Amiga questions you've always been too embarrassed to ask

[Daedalus]

Well, maybe. That size simply doesn't have enough pixels to show the subpixel scrolling, so there probably are points where pixels go missing. It's also possible that WinUAE just doesn't bother with it when the window size is too small, in which case you'd get standard OCS-style scrolling - I haven't tried it.

Edit: I tried it just there with a 720x576 window and it does half-pixel scrolling instead of quarter-pixel scrolling, but still looks artifact free.

[ross]

Originally Posted by Daedalus

Edit: I tried it just there with a 720x576 window and it does half-pixel scrolling instead of quarter-pixel scrolling, but still looks artifact free.

Yes, i've also noticed this.
From an old post you can grab a quirky test program written for Toni for an unrelated beta test:
http://eab.abime.net/showpost.php?p=...7&postcount=44

If i'm not wrong contain a 1/4 pixel scrolling.

[Toni Wilen]

Amount of possible subpixels depend on "resolution" selection (in Display panel) which selects internal base resolution: Lores=no extra subpixels possible (1x), hires=2 subpixels (2x), shres=4 subpixels (4x).

Note that subpixel emulation does not fully emulate them, it is currently a hack: only scan line based (mid line changes can't work 100% correctly) and both odd and even plane shift values must be same. Full emulation would need major redesign of internals, fortunately this handles most common situations well.

If you want non-pixel perfect shres artifacts/mixing, tick "filtered low resolution". It averages color of two pixels horizontally.

[ross]

Originally Posted by Toni Wilen

If you want non-pixel perfect shres artifacts/mixing, tick "filtered low resolution". It averages color of two pixels horizontally.

Toni this filter even hi-res pixels resolution?
I remember in the '80 my A500 connected to a poor 14" CRT TV where the RGB "shadow mask" was so grainy that the "640" resolution was blurry.
So a color blend was possible

[Amiga1992]

Thanks for all your answers about this subject!

I have a new one: what is self-modifying code and what is it used for/why?

[demolition]

Originally Posted by Akira

I have a new one: what is self-modifying code and what is it used for/why?

Self-modifying code is just that - self-modifying. The simplest type of SMC is just code which changes hard-coded 'constants' in the code, so instead of using using an instruction which uses a register to point to a memory address, you can save a bit of running time by using an immediate inside an often-used code path, as long as you do not need to change that constant very often as that will add some overhead. More advanced types of SMC will dynamically create entire segments of code. This is done e.g. by many C64 demos where the demo effects are basically procedurally generated code.

[Jope]

It can be used for obfuscation, for a simple example you have a NOP in the disassembly that gets modified into a real opcode a bit earlier, so when the CPU gets to that position it will never actually see a NOP. Or a BNE could be changed into a BEQ on the fly before execution gets there, or whatever. :-)

The aforementioned code generation is also good for keeping the executable size and loading/decrunching times down.

To give an example for the running time scenario above, especially on the C-64 and other similar small architectures, self modifying code is often used in place of pointers to save a lot of execution time. Think about a smooth scroller, you can use a pointer or you can just modify the address where the scroll text is copied from directly into your code, then it is automatically going to copy from the next character when that copy code is run during the next frame. In a C-64 pointers are quite a big performance hit and you have precious little time per frame.

[Amiga1992]

I don't think I quite understand it.
How exactly is this possible? How does the code modify itself? Wouldn't that take lots of extra space?
Again I think I aimed too high with my question and the explanations do not get through because I lack pre-required knowledge.

[daxb]

Maybe an example of pseudo code could it make more clear?

[Jope]

Originally Posted by Akira

I don't think I quite understand it.
How exactly is this possible? How does the code modify itself? Wouldn't that take lots of extra space?
Again I think I aimed too high with my question and the explanations do not get through because I lack pre-required knowledge.

In a Von Neumann architecture, code and data share the same memory space. So you just write stuff into memory before or after it has been executed. The CPU's program counter ticks on relentlessly and the CPU fetches the next instruction decided by the program counter, but any instruction it is not currently fetching/executing can be modified, even right before you get to the instruction you just modified. Naturally instruction and data caches mess this up, hence self modifying is rather frowned upon in systems like the Amiga. Still, it's possible given that you flush the caches.

Yes, it takes time to execute and space to store the instructions for changing the memory. However your code generator can be in a loop and thus can generate a lot more code than what the code for the generator itself takes up.

