Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Warp wrote:
The two famous postulates of special relativity are that the laws of physics are the same for all inertial frames of reference, and the speed of light in vacuum is the same for all inertial frames of reference.
Do we know that a frame of reference is inertial by observing that there's not a single body moving at speeds higher than c? For instance, the Earth is not an inertial system by the simple observation that, relative to it, the stars (except the Sun) travel at a much higher speed.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Take a look at memory.settrace(filename)If filename is non-empty, output trace to that file. If "", stop tracing.
For some reason it's not in the pdf.
You can also tracelog to a file with the internal command
tracelog:
<cpuid> <file>: Start tracing <cpuid> to <file>
<cpuid>: End tracing <cpuid>
Internal commands can be manually typed in the message window and invoked from Lua using exec(string command).
I hope that helps. If not, the best place to ask is at #lsnes IRC.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Anyone who speedruns and has a WR (or even top 10) thinks so? Any experienced TASer would say that TASing has nothing to do with skills (not reflexes)?
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
A new glitch (more of a technique, really) allowed 2 new exits in the smallonly TAS: Cheese Bridge Area and Soda Lake (that was unreachable before).
Link to video
The technique relies on those facts:
> sprites are spawned when the camera reaches a certain range of 16 pixels;
> the spawn routine happens every other frame;
> pausing the game allows a person to add one more frame of inactivity;
> the camera can travel 2 pixels per frame faster if Mario was facing the opposite side and then turn around the direction he's traveling. This might happen even he's very slow.
> the rope can boost Mario. Without any boost, this tech is impossible with speeds like 49 or 51.
Therefore, in order to skip a certain range, one can try to make the camera follow this pattern:
Frame 0 (active) | Camera is 15 mod 16 (just before the range)
Frame 1 (inactive) | Camera is 06 mod 16 (game is paused)
Frame 2n + 1 (inactive) | Camera is 12 mod 16 (game is unpaused at appropriate frame parity)
Frame 2n + 2 (active) | Camera is >= 0 mod 16 (just after the range)
This whole combination makes it extremely unlikely to ever happen on RTA, but who knows?
Remark 1:
Later, I found out that it can happen without pausing with those platforms, because they boost mario very quickly.
Remark 2:
Sadly, the same cannot happen in vertical levels, because the vertical camera has the maximum speed of 5 pixels/frame. So 15 -> 4 -> 9 -> 14 is the best we can do, only 2 pixels away...
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Bobo the King wrote:
No joke, I was thinking of doing the same thing. Then you actually went there.
(By the way, is there any reason that this polynomial evaluates to an integer when x=19? Is that true in general? It seems like a wild coincidence.)
I just added the value that I wanted as the 19th term and incremented the degree of the polynomial.
function readAsAscii(address, length, domain)
local q = memory.readbyterange(address, length, domain) -- what if address == 0?
table.move(q, 0, #q, 1, q) -- this shifts the elements of q to be 1-based, not sure about performance
q[0] = nil
for index, byte in ipairs(q) do
q[index] = string.char(byte)
end
return table.concat(q)
end
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
partyboy1a wrote:
Is this game compatible with lsnes or Bizhawk?
Very late to the party...
I have verified this run on lsnes without any issue.
It was believed that this run would crash on the final boss, but it doesn't. Speedrunners even created version 3.2 because of this issue...
Anyway, a new run that was started (not by me) long ago can go on using v. 3.1...
http://tasvideos.org/userfiles/info/44570109917081025
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
A TAS is supposed to be good. Yours doesn't even try to beat the human records... from 20 years ago.
Edit: I think you misunderstood everything about TAS and this site. Please, read this first and, in doubt, ask in the portuguese area.
http://tasvideos.org/PT/WelcomeToTASVideos.html
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Link to video
This game is officially broken. This time, I used the stun glitch to beat the level prematurely. The last room, for some reason, has sprite buoyancy enabled. That means that the fish will swim and activate the stun timer if he gets in the water. However there's no water in the room, but the offscreen area acts as water. The setup is similar to YI4 in the SMW any%.
P-Switch -> Horizontal Fish -> Reznor (which ends the level)
Movie:http://tasvideos.org/userfiles/info/44187441227331633Past recordsDownload Difficulty: ★
5:44.28 by Amaraticando [original glitched video with description]
7:06.10 by Dawn[Regnum0nline] (smv)
7:12 by PangaeaPanga [YouTube strike]
7:15 by TheFinalBoss726 and Dawn[Regnum0nline] (smv)
7:35 by PangaeaPanga [YouTube strike]
7:41 by TheFinalBoss726
8:05 by xKingBulletBillx
8:14 by ISM
8:46 by xKingBulletBillx
10:34 by dainegai
PS: in the old records, the timing ends when the credits scene appears.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Tremane wrote:
I saw some SMW:Hacks here already "The legend continues" or "Kaizo"
NOT A JUDGE, BUT: Notice that the hack's popularity is not a determinant factor. Kaizo Mario World is by far the most famous hack and the any% was canceled here some years ago, while the 100% wasn't even submitted. KMW3 was accepted here with good reception, because the hacks is not ugly as hell like its precursors and the TAS itself is full of WTF moments. J.U.M.P. is a very good hack that wasn't accepted, so look at those runs to see if the VIP run will be similar and likely to be rejected.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Edited, adelikat. My final suggestion is: try both, pick the one that gives you better productivity. We ultimately care for good (and reliable) speedruns, the software that you use is secondary.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Danfun64 wrote:
Umm... my question was never really answered...
The answer depends on what better is. If you wanna support for many controllers, different kinds of controllers (mouse, multitap, etc), support for subframes/hard resets, if you want to produce a console verification, or portability, then go with lsnes.
If you wanna the favorite movie format of this site, the most maintained emulator, the fancy tools like TAStudio, or don't mind the small probability of your movie not working in version 2.5, then choose BizHawk.
Editor, Experienced Forum User, Published Author, Player
(158)
Joined: 1/10/2012
Posts: 673
Location: Brazil
Telephia wrote:
I'm having trouble with lua. I'm trying to get a picture to display and i keep getting attempt to call field 'drawImage' (a nil value). Getting the same kind of error whether I use forms or lua canvas and I can't figure out why.
Remark 2:
Sadly, the same cannot happen in vertical levels, because the vertical camera has the maximum speed of 5 pixels/frame. So 15 -> 4 -> 9 -> 14 is the best we can do, only 2 pixels away...
