Heroes of Might and Magic II

[Raybeez]

Understood, if I will find any help I will write You.

[zzbylu]

HoMM II on the real Amiga 1200/030 - gameplay:
[ Show youtube player ]

[saimon69]

Beside the 128 megabytes and the 50 mhz FPU that is mostly my A1200 config in italy (viper EC030/42mhz+FPU+32 MB ram) - now Meynaf would go mad at me more - since i still need to finish all audio tracks - but i wonder if that code could be a bit optimized, in-field graphics could be reduced of colors, audio rediced to lower rate or the C2P could use an approach similar to shapeshifter EVD - but that is just me since am no coder

[meynaf]

Originally Posted by saimon69

Beside the 128 megabytes and the 50 mhz FPU that is mostly my A1200 config in italy (viper EC030/42mhz+FPU+32 MB ram) - now Meynaf would go mad at me more - since i still need to finish all audio tracks - but i wonder if that code could be a bit optimized, in-field graphics could be reduced of colors, audio rediced to lower rate or the C2P could use an approach similar to shapeshifter EVD - but that is just me since am no coder

Of course the code can still be optimised, but not in the ways you suggest.

Reducing colors isn't realistic because many effects expect 256 colors, and these particular 256 colors. There is notably palette cycling that would fail miserably, shadowing done by reading on-screen pixels and darkening them, etc.

Decreasing the rate of audio would only bring quality loss - audio does not currently take significant cpu time.

And the C2P already updates partial screen regions (else it would be real damned slow, believe me - even the mouse pointer would be so sluggish you'd want to just throw the game !).

But it's C++, and the code is largely sub-optimal.
Some parts have been rewritten in asm already.
However it's basically huge...

[NovaCoder]

Originally Posted by meynaf

This is in theory possible, but I have no knowledge about these. The game has mainly been made for A1200/030 with just AGA.

Last version has had some code cleanup in this area (c2p in include file rather than being buried deep in the game itself) so now it's just a matter of how to use these graphics cards.
If someone can give me simple code for handling some (often partial) copy of 640x480x8 chunky frame buffer to the video mem, then i might do this.
Also i need to setup this under Winuae so i can test if the code works.

You would need this kind of thing..

PHP Code:

#include <proto/exec.h>
#include <proto/intuition.h>
#include <proto/graphics.h>

#include <cybergraphics.h>
#include <proto/cybergraphics.h>
 
struct Library *CyberGfxBase = NULL;


/** Global Hardware window */
static struct Window *_hardwareWindow = NULL;


/** Hardware screen */
static struct Screen *_hardwareScreen = NULL;



CyberGfxBase = OpenLibrary ("cybergraphics.library", 0);
if (CyberGfxBase == NULL) {
    SDL_SetError("Can't open cybergraphics.library\n");
    return -1;
}  

<SNIP - Open screen and window, choose display mode>    
    
static void RTG_UpdateRects(int numrects, SDL_Rect *rects)
{
    int i;
    Sint16 dirtyStart, dirtyEnd;
    UBYTE* src;
    UBYTE *base_address;
    APTR video_bitmap_handle = NULL;
        

    // Seed.
    dirtyStart = SDL_VideoSurface->h;
    dirtyEnd = 0;
    
    
    // Find dirty strip extents.
    for (i = 0; i < numrects; i++) {
        SDL_Rect *r = rects + i;
        if (!r) {
          continue;
        }

        if (r->y < dirtyStart) {
            dirtyStart = r->y;
        }
        
        if ((r->y + r->h) > dirtyEnd) {
            dirtyEnd = (r->y + r->h);
        }
    }

              
    src = (UBYTE*)SDL_VideoSurface->pixels + (dirtyStart * SDL_VideoSurface->w);
    
     
    // Check whether the palette was changed.
    if (_paletteDirtyEnd != 0) {
        RTG_updatePalette();
    }
        
    
    video_bitmap_handle = LockBitMapTags(_hardwareScreen->ViewPort.RasInfo->BitMap,
                                         LBMI_BASEADDRESS, (ULONG)&base_address,
                                         TAG_DONE);
    if (video_bitmap_handle) {
        base_address = (base_address + (dirtyStart * SDL_VideoSurface->w));
        CopyMemQuick(src, base_address, (SDL_VideoSurface->w * (dirtyEnd - dirtyStart)));
    }

    if (video_bitmap_handle) {
        UnLockBitMap (video_bitmap_handle);
    }
} 


[meynaf]

Originally Posted by NovaCoder

<SNIP - Open screen and window, choose display mode>

That part may well be the biggest problem, in fact.

