I always expect old games to have stood the test of time and performance-wise I guess be perfect for what they go for at least. (No idea why to be honest) Even though they were made with short deadlines with possibly not the best tools... But even then agreed it just hurts to know the game doesn't even wait x frames or keep count but instead the cpu is taking that long and then does all that extra work ;\
Ahh! I definitely needed a lot more proofreading on this one. I completely forgot there *was* a second boulder. I also forgot about the first pipe in world 5 which you can save a frame on with 486 cycles to spare. That's pretty important information... I might just fix all that and upload again.
@@100thCoin I'm also curious about W8, where you mentioned that no frame could be saved. Though we first discovered this in 8-Navy, that we got an extra lag frame there somehow due to the amount of points we had. Was that something separate compared to what you have talked about in this video?
@@TompaA I briefly mentioned that exiting "non-standard" levels, such as Hammer Bros, or in this case the Navy have several other factors that make it annoying to determine if you get lag or not. The NMI before the level exits includes seeding RNG and running the sound engine. I didn't do too much experimenting since I figured it would be way too much to explain, though I only tested it on a few hammer bros and it lost only around 600 extra cycles. I know world 8 loads slightly differently to check for the dark room, though I don't know where the extra few thousand cycles are being lost, (or maybe even saved?) since you're certainly not going to lose that much time from your score. I just tested in-game, and I have no idea how you could create/save an extra lag frame exiting the navy. Perhaps I need more research on this.
I made a custom emulator that I use for a lot of visuals throughout this video (and a few others). For that image specifically, instead of rendering the screen how it should look, I simply color the pixel by what instruction is happening during those ppu cycles.
Its always fascinating to see the inner workings of a game, and the consequences of developer decisions that nobody ever would've really noticed. Like lag being common because Sprites are cleared twice. That could save like. 10 whole frames! xD Also, you might consider linking Retro Game Mechanics video on lag for the SNES for anyone confused about NMI and such. Explains all you need to know for this to make sense.
it's criminal just how underrated you are. you're the ZZAZZ of Mario 3.
return of the king
This race against the CRT beam is much more intense!
Amazing work has always. I'm always blown away with the amount of effort you put in these videos.
The takeaway from this is that I'm kinda pissed off at the fact that sprites are cleared twice.
The nametable is cleared twice when loading the map. That's so much wasted time!
@@100thCoin aaaaaaaaahhhhhhhhhhhh
I always expect old games to have stood the test of time and performance-wise I guess be perfect for what they go for at least. (No idea why to be honest)
Even though they were made with short deadlines with possibly not the best tools... But even then agreed it just hurts to know the game doesn't even wait x frames or keep count but instead the cpu is taking that long and then does all that extra work ;\
that footnote at 9:24 cracked me up. it's funny that MOVING A SUBROUTINE could cause an extra frame of lag
Worth the wait
I sure hope so.
Love it! =D
One note about the W4 map though: Seems like you forgot to add the cycles time for the second rock?
Ahh! I definitely needed a lot more proofreading on this one. I completely forgot there *was* a second boulder. I also forgot about the first pipe in world 5 which you can save a frame on with 486 cycles to spare. That's pretty important information... I might just fix all that and upload again.
@@100thCoin I'm also curious about W8, where you mentioned that no frame could be saved. Though we first discovered this in 8-Navy, that we got an extra lag frame there somehow due to the amount of points we had.
Was that something separate compared to what you have talked about in this video?
@@TompaA I briefly mentioned that exiting "non-standard" levels, such as Hammer Bros, or in this case the Navy have several other factors that make it annoying to determine if you get lag or not. The NMI before the level exits includes seeding RNG and running the sound engine. I didn't do too much experimenting since I figured it would be way too much to explain, though I only tested it on a few hammer bros and it lost only around 600 extra cycles. I know world 8 loads slightly differently to check for the dark room, though I don't know where the extra few thousand cycles are being lost, (or maybe even saved?) since you're certainly not going to lose that much time from your score. I just tested in-game, and I have no idea how you could create/save an extra lag frame exiting the navy. Perhaps I need more research on this.
0:53
How did you make that image???
I made a custom emulator that I use for a lot of visuals throughout this video (and a few others). For that image specifically, instead of rendering the screen how it should look, I simply color the pixel by what instruction is happening during those ppu cycles.
So, weirdly, speedrunners care about SM3 scores.
And are probably the only ones to, ever.
I became more confused with every paragraph, but cool video nonetheless.
Very good videos!
Its always fascinating to see the inner workings of a game, and the consequences of developer decisions that nobody ever would've really noticed.
Like lag being common because Sprites are cleared twice. That could save like. 10 whole frames! xD
Also, you might consider linking Retro Game Mechanics video on lag for the SNES for anyone confused about NMI and such. Explains all you need to know for this to make sense.
Having this issue emulating the game, any suggestions?
have you tried fceux or mesen
do anyone know why crouch make racoon break blocks backwards?