Indivisible on NES

It's been a while. I demade the Indivisible Prototype by Lab Zero Games on NES.


Get it here: https://kasumi.itch.io/indivisible
If you're unfamiliar with the original game, this tweet shows a side by side: https://twitter.com/KasumiDirect/status ... 2007395339
And this shows off some of the animation in the game: https://twitter.com/KasumiDirect/status ... 1388644353

I will admit to sitting on post 99X for a while so this could be post 1000 >_> Anyway, enjoy! It's harder than I'd like, but I believe in you all!

This is great! Looks excellent. And, as someone who's made a demake before, I know how hard it can be to capture an original, but you've done really well!

Use of color and shape in those backgrounds is gorgeous.

Wow, I've been wondering what you were working on (if you posted about it, I missed it). I've never seen the original, and have only played up to the first boss so far. But this is awesome.

Congrats on finishing it!

My first normal run of this version had a clear time of 14:24.20

(The jump indicator for the boss helps a lot!)

Holy carp.

It looks amazing, it plays well and your devlog covered tons of little details I never even thought about. Thanks for posting, and great work!

This looks absolutely, stunningly beautiful! :O WELL DONE!

I very much like that you went with large meta-sprites despite the per scanline limit. It helps the visual presentation having its own identity.

The 1st boss fire attack seems more aimed at people who've played it/tested this game quite a lot. Switching to bare hands helped a bit with timing evasions, but i've yet not beat it. Maybe i'm just not good enough. Are there more bosses after this one?


EDIT: Ok, beat it at last! 36 minutes, including the pause to write the previous. I learned the timings using bare hands, then translated that into getting more careful/conservartive using the axe.


EDIT 2: I noticed you get to start with the axe on the second round!

Thanks all!

Memblers: I didn't post because I didn't think I'd work on it so long, and the early stuff was not that interesting to this audience. (People who see NES stuff often.)

There is a devlog: http://skullgirls.com/forums/index.php? ... eavy.8488/

My favorite post is this one about button input.

I totally regret not posting updates in more places once I realized I was in it for the long haul.

rainwarrior: You're the first person to publicly post a time (that I know about)!

Rahsennor: There's so much more to write about! In theory. I have post drafts and over 300 gifs recorded over the course of development. Most didn't make it to that thread. Here's one gem: God Mode (Light Spoilers, don't click until you've played!) (You can't actually unlock this, it's just a silly thing I did during development because it was easy.)

But there's some things I need to take care of before I really can write much more. Just getting this done was a struggle with current going-ons.

Thanks! This game even handles >64 sprites, but there's only one way to really see that happen and it ain't the vines. I sort of regret the graphical style, but that's a thing I'll probably write a whole giant post on.

Other people, definitely don't use barehands because then that fight takes like twice as long. (Axe does more damage). Glad you beat it, though!

"Just like the original game." is probably the best answer to that question.

Yep. And if you beat it again that way you'll get different clear text showing it.

I obviously didn't do the battle system, but this game took some pretty great pains to be close to the original in most other ways.

Extremely solid work! I happened to play the Indivisible prototype on PS4 not a long ago, so it was fun to compare the two.

(One minor difference I noticed was that if you die in the boss, the boss music keeps playing. I think having silence or some mellow music (as in the original) before the boss appears would be better.)

I was able to try it a little bit and it looks great! I will try to see how far it goes once I find more time.

Good work!

This is amazing. Congrats!

Well, maybe it's not the most meaningful time either, ha ha. Basically I was just saying it took me 15 minutes to beat that boss, even though it's slightly easier and I had a bunch of practice a few months back.

Trying not to say anything too spoilery but I managed to do the "other" thing on my first try this time (not sure if it changed), though it took a lot of savestates to get there.

thefox: Good Ear! This is for slightly selfish reasons. The boss fight is shorter than the song, that allows it to loop. I have less of an excuse for the music difference AFTER you kill him, though.


Unchanged.

If you choose to write about it, I'm going to read the hell out of it.

I can't beat the boss, by the way. Not enough health to the endure the unavoidable fire attacks any further than the two Belus.

Wow a new NES game in 2017?!?! Congratulations on making it.

BTW, how long does it take to beat the boss??
Whoops! Confirmed boss is beatable

Qualifier: unavoidable with my impaired reflexes. Obviously everyone else has no problem with it.