Originally Posted by NovaCoder

    // Check whether the palette was changed.
    if (_paletteDirtyEnd != 0) {
        RTG_updatePalette();
    }

How to transmit actual palette data ?

[NovaCoder]

Originally Posted by meynaf

That part may well be the biggest problem, in fact.



How to transmit actual palette data ?

PHP Code:

int RTG_SetVideoMode(int width, int height, int bpp, Uint32 flags)
{
    ULONG modeId = INVALID_ID;

    
 
    // Automatically choose the best mode.
    modeId = BestCModeIDTags(
                CYBRBIDTG_Depth, bpp,
                CYBRBIDTG_NominalWidth, width,
                CYBRBIDTG_NominalHeight, height,
                TAG_DONE);


    // Verify the mode choosen.
    if (modeId != INVALID_ID) {
        if (GetCyberIDAttr(CYBRIDATTR_DEPTH, modeId) != bpp) {
            modeId = INVALID_ID;
        }
    
        if (GetCyberIDAttr(CYBRIDATTR_WIDTH, modeId) != width) {
           modeId = INVALID_ID;
        }
    
        if (GetCyberIDAttr(CYBRIDATTR_HEIGHT, modeId) != height) {
           modeId = INVALID_ID;
        }
    }

    if (modeId == INVALID_ID) {
        SDL_SetError("Couldn't find a CyberGFX Screen for requested mode");
        return(-1);
    }
    
    // Create the hardware screen.
    _hardwareScreen = CreateHardwareScreen(modeId, width, height, bpp);
    if (!_hardwareScreen ) {
        SDL_SetError("Couldn't create a Hardware Screen for requested mode");
        return(-1);
    }
    

     
    _hardwareWindow = CreateHardwareWindow(width, height);
    if (!_hardwareWindow) {
        SDL_SetError("Couldn't create a Hardware Window for requested mode");
        return(-1);
    }                
   
    



    /* We're done */
    return(1);
} 


PHP Code:

static struct Screen* CreateHardwareScreen(ULONG modeId, int width, int height, int depth) {

    return OpenScreenTags(NULL,
                     SA_Depth, depth,
                     SA_DisplayID, modeId,
                     SA_Top, 0,
                     SA_Left, 0,
                     SA_Width, width,
                     SA_Height, height,
                     SA_Type, CUSTOMSCREEN,
                     SA_Quiet, TRUE,
                     SA_ShowTitle, FALSE,
                     SA_Draggable, FALSE,
                     SA_Exclusive, TRUE,
                     SA_AutoScroll, FALSE,
                     TAG_DONE);        
}

static struct Window* CreateHardwareWindow(int width, int height) {

    return OpenWindowTags(NULL,
                      WA_Left, 0,
                    WA_Top, 0,
                    WA_Width, width,
                    WA_Height, height,
                    WA_PubScreen, (ULONG)_hardwareScreen,
                    WA_Backdrop, FALSE,
                    WA_Borderless, FALSE,
                    WA_Activate, TRUE,
                    WA_SimpleRefresh, TRUE,
                    WA_NoCareRefresh, TRUE,
                    WA_ReportMouse, TRUE,
                    WA_RMBTrap, TRUE,
                      WA_IDCMP, IDCMP_RAWKEY|IDCMP_MOUSEMOVE|IDCMP_DELTAMOVE|IDCMP_MOUSEBUTTONS,
                      TAG_END);
} 


PHP Code:

// Palette data
static SDL_Color *_currentPalette;
static int _paletteDirtyStart, _paletteDirtyEnd;

static void RTG_updatePalette() {
    
    int i, numberOfEntries;
    ULONG  _rtgPalette[256*3 + 2];
     
#ifdef DEBUG_PALETTE    
    SDL_LogDebugMsgchar ("RTG_updatePalette()\n");

