Posts for True


1 2
9 10 11
26 27
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Dimon12321 wrote:
Seems like there is no warp tree and warp tunnel to 8th world on 4-2.
The vines are not implemented at all, and the upper levels that they would reach aren't implemented either. Only two pipe areas are implemented IIRC. So yes, unless some major break is found that means going through half the game...
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
But if it was a licensed game, especially a popular one, would the thought be different? Probably :) Edit: fixed the glaring mistake in 7-4; saves 26 frames. no substantial visual changes other than unnecessary screenscroll speedup wait removed in 7-4. and resync in 8-1. new inputfile is here; please update, thanks Also, for those that don't know / if it isn't clear, this is basically the exact opposite of Somari. WST brought this up but both games are universally bad in their controls and music. SJ6 visuals are better though, not needing to go from a 16-bit to 8-bit helps. (though what help it is ultimately doesn't matter)
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Big oversight in 7-4, can do it 31 frames faster. Don't TAS tired. No big visual difference, but need to redo world 8, so I'll work on this and have this done tonight/tomorrow/day after tomorrow.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
You link to a Fences screenshot. Try Fences? KDE has this type of functionality built-in.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
If you haven't already, when you get some time, open the station up. Most are assembled poorly and are fire hazards. The plastics are also the wrong type and can also catch on fire. Many of the fan-in-handle type of reflow devices are particularly poor quality. Just safety shit, hopefully you don't run into problems but always check the wiring of cheap soldering gear when you get it.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
How is that Lukey/Otherbrand 702 treating you? Hasn't caught fire yet I hope? If you want others to guess change the image filenames. Otherwise have fun with that pilot light transistor.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
I will try to verify whatever, but if there's a spam of input, it very likely won't verify.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Boohoo, everyone keep crying about it. But one legitimate concern I have with the scrolling is that because DPCM bug doesn't seem to be emulated (or at least it doesn't match console), and because of the type of controller handling this game has, this specific movie may be console unverifiable. I'll try later, but will editing out the scrolling still sync?
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
HDDs. OK.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
5-4 requires a mushroom to complete. There is no way to avoid the flames. There are no mushrooms in 5-4 and 5-3 as there are on the NES port. Need to keep working my way back, or writing a script to show what will come out or dumping the maps just in case there is a hidden mushroom somewhere (unlikely). edit: run is done, would like for someone experienced to look at it.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
What is Reset used for? Is it timing-sensitive? This is something as yet untested and could be problematic.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
dwangoAC wrote:
True wrote:
dwangoAC wrote:
We have additionally committed to tithing on any donations we receive
What? There goes the surprise from me. I'm not going to pay for your church.
The word "tithe" unfortunately has an overly religious connotation to it in the English language. To clarify, I mean that we will literally give 10% of what we receive to charities and worthwhile foundations such as the EFF or PCF. On a personal level, yes, my wife and I do give to local churches, but any money we receive from donations here will not be used for that purpose. Does that help?
I still disagree quite severely with the actions (or inactions) of one of those listed. If you need to go somewhere, giving my money away to expressly not do that doesn't make sense. This isn't suffered income. If I wanted those organizations to have my money, they would have it by me directly.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
dwangoAC wrote:
We have additionally committed to tithing on any donations we receive
What? There goes the surprise from me. I'm not going to pay for your church.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
FYI, console verified 100% run. Requires the console RAM to be cleared with the clearing cart (writes FCEUX 00/FF pattern). Will put videos up when I get my visualization boards in and I write the firmware to utilize them.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
My bot is wholly framerate independent. But I need a clocking input to sync the input signal to. As long as there is a vsync or other frame-synced output then it should be doable. My current bot can do at least 16ch of output with an independent clock input, with less than 2usec delay for all 16 channels set from clock interrupt, as long as IO is 5.5V or less. Just need to write code for a generic clocked output driver. However the only cabs I have access to are owned by people who like old fighting games, not tetris games. Most specifically I probably have easiest access to an MVS cab. Emulation needs to be accurate enough for playback to work as expected and I don't think it is with MVS, could be wrong. Regardless it isn't tetris and I don't know enough about arcade to know what tetris runs on.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Can you really please stop doing that?
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Tested live with TheAxeMan. The method he has for FF sync works. Final Fantasy syncs on console, so long as it is an empty cart and sometimes even with console fresh from power off memory (all typically 0) as opposed to the 00/FF pattern in FCEUX. And without an empty cart, the movie will need to be set up to sync state, but if it does that then it will sync (as seen with TheAxeMan's script). So yeah, this is doable. If it won't be live, then once a run is complete, I can do a video of the whole process.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
1) When you find something weird, try everything you can with it. Everything. You might be surprised. 2) Re-check everything. 3) Find out why everything happens. The glitch I use to save ~900 frames is very obvious and refers to #1. You already know what I am talking about, now think of where you can apply that to save some time. The savings in frames in my run refer to #2. I'm not optimized yet because I myself haven't applied #3 to my TAS. Before getting too far I will do this. Here's a script to show you some important data. I need to add more. This game is simple enough that it doesn't need anything more fancy than this, but there is some important data (at minimum Y boundary areas) that should be displayed.
char_x = 0

while true do
	lev_x = memory.read_u16_le(0xFFB6) 		-- not correct

	screen_x = memory.read_u16_le(0xF006) - 0x80 	-- this is first actor, not sanik
	screen_y = memory.read_u16_le(0xF000) - 0x80 	-- this is first actor, not sonique

	jumptimer = memory.readbyte(0x7150)

