Author
|
Topic: Screenshots (Read 67874 times)
|
vivify93
Guest
|
 |
« Reply #585 on: July 03, 2008, 12:08:47 am » |
|
Er, I stupidly didn't insert a quote in that post without reading the rest of this topic. My post, in which I was incorrect anyway, should be gone by now. It was directed to Dragonsbrethren, by the way.
With regards to the other game, yes, the font looks fine.
|
|
|
|
|
Googie
Guest
|
|
« Reply #586 on: July 06, 2008, 08:23:53 pm » |
|
Okay, here's two screens of my SMB3 hack "Toad's Trauma Show"...   I'll have a video up soon, now to edit the airship level... ^^' |
|
|
|
|
|
|
Rockman
Guest
|
|
« Reply #588 on: July 15, 2008, 02:23:28 pm » |
|
I like the graphics.  |
|
|
|
|
Googie
Guest
|
|
« Reply #589 on: July 15, 2008, 09:00:26 pm » |
|
I'm wondering what're the gems about? Good to see that you took out the timer though... :thumbsup:
|
|
|
|
|
Darkdata
Guest
|
|
« Reply #590 on: July 16, 2008, 12:05:54 am » |
|
I'm wondering what're the gems about? Good to see that you took out the timer though... :thumbsup:
Look again, I just moved some things around.  |
|
|
|
|
Googie
Guest
|
|
« Reply #591 on: July 16, 2008, 07:36:02 am » |
|
I'm wondering what're the gems about? Good to see that you took out the timer though... :thumbsup:
Look again, I just moved some things around. Eh, it's not my fault I was sleepy while typing lol!  |
|
|
|
|
|
|
GenoBlast
Guest
|
|
« Reply #593 on: July 19, 2008, 08:28:42 pm » |
|
Nice screens, Darkdata.
Does the title have something to do with the Opera browser sucking at playing flash? <_<
|
|
|
|
|
Darkdata
Guest
|
|
« Reply #594 on: July 20, 2008, 10:39:27 am » |
|
Nice screens, Darkdata.
Does the title have something to do with the Opera browser sucking at playing flash? <_<
Nope: Story is basically "Trapped in a crappy flash game, can you escape?" Edit:  |
|
|
|
« Last Edit: July 20, 2008, 11:35:13 pm by Darkdata »
|
|
|
|
GenoBlast
Guest
|
|
« Reply #595 on: July 24, 2008, 12:44:00 am » |
|
It's such a pain in the ass to get FEIDIAN to work, which is ironic considering what it stands for. But, yay, Japanese fonts!  I'm going to attempt to translate the Japanese Super Mario RPG's menus. (by "translate" I mean copy the menus from the US version) Hopefully it won't require too much ASM hacking. (by "too much" I mean any ) |
|
|
|
|
Chester
Guest
|
|
« Reply #596 on: August 05, 2008, 11:00:20 am » |
|
I have restarted the work on my Bahamut Lagoon's hacking/translation project just a few days ago. Some of you will probably remember what I had already done and what I was going to do. For those that do not remember: the most noticeable hack was the 6-bit encoding that allowed to use 10 characters instead of 8 for the names, making possible to set as defaults names like "Ice Dragon", more adherent to the original (Japanese) ones. I chose that method because the names are tightly packed into RAM, loaded/saved directly into SRAM, where there is no free space safely available. Byuu pointed out that "Thunderhawk" needs 11 characters, and suggested a way to improve even more the "compression": a base-53 encoding, which involves a large use of 64-bit arithmetic operations, not natively supported by the SNES which, as you know, has just a 16-bit accumulator and some registers for 8 and 16 bit multiplication and division. So, it was possible, but "definitely not going to be easy", and there was the possibility that, even if implemented, it would have been slow, or not quick enough. In these days, I have coded a library that provides full support for 64-bit operations on the SNES. I plan to make the source available, but I have to go back over it, firstly. What is important, it is that I have managed to translate Byuu's algorithm in 65c816 assembly; Afterward, applying it to the game has been a joke:   And without any noticeable delay. |
|
|
|
« Last Edit: August 05, 2008, 11:57:13 am by Chester »
|
|
|
|
byuu
Guest
|
|
« Reply #597 on: August 05, 2008, 11:48:46 am » |
|
Wow, you actually pulled off the base-53 encoding on the SNES?
I wrote a reference 64-bit mul/div lib in C, and it was pretty painful even there.
Multiplication is trivial. You can add and shift over as many bytes as you want, so you just do:
n += char; //char = 0-55
n = (n << 5) + (n << 4) + (n << 3); //n*32 + n*16 + n*8 == n*56
For division, 56 is the most favorable number that fits 11-chars into 8-bytes, so you can break that back down to n = n / 8 + n / 7. Division by 8 is easy, n >> 3. But division by seven ... division isn't like multiplication, you can't chain them together.
Eg, n = (n >> 3) +/- (n >> 2) +/- (n >> 1) +/- n; will not work.
But what you can do is to keep subtracting by 56^11 until your value is too low for another and store that value. Do the same thing all the way down to 0. Then multiply all your values by their places and add the sum together. Now take your original value and subtract your new value to get the result of division.
But that is absolutely painfully slow.
The other idea is to take advantage of the fact that:
n/7 - n/8 = n/56. But you still can't divide by 7, so you still need some really intense processing to pull it off using subtraction loops.
So, how'd you get around the division issue? I was thinking you'd be best to just use base-64 anyway, and limiting character names to 9, and dragon names to 11. Eg use some of the player space to store bits of the dragon names.
|
|
|
|
« Last Edit: August 05, 2008, 11:57:25 am by byuu »
|
|
|
|
Chester
Guest
|
|
« Reply #598 on: August 05, 2008, 12:26:06 pm » |
|
Yes, I am definitely daring to the information entropy, and to the SNES' processor as well I am using SNES' registers to do 8/16-bit multiplications and divisions. Think about it in this way (I'm simplifying it a lot due to lack of time): In the multiplication, take the first byte of the 64-bit integer, and multiply it by 55, using the 0x4202 for the multiplicand, 0x4203 for the multiplier, then get the result from 0x4216/17. Store it in $7f0000/01 (for example), then do the same thing with the next byte; instead of storing the product, add it to $7f0001/02... and so on until the end of the number. The division is a lot more difficult. Basically, I use 0x4204/05 for the dividend and 0x4206 for the divisor, getting the results from 0x4214/15 (quotient) and 0x4216/17 (remainder). But I also have to handle the remainders of each division, taking account of them together with the next byte of the 64-bit number. Excuse me if I have not made the point clear, and if I have made some (technical or grammatical) mistake, but in this moment I'm in a hurry. One of these days I'll post all the source code, well commented. |
|
|
|
« Last Edit: August 05, 2008, 12:55:21 pm by Chester »
|
|
|
|
DarknessSavior
Guest
|
|
« Reply #599 on: August 05, 2008, 01:45:07 pm » |
|
Anyone else get a migraine reading that?
Chester, that stuff looks excellent. I can't wait to see more. ^.^
~DS
|
|
|
|
|
|
RHDN Forum Archive
