to be reminded of something. I'm not sure where my error is for this.
Where is my newbie mistake? I have a hard time seeing it, but I believe it's the "XY".
Maybe you wanted something like this:
Language: lua
local function position(XPosA, YPosA, XPos2A, YPos2A)
--A: Sprite position low byte
XPos = memory.readbyte(XPosA)
YPos = memory.readbyte(YPosA)
--B: Sprite pos high byte
XPos2 = memory.readbyte(XPosA)
YPos2 = memory.readbyte(YPosA)
--D:Final position
X = XPos + XPos2*256
Y = YPos + YPos2*256
gui.drawText(x-xcam ,y-ycam,"X:"..X.."\n".."Y:"..Y)
end
end
You gave the ram addresses to D, not the values, and you need to assigne the value from readbyte to a variable. Also give your variables meaningful names.
I don't know what it's supposed to do or what your Map positions are, I guessed that they're position and cam related. When you want to make that the drawed text follows the sprite, you need to subtract camera position from object position. And when you want to combine X and Y into one variable you need a table.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
Here's a WIP up to 6-2: User movie #29299530440886933
I needed go back to world 5 to kill an additional enemy for a ~10 frame improvement in 6-1. In the next room the sparky would have attacked, instead of jumping down.
Being 6841 frames faster overall.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
dwangoAC wrote:
How would you differentiate this from the Mega Man explanation from SGDQ? I overall agree with the need for this but I'm concerned that pitching the same thing a second year in a row would get rejected by the submission committee. At the moment my plan is to do a fair bit of this explanation via the Anatomy of an Arbitrary Control Exploit idea, if possible. Thoughts?
I don't mean that you need to show in emulator how it is made, just explaining while a TAS plays back.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
Considering that AGDQ lacked an explanation of what a TAS is and how it's made, I suggest picking up a TAS that is impressive to watch, but doesn't have a lot of tricks to explain, so the presenters can focus on explaining the concepts of TASing.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
You can make a N64 TAS with TAStudio, but it's going to be rough with the current version. I made two levels of Quake64 with TAStudio on a version that was better with the TAStudio controls. The current version has conflicts with analog input controls.
Jogapher: The UDLR is the dpad, not the analog stick. You need to double click into the X-axis or Y-axis and then you can either type in a number or use arrow keys or hold and move mouse. But as I already said it has conflicts and the playback courser often doesn't jump back when the greenzone is invalidated.
Dimon: TAStudio can the used to both make and edit movies, it's up to the TASer how well he can use it.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
ALAKTORN wrote:
It’s unfair that a user can get a rejected movie mark through no faults of his own. How was he supposed to know BizHawk was inaccurate?
I don't think it's unfair that this movie got rejected due to innaccurate emulation, this movie didn't take a long time to make anyway. And having rejected movies is not the end of the day.
With the current build the fastest I got so far is 3 frames.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
feos: Updating branch, then saving immidiately after it seems to do some nasty things. The progress bar goes really fast when saving, I'm not sure if it is the same, as it happened to me on v1.11.3, all input files files in the tasproj gone. Since then I look at the progress bar when saving.
Editor, Experienced Forum User, Published Author, Skilled player
(1529)
Joined: 7/9/2010
Posts: 1322
NitroGenesis wrote:
arandomgameTASer wrote:
I'm voting Yes. This is obviously different enough from RTA play because of how soon input ends, which would justify it having a TAS. Just because it's short doesn't mean much.
How could two runs that both have no input in them be distinguishable?
RTA is timed by gameplay, so they would need to actually play the game to score 10 points as quickly as they can.