    SDL_LogDebugMsgchar("RTG_updatePalette() - _paletteDirtyStart = %d\n", _paletteDirtyStart);
    SDL_LogDebugMsgchar("RTG_updatePalette() - _paletteDirtyEnd = %d\n", _paletteDirtyEnd);
#endif

    numberOfEntries = (_paletteDirtyEnd - _paletteDirtyStart);

    for ( i=0; i < numberOfEntries; i++ ) {
        _rtgPalette[i*3+1] = _currentPalette[_paletteDirtyStart + i].r<<24;
        _rtgPalette[i*3+2] = _currentPalette[_paletteDirtyStart + i].g<<24;
        _rtgPalette[i*3+3] = _currentPalette[_paletteDirtyStart + i].b<<24;
    }

    _rtgPalette[0] = (numberOfEntries << 16) + _paletteDirtyStart;

    // Terminator: NEEDED
    _rtgPalette[((numberOfEntries * 3) + 1)] = 0x00000000;

    LoadRGB32(&_hardwareScreen->ViewPort, _rtgPalette);


    // Reset.
    _paletteDirtyStart = 256;
    _paletteDirtyEnd = 0;
} 


[meynaf]

Ok.
Now i need to setup RTG under winuae. Just adding rtg board is easy, but how to setup it so that the above can work ? I don't even have cybergraphics.library (btw, do all rtg setups use this ?).

[Seiya]

hello, i reopen this thread to ask if there is a way to play music? i see from option in the game that is possible to play with MIDI. This porting can play music?
By the way, excellent porting for Amiga.

[meynaf]

Originally Posted by Seiya

hello, i reopen this thread to ask if there is a way to play music? i see from option in the game that is possible to play with MIDI. This porting can play music?
By the way, excellent porting for Amiga.

No, unfortunately there is no music. It's on the todo list, maybe one day (when i find a good solution).
Options exist to play either MIDI or CD but changing them will have no effect. Note that Mac version could not play MIDI either (actually, XMI).

[Seiya]

ok, thanks

[saimon69]

well, i was tasked to do the music, and seems still got a lot to do; i did listen cold lately at what i did and i admit more than half of it is not as good as it should be

Am busy with two jobs right now and several gigs for other projects but would love to have this on my pedigree too, and done better than is now, hoping that all those projects give me better skills than in 2016

[DamienD]

Hi meynaf,

Not sure if you're aware of this or not... someone over at https://www.ppa.pl/ updated the game file and editor:

Originally Posted by Aladin

Hi DamienD

Heroes 2 meynaf have an update: v0.5

https://www.ppa.pl/forum/strefa-gier...gry/15#m714550

For your english version, i think you have juste to replace the file heroes2 and editor2

This is what the Polish post translated says:

In response to the #444

Heroes of Might and Magic 2 ported to the Amiga: link

As roman Kra?niewski
writes
on facebook: Configuration to launch the game: Minimum
requirements
AGA - 68030 - 2
MB Chip - 8 Fast Recommended Requirements AGA 68040 2 MB Chip - 16 Fast Link to download my version of the game to the Amiga;
https://files.fm/u/9jwd7x2xy#/view/tzhy6dwut
NOTE: The game fires from the file - heroes2 and not from
the file Heroes2w.exe Configuration file to launch the game was designed by Meynaf's encoder who is famous for moving games from AtariST to Amiga. But the project was abandoned years ago, I undertook the task to complete this project. The basic version allowed you to fire up the game, but the problems were with animations between campaign missions.
I tried to make an additional Polish language version and I succeeded but not entirely. that is, menus and options and any button are in Polish but all descriptions are still in English, I used a dozen versions of PL and the result is always the same, and strange because the version from CDProjekt firing on Windows is normally pl version.
I tested the game in every way to suspend it and wait for an error but i did not suffer any problems.
So I say I tweak it a little to make sure there are no problems adding animations and other features in the game. But the game works very well. I completed 3 campaign missions each time after completing the mission fires up an animation of the description of the next mission with a Polish teacher and fires the mission. In a word SUCCESS!!
Animations for weaker configurations can be set to Interlace when video function and animation will be smoother.

