Protracker player with support for external sound fx

[Galahad/FLT]

Originally Posted by Galahad/FLT

Thanks for doing this, assembles properly in Devpac now

However, not sure if its something i'm doing wrong, or if theres a problem with the routine, but i've tried two separate modules now, and after playing about 20 seconds or so, both stop playing using this routine.

Is it just me?

Anyone?

[phx]

Just tried it again with some different mods. Stringray's source and my updated version both seem to work here.

Which one did you use? Note, that when using the updated version (2.0 on Aminet) the API for mt_init changed. You have to pass additionally A1=0 (no separated samples) and D0=0 (start playing at position 0).

Otherwise, on which hardware did you test. And which mod? Maybe you can provide a test source or binary?

[spud]

I've been trying to read up and figure out amiga interrupts because this is one of many areas I'm pretty ignorant on so I thought I'd edit it the code to show at which scan line the player starts and for how long it runs. It seems to start slightly earlier each frame, should this be the case?

[Galahad/FLT]

Originally Posted by phx

Just tried it again with some different mods. Stringray's source and my updated version both seem to work here.

Which one did you use? Note, that when using the updated version (2.0 on Aminet) the API for mt_init changed. You have to pass additionally A1=0 (no separated samples) and D0=0 (start playing at position 0).

Otherwise, on which hardware did you test. And which mod? Maybe you can provide a test source or binary?

I used Stingrays attachment, all conditions were correctly set, and I'm using WinUAE in an A1200 environment and using cycle exact.

The mods are for Where Time Stood Still. I thought it was an error in one mod that had an unsupported effect, but then I tried it with another mod and the same problem occurred, 20 seconds or so into playback, it stopped playing.

I will download your version from Aminet and see if that makes a difference.

[phx]

Originally Posted by spud

I've been trying to read up and figure out amiga interrupts because this is one of many areas I'm pretty ignorant on so I thought I'd edit it the code to show at which scan line the player starts and for how long it runs. It seems to start slightly earlier each frame, should this be the case?

Yes, that's normal. The player doesn't use the VERTB interrupt, which would be synchronized with the raster beam. It uses a CIA-B timer interrupt at a speed close to 50Hz, but it cannot and must not match perfectly.
You could run a 60Hz screen mode and the playback still has the expected speed.

[phx]

Originally Posted by Galahad/FLT

I used Stingrays attachment, all conditions were correctly set, and I'm using WinUAE in an A1200 environment and using cycle exact.

I tried to verify that: WinUAE, A1200, 3.1, 68EC020, 2MB Chip RAM, cycle exact CPU and chipset. No problem.

The mods are for Where Time Stood Still. I thought it was an error in one mod that had an unsupported effect, but then I tried it with another mod and the same problem occurred, 20 seconds or so into playback, it stopped playing.

I wouldn't exclude the possibility that there is still a strange effect which is unsupported, although I tried a lot of different mods. The Sqrxz-mods, for example, are quite complex.

Does it stop at exactly the same position everytime? What does "stopping" mean? The player becomes silent? The last notes continue to play forever? Something crashed?

I really would like to help, if I had something to reproduce.

[spud]

Originally Posted by phx

Yes, that's normal. The player doesn't use the VERTB interrupt, which would be synchronized with the raster beam. It uses a CIA-B timer interrupt at a speed close to 50Hz, but it cannot and must not match perfectly.
You could run a 60Hz screen mode and the playback still has the expected speed.

Makes sense, thanks for the explanation! Kind of obvious I suppose.

[Galahad/FLT]

Originally Posted by phx

I tried to verify that: WinUAE, A1200, 3.1, 68EC020, 2MB Chip RAM, cycle exact CPU and chipset. No problem.


I wouldn't exclude the possibility that there is still a strange effect which is unsupported, although I tried a lot of different mods. The Sqrxz-mods, for example, are quite complex.

Does it stop at exactly the same position everytime? What does "stopping" mean? The player becomes silent? The last notes continue to play forever? Something crashed?

I really would like to help, if I had something to reproduce.

EDIT: I'll punch myself in the face, save anyone else having to do it. I had some debug interrupt code still present and one tiny piece of keyboard code was wrong, which meant spurious results were interpreted as a key press and restoring all the interrupts back as they would be had AmigaDOS been reactivated.

