>Couldn't you get the orginal source code and replace sprites with a blue coloured bitmap.
You can replace all the sprites with bitmaps of course, but the sprite is not always the same size as the hitbox.
Something else: A feature to reduce the amount of work for luck manipulation and lag removal would be nice... i have no idea how that would work though.
You can replace all the sprites with bitmaps of course, but the sprite is not always the same size as the hitbox.
I wonder how often this is the case.. NES hardware has sprite collisions built in, so I imagine cases where hitboxes and sprites aren't equal are rare.
someone is out there who will like you. take off your mask so they can find you faster.
I support the new Nekketsu Kouha Kunio-kun.
NES hardware has sprite collisions built in, so I imagine cases where hitboxes and sprites aren't equal are rare.
NES hardware only has sprite-0 vertical scanline hit built-in. It is mostly used for split screen/parallax/palette effects, by triggering an IRQ when a certain scanline is being rendered.
Inter-sprite collisions, and horizontal collision checks of all sorts, must be done with software.
It also has means to indicate whether 8 sprites on the same scanline were rendered, but barring special cases, it cannot be used for collision checking.
The lack of hardware sprite collision checks is precisely the reason why collision checks bear only a superficial resemblance to the onscreen object (either the hero can safely hug the enemies, as in Solomon's Key, or the hero can die from merely smelling the enemy, as in Astyanax).
In Rockman 1 and 2, the collision check is done using bounding boxes stored in the game ROM for each different pose of each of the objects.
A screen that tells you the fastest way to beat the game. It would take all the fun out of making one, but it would be pretty amazing to watch. If you didn't tell anyone how you did it, they would all marvel at your skills...
Joined: 8/9/2004
Posts: 139
Location: Washington State
Chamale wrote:
A screen that tells you the fastest way to beat the game. It would take all the fun out of making one, but it would be pretty amazing to watch. If you didn't tell anyone how you did it, they would all marvel at your skills...
As in watching someones previous TAS on a secondary monitor?
I'm not sure I understand.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
There is a very easy way to get around this, with the tools we already have...
Frame Advance.
Just get close to a sprite, choose one frame that you want to go around it, and if you get hit, try one frame earlier. and again, keep going down/up one frame until you find the best possible frame to go around something without getting hit.
As in watching someones previous TAS on a secondary monitor?
He said it tells you the "fastest way to beat the game". Most TAS's aren't the fastest way, any new discovery will prove a given one not to be. He meant... some magic crystal ball of TASing.
comicalflop wrote:
Just get close to a sprite, choose one frame that you want to go around it, and if you get hit, try one frame earlier. and again, keep going down/up one frame until you find the best possible frame to go around something without getting hit.
I think we all know that, but they're hoping for an easier way of doing it .
Joined: 8/26/2006
Posts: 1139
Location: United Kingdom
> As in watching someones previous TAS on a secondary monitor?
Although it was a misunderstanding the thought of watching a new run whilst also comparing it to the movie it obsoleted in AVI format would be cool, if that's what you meant, unless i have misunderstood.
I wonder how many movies are actually perfect. There can't be more that 6 or 7. Although there is a lot that can't be improved by more than two seconds. Please don't mention the SMB1 movie as being perfect, then again I doubt the small number of perfect runs to watch are actually the best to watch.
Joined: 6/13/2006
Posts: 3300
Location: Massachussetts, USA
I watched Sami's any% run in a second emulator as I played my run, in order to learn how to TAS DKC2.
KNowing the fastest way to TAS is a game, I agree there is no way to do that.
It's not too hard to use Frame Advance to do what I describe- it only took me anywhere from 3-15 rerecords to dodge around by one frame one sprite/monster/wall.
Joined: 5/22/2006
Posts: 58
Location: Denver, CO USA
Saturn wrote:
Regarding Snes9x:
1. Sound channels enabling/disabling
2. Save/Load of the memory watches
3. Allowance to select 2 games in the same time, seting the starting frame for both seperately and dual encode them side by side for comparisons
4. Perfectly working Net-Play support (like ZSNES already have)
Thats all I could think of currently.
regarding # 2, i think there should be a centralized database of pertinent memory addresses for each game, kind of like cddb for CD track information with music. each time you load a rom, the emulator checks the database for new memory addresses and downloads them with a description to your local machine...just a thought
What would Mr Belvedere do? Probably eat some butter.
Joined: 8/26/2006
Posts: 1139
Location: United Kingdom
AKA wrote:
I wonder how many movies are actually perfect. There can't be more that 6 or 7. Although there is a lot that can't be improved by more than two seconds. Please don't mention the SMB1 movie as being perfect, then again I doubt the small number of perfect runs to watch are actually the best to watch.
I guess we'll never know, or could we? Is there any way we can really sit back and say, "there is not one single frame in that movie that can be improved, it is perfect."
Barring any new glitches discovered, I think there are a few runs we could assume to be 100% perfect.
There are sure some runs that look absolutely perfect already, but unless you TAS them yourself very careful and optimized you can't really judge them correctly.
The only run I can say for sure that is technically 100% (or at least 99,99%) perfect is Genistos Chip 'n Dale - Rescue Rangers 1-player TAS. I would be extremely surprised if somebody could improve it by more than 10 frames during the entire run. With this run (which obsoleted my - already considered perfect run) Genisto really surprised me and showed that he is a true and absolutely deserved "Ninja".
I also consider my Chrono Trigger "New Game+" TAS to be absolutely frame perfect. I really doubt one can save more than 5 frames during the entire TAS. Only possible but minimal frame saver here would be by entering one letter in each of Cronos and Marles name to save time during the few text dialogues which I intentionally skipped though, to not destroy the names.
Emulator Coder, Site Developer, Site Owner, Expert player
(3573)
Joined: 11/3/2004
Posts: 4754
Location: Tennessee
I believe these to be absolutely frame-perfect. I don't believe even a 1 frame improvement is possible (notice the word "believe").
Smb1 warped
smb3 warped
donkey kong
circus charlie
mike tyson's punchout
In addition, I think there are several others that are "near" frame perfect that probably couldn't be improved by more than 10 frames.
Emulator Coder, Site Developer, Site Owner, Expert player
(3573)
Joined: 11/3/2004
Posts: 4754
Location: Tennessee
yeah, what JXQ said.
and as i recall in with smb, phil made on 3 frames quicker. It was cancelled and then pom made a new run with the imporvements.
Since then, I know of no improvements to smb.
I'm sure Pom admitted that his run was unoptimized in places. It kind of makes me want to do it just to prove to everyone that its not perfect. Unless certain delays were caused to avoid transitional lag.
Well you announce on the forums that this run or that run is improvable, so I thought it would be appropriate to bring up that trend on the forums as well.
<Swordless> Go hug a tree, you vegetarian (I bet you really are one)