Indeed the small issue that sometimes occured with animations between Campaign missions is fixed

[Aladin]

the readme:

0.5
. both: c2p slightly optimised (but this is probably not noticeable at all)
. game: some heroes were nearly never available in the player's castle and the
first hero wasn't always of the player's type (but of another player)
. game: spell "elemental storm" triggered an abort in 0.4
. game: due to misuse of sprintf in original code, signposts consisting of a
text having the "%" symbol in them weren't shown properly
. game: no longer stops the color cycling when moving a hero - why they did
this (it was intentional) is beyond me
. game: while the game explicitly forbids giving your spellbook to someone
else, it no longer prevents you moving it inside your own inventory
. game: titans located at the top of the screen and firing with a narrow angle
could destroy some off screen mem (negative start y coord), ultimately
leading to random crashes
. game: could not build something after building a boat, where a boat could be
built anytime - now the boat is ignored but if you think the game should
have always disallowed building a boat and something else at the same turn,
well, then, just don't do it
. game: you can now select midi output again (but no effect, of course...)
. game: path finding made faster (twice as fast as before), should speed-up
computer moves too

[DamienD]

Originally Posted by Aladin

the readme:

0.5
. both: c2p slightly optimised (but this is probably not noticeable at all)
. game: some heroes were nearly never available in the player's castle and the
first hero wasn't always of the player's type (but of another player)
. game: spell "elemental storm" triggered an abort in 0.4
. game: due to misuse of sprintf in original code, signposts consisting of a
text having the "%" symbol in them weren't shown properly
. game: no longer stops the color cycling when moving a hero - why they did
this (it was intentional) is beyond me
. game: while the game explicitly forbids giving your spellbook to someone
else, it no longer prevents you moving it inside your own inventory
. game: titans located at the top of the screen and firing with a narrow angle
could destroy some off screen mem (negative start y coord), ultimately
leading to random crashes
. game: could not build something after building a boat, where a boat could be
built anytime - now the boat is ignored but if you think the game should
have always disallowed building a boat and something else at the same turn,
well, then, just don't do it
. game: you can now select midi output again (but no effect, of course...)
. game: path finding made faster (twice as fast as before), should speed-up
computer moves too

No, that's meynaf's "0.5" changes

Here's what I added to the meynaf's "History.txt" file:

I've tested further now and... although the intermittent slight speech issues with animations between the Campaign missions has been fixed; I have a major issue with the intro animation after clicking on "New Game" --> "Campaign" jittering like hell

[ Show youtube player ]

I'm running a 68040 / Fastest possible configuration... if I tick "Jit" then this fixes the issue but the game then plays way too fast...

So I'm afraid that I can't use this unofficial update.

[Aladin]

So I still had the version 0.4 readme on my installation, I do not have have replaced the history file, thank you for the clarification

[DamienD]

Originally Posted by Aladin

So I still had the version 0.4 readme on my installation, I didn't have to replace the history file, thank you for the clarification

Do you have the same issue as me though with the "unofficial update"?

Originally Posted by DamienD

I've tested further now and... although the intermittent slight speech issues with animations between the Campaign missions has been fixed; I have a major issue with the intro animation after clicking on "New Game" --> "Campaign" jittering like hell

[ Show youtube player ]

I'm running a 68040 / Fastest possible configuration... if I tick "Jit" then this fixes the issue but the game then plays way too fast...

So I'm afraid that I can't use this unofficial update.

[Aladin]

No with cpu max and nojit, i have first campaign video (the same that your video) running good. (winuae 68060 and 68040)

the only video that jerks a little is the intro video at the launch of the game

I'm using Homm2 french

[DamienD]

Originally Posted by Aladin

No with cpu max and nojit, i have first campaign video (the same that your video) running good. (winuae 68060 and 68040)

the only video that jerks a little is the intro video at the launch of the game

I'm using Homm2 french

Hmmm, without "Fastest possible" all videos jitter like the Campaign intro and the game far too slow

I guess my laptop is too old...

I can get around the "Campaign intro" issues by setting "Jit" and "CPU Speed = -60%":

[meynaf]

Seems someone has made some hacks to try to fix something.
But if a fix breaks something else, of course not a good fix.