	-- if memory.readbyte(0x6D95) == 255 and memory.readbyte(0x7151) == 255 then
	if jumptimer == 255 then
		djump_ok = 'Y'
		gui.drawBox(screen_x, screen_y + 31, screen_x + 20, screen_y + 32, white, white)
	else
		djump_ok = 'N'
	end

	gui.text(120,  0, "DJumpOK? " .. djump_ok)    -- NOTE: not always reliable, especially near the floor
	gui.text(120, 12, "Screen-X " .. screen_x)

	gui.text(256,  0, "JmpTimer " .. jumptimer)
	gui.text(256, 12, "Screen-Y " .. screen_y)

	gui.text(392,  0, "Lev-X  " .. lev_x)
	gui.text(392, 12, "Spd/px " .. memory.readbyte(0x00E4))

	gui.text(424, 40, (memory.readbyte(0x97) * 256) + memory.readbyte(0x96))

	if (lev_x ~= 0xffff) and (lev_x - char_x) <= 3 then
		real_speed = lev_x - char_x
		char_x = lev_x
	elseif screen_x - char_x <= 3 then
		real_speed = screen_x - char_x
		char_x = screen_x
	else
		real_speed = 0
		char_x = lev_x
	end

	-- show actual speed
   gui.drawText(screen_x + 7, screen_y - 15, real_speed, 0xB0000000)
	gui.drawText(screen_x + 6, screen_y - 16, real_speed)

	emu.frameadvance()
end
Edit: 4-1 is 873 frames faster. Not including the skip it is 7 frames faster. Approaching 2K frames saved out of about 3.6K frames. Can maybe be improved, but I don't think I have any hitch points in the recording, will have to check carefully. Is anyone else other than Kurabupengin interested in a WIP of this shitty game?
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
World 1-1: Saved at least 4 frames in the level itself, possibly more except for the skip execution. So I am currently 909 frames ahead of Kurabupengin after just one level. Should I continue? Edit: World 1-2 is 28 frames faster; total difference is now 937 frames. Not even trying... Also, I believe most of the lag in this game is simply input lag and the game renders every frame. Unless there is something with this emulator I don't know about... The biggest contributor to this lag is coins, and only when there are large numbers will it actually slow the game down IIRC - I could be wrong. A far bigger problem are the transition points, probably every 256 pixels or so but I will need to check again, where if you are in the air going through them you will slow down to speed 1 for a frame. Need to figure this out and put this in my display script. Edit 2: Another source of the lag is the bass notes in the music. But again it seems to just be input lag, the game continues to render without any lagged render frames.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Because of time constraints the one on the site likely wouldn't be there; a modified run without hard/soft resets is being worked on by TheAxeMan it seems. We're trying to get it to run. The concept itself really shows how far TAS can go though. Using a script to generate input based on a battle in the game in order to get it to sync with console. Pretty cool...
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
I've been waiting for the VDP glitch to be fixed to run this - otherwise the Lakitu spawns infinite spinys causing a lockup. Why do you keep posting this? This run is very suboptimal. You are some seconds behind after just the first stage, and the second stage can also be done faster. I have work to do tomorrow it looks like...
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
jlun2 wrote:
Also for the FDS, what exactly is stopping it from verification?
Disk access time, and not having one to know if it continues polling the controller during disk reads (may depend on the game). If I had one, I could work to making it verifiable, as for some games it very well may be.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Corrodias wrote:
I have searched the site and the forum and I cannot find information related to verifying Super Metroid runs. The "Game End Glitch" run was recorded in lsnes. What sort of console modification would be required to make it work? dwangoAC did not seem concerned about any difficulty in getting the bot the ability to hard reset the console. I did not search the entirety of the 250+ page general Super Metroid thread, as I see no function for searching within a thread on this forum.
SNES has two crystals, clocking the CPU and APU repsectively. There is likely either code that tries to access the bus at the same time but the bus is busy, or timing-critical interrupts between the two. The clocks have been tested on emulator down to 1ppm or so and the run still ends up desyncing. The issue isn't the absolute clock speed, but rather the clocks always running at the exact same ratio, in this case the "ideal" ratio as in the emulator. The modification will require either injecting two accurate phase-locked clocks or some other method. dwangoAC said he is sending me a less than ideal SNES so I will modify this for clock injection and see if I get better results.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
You could check the threads or the console verification progress page for some of these answers. River City Random does not sync. The enemies perform different actions on console than on emulator. International Super Star Soccer is untested. Super Metroid does not sync and will not sync without modifying the console.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
Editor, Experienced Forum User, Published Author, Player (68)
Joined: 1/18/2008
Posts: 663
Yeah, well people said there's no way in hell Rockman 2 would sync. Or Mega Man 3. Or Super Mario World executing code input by controllers. Or... The result appears to be deterministic. Emulation accuracy needs to improve.
true on twitch - lsnes windows builds 20230425 - the date this site is buried
1 2
9 10 11
26 27