Nah, Rahsennor, you're definitely not alone. You'll probably find yourself in good company. The boss is one of those things for a long devlog post.

I was totally astonished. Magnificent mastery of physics.

Sprite animations looks great! Gameplay is amazing.

Didn't have time to play it properly yet, but congratulations on the great looking game! I particularly like the complex fading animations! Few NES games bothered changing the hue to add more steps to fading sequences.

Thanks for more kind words!

Tokumaru: I figured you'd notice the fades! It's one of many things this game steals from my secret advanced slope game. It's a really simple algorithm that's something like <= $X7 move toward $X1, >= $X8 move toward $XC. Once at $X1 or $XC, subtract $10 and make black on carry clear.

But as simple as the algorithm is, the code for it is HUGE and a mess, I'd like to simplify it sometime.

The thing I didn't steal was my other game's level format, this game is basically uncompressed. I wanted to be lazy with RAM, the good format uses a whole page for it.

Maybe in a couple of weeks I'll upload a video of a runthrough that shows lots of stuff off. (As well as getting a fast, fast time.) I've got a good recording already, but it has two mistakes that keep it from a time milestone (Sub X:00.00) in my favorite category. I want to knock the two seconds off. You'll have to wait to see what X is.

But of course... the game can be beaten even faster than that in other categories.

edit: see note below

Larger than 64 bytes? Sounds like a job for a lookup table.
01 0F 01 02 03 04 05 06 09 0a 0b 0c 0F 0c 0f 0f
11 01 11 12 13 14 15 16 19 1a 1b 1c 0c 1c 0f 0f
21 11 21 22 23 24 25 26 29 2a 2b 2c 1c 2c 0f 0f
31 21 31 32 33 34 35 36 39 3a 3b 3c 2c 3c 0f 0f
Which isn't the same thing as is being done, so…oh well. Maybe I can use it later.

The code I could avoid by using a lookup table for fade to black seems to win over a 64 byte lookup table. It's also not quite as simple as beating 64, because apparently the hue shift and value shift are separate steps. I was wrong, value is first, any color that reaches $0X doesn't become black until the hue shift step. You can see this with $0C on the title screen. (It stays $0C until everything else is $0X.)

So you'd need two tables for a purely table based approach, but one could be 16 bytes since after value the high nibble is guaranteed zero.

Hybrid's probably best, the subtract $10 and do a thing on carry is easy, and then you only need the 16 byte table. Looks like it'd beat what I have by about 8 bytes.

Fade in is a tougher call, since the destination color is not always the same it's trickier to find the code I could certainly avoid using lookup tables. I'll say had I used tables the fade in would probably look better because I'd have iterated more. I'll probably play with the algorithm for fade in for future games.

Well, speed isn't super-important, you're not updating much else, so yeah, hybrid with 16-byte table and ANDing off the value nyb.

(…thought you said it hue-shifted, going to hue 1/C, *then* value-shifted to 0 and then black?)

I said it was something like that, yes, hence the "I was wrong".

No worries! My table is for hue-then-value (though all in the same process).

Spoilers: This video reveals one new thing, but not everything: https://www.youtube.com/watch?v=p86jeacic9Y

You can whack the floor attack back at the boss??

Finally got a chance to play this. Wow, well done. I'm impressed.

rainwarrior: I expected more wouldn't find it than would, but it was a fun idea I had and it was easy to do. It definitely makes the fight faster, whether it makes it easier is up for debate.

Thanks, gauauu!

I mean to post one of these much sooner. Here's my own speedrun of the game, showing a secret. There's still one secret yet to be found, which I'm still debating if I want to spoil. (It's not that exciting.)

In any case, here's the video: https://www.youtube.com/watch?v=H9JIZmJAiXA (The death is on purpose)

So this thing is a year old. I wrote a post about it here: http://skullheart.com/index.php?threads ... ost-408594

But, that's not why I'm bumping this. Can anyone try this on a PAL NES? Maybe from an infiniteneslives cart and not a powerpak?

Does the "axe text" have any visual errors?

I've always been curious, but still haven't seen any evidence of anyone trying it. If there aren't errors, it'd be kinda cool because I actually don't do anything different between NTSC and PAL for that split.

But I only ever checked on Mesen.

I just tested it and no problem here !

edit: tested with an Everdrive.

HAH AWESOME! Thanks. Then I'm definitely doing the same thing for my next game.