[d4rk3lf]

I am not afraid to ask this question, but I don't know where else I can ask.
Mortal Kombat 2 for the amiga has 4 floppy disks (cracked), yet, I've read somewhere that original had only 3 disks.
Why?
Crack (in other games) is just a braking security, not adding to the size of the game.

Apart from that, I want to ask owners of original MK2 on Amiga, is swapping disks was so intense as with 4 disks versions? I remember, between each fight, you must swap 2 disks (with pretty long loading time). Logic say 3 is less then 4, so swapping should be less intense.

I know this is not something that important, but I am pretty curious.
If I find, I might buy MK2 original disks to check out myself, but that doesn't answer my first question.

Thanks in advance.

[slaapliedje]

Originally Posted by d4rk3lf

I am not afraid to ask this question, but I don't know where else I can ask.
Mortal Kombat 2 for the amiga has 4 floppy disks (cracked), yet, I've read somewhere that original had only 3 disks.
Why?
Crack (in other games) is just a braking security, not adding to the size of the game.

Apart from that, I want to ask owners of original MK2 on Amiga, is swapping disks was so intense as with 4 disks versions? I remember, between each fight, you must swap 2 disks (with pretty long loading time). Logic say 3 is less then 4, so swapping should be less intense.

I know this is not something that important, but I am pretty curious.
If I find, I might buy MK2 original disks to check out myself, but that doesn't answer my first question.

Thanks in advance.

That's like the opposite of what usually happens with a crack, most of them tend to compress the data as well. Like Frontier: Elite 2 on the ST went from two 720kb floppies to a single 300kb executable in the cracked version.

Now I'm going to have to check that out, never played any of the MK games on the Amiga.

[ransom1122]

The 3 disk and 4 disk version of MKII are available here
http://grandis.nu:81/eabsearch/searc...xclude=&limit=
Compare the two via winuae

[slaapliedje]

That looks like 4 disk Mortal Kombat 2 and 3 disk Mortal Kombat 1?

[ransom1122]

Originally Posted by slaapliedje

That looks like 4 disk Mortal Kombat 2 and 3 disk Mortal Kombat 1?

Oops your right. And yes original is definitely 3 disks cracked is 4 disks. Interesting..

The only other suggestion is to try the ipf version that should be a clone of the 3 disk original.

<please don't link to sites that contain IPFs>

[demolition]

Originally Posted by d4rk3lf

Mortal Kombat 2 for the amiga has 4 floppy disks (cracked), yet, I've read somewhere that original had only 3 disks.
Why?
Crack (in other games) is just a braking security, not adding to the size of the game.

The cracked versions use standard sector formats so they can be copied using regular tools, but originals sometimes used custom sector formats which not only serves as a means of copy protection but would also allow them to put more data on each disk. Sometimes a cracker would remove data like an intro to make it fit on the cracked disks or some times they would convert a 3-disk game to a 4-disk version. That would also leave them a bit of room for a cool cracktro.

[Daedalus]

Originally Posted by slaapliedje

That's like the opposite of what usually happens with a crack, most of them tend to compress the data as well. Like Frontier: Elite 2 on the ST went from two 720kb floppies to a single 300kb executable in the cracked version.

It's similar on the Amiga - the Frontier disks are in standard Amiga format so can be read just fine in Workbench. The main executable is less than half one disk, and the second floppy just contained some extra save games and playable scenarios, nothing required by the game at all and so usually not bothered with for cracked versions. The floppy capacity is irrelevant in this case really.

[idrougge]

Super Skidmarks offers sub-pixel scrolling.

[idrougge]

Turrican II used a special "long" floppy format which couldn't be written on standard Amiga disk drives (but could be written using modified drives and copied using industrial disk copiers), so the cracked versions had to be moved to two disks or it couldn't have been copied using X-copy.

[d4rk3lf]

Originally Posted by demolition

The cracked versions use standard sector formats so they can be copied using regular tools, but originals sometimes used custom sector formats which not only serves as a means of copy protection but would also allow them to put more data on each disk. Sometimes a cracker would remove data like an intro to make it fit on the cracked disks or some times they would convert a 3-disk game to a 4-disk version. That would also leave them a bit of room for a cool cracktro.

Thanks.
That explaining it. However, it raises another question at how much trouble people that cracked the game had with re-programming the "insert disk x", so appropriate warrior loads.
And still want to see how swapping is intense on original game.

Here's some interesting interview I found few months a go.
Interview with Richard Costello - coder of Mortal Kombat
https://www.ppa.pl/gry/interview-wit...auntlet-2.html