I.AM.A.TWAT


One thing I had to do though was put an EVEN statement before MT_VAR otherwise there was some addressing problems for 68000 Amigas.

[phx]

Originally Posted by Galahad/FLT

I had some debug interrupt code still present and one tiny piece of keyboard code was wrong

LOL

One thing I had to do though was put an EVEN statement before MT_VAR

That was already fixed in the Aminet release. Stingray removed one CNOP too much while optimising mt_MasterVolTabs away.

[StingRay]

Originally Posted by phx

That was already fixed in the Aminet release. Stingray removed one CNOP too much while optimising mt_MasterVolTabs away.

Indeed, happened when I changed the mt_mastervol routine and deleted the pointer tables.

[Galahad/FLT]

Well the routine works a treat gents, and you both get a credit in the game, so now its nearly finished

[clenched]

Originally Posted by Coagulus

Might be a long shot but can anyone (or is it even possible to!) make a Blitzlib out of this?

This would be good substitute player for what is available for Blitz. Check out the rudimentary library here. It already scores three points ahead of others because:
- no KS2 dependency
- no guru (yet) on 68000
- doesn't demand 1meg to run.

Experimental commands:
PHXInst vbr,pal or nts - Install the player (if vbr is not 0 the value can be found in the debugger)
PHXInit address - initialize mod at address
PHXEnable 0=no music not0=music
PHXRem - remove the player.

How to install:
Put phx.obj into Blitz2:Blitzlibs/Otherlibs drawer.
Run the makedeflibs program.
Shutdown and restart Blitz.
Paste the test program into the editor.

If all went well the PHX commands will be highlighted the same as BASIC keywords. Fill in the path and name of a module and try to compile and run it.

To put Blitz back the way it was delete phx.obj and re-run makedeflibs.

chip.l=2
vbr.l=0
pal.l=1
nts.l=0
LoadBank 0,"some.mod",chip
a.l = Peek.l(Addr Bank(0))
NPrint a
PHXInst vbr,pal
PHXInit a
PHXEnable 1 ;0 no music
For j.l=1 To 25000
NPrint j
Next j
PHXRem

[sandruzzo]

are there some example to show how it work?

[alpine9000]

Originally Posted by sandruzzo

are there some example to show how it work?

I just did a quick hack to play a module and it seemed to work. Haven't tried SFX yet.

Note: I didn't use the small data model version. I used ptplayer_noa4.asm.

        ; mt_init(a6=CUSTOM,a0=TrackerModule, a1=Samples|NULL,d0=InitialSongPos)                                            
        lea CUSTOM,a6
        lea module,a0
        move.l #0,a1
        move.l #0,d0
        jsr mt_init

	; mt_install_cia(a6=CUSTOM, a0=AutoVecBase, d0=PALflag.b)                                                               
        move.l vectorBase,a0
	move.b #1,d0
        jsr mt_install_cia

	move.b #1,mt_Enable ; Start playing

module: incbin "module.mod"
vectorBase: dc.l 0 ; make sure this has the correct vector base

[phx]

Just a note, because the order of mt_init and mt_install_cia in alpine9000's example might be confusing:

Usually you will do mt_install_cia only once at the start of your program. It install a level 6 CIA-B interrupt handler and enables TimerA/B interrupts.

mt_init is called for every new module you want to play. Also for sub-modules, as you can pass the start position.

[Photon]

Originally Posted by phx

From the ReadMe:

The source is tested with the vasm assembler, but should work with most
assemblers after minor modifications (e.g. NEAR directive). Static
variables are accessed using the Small Data model, with base register
in A4.

Perhaps I should release a more portable source without using the small data model, as many people seem to have problems with it. Sorry for that.

The "minor" modifications would include:
- remove "xref _LinkerDB" and "near a4"
- remove the two "lea _LinkerDB,a4"
- remove the "(a4)" in all base-relative addressing modes
- optionally rename or remove "section __MERGED,bss"

The resulting code will be bigger and slower, but it might assemble with AsmPro after after that.

Or um just replace with BASEREG

If you need the compression for the module then P6112 supports sound effects by way of setting the number of channels, freeing the other channels for playing samples at any time.

[Tigerskunk]

I tried to include this in my code, but get lots of errors when compiling (./vasmm68k_mot -Fhunkexe -o e -nosym Innviya/main.asm):

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_Enable + 0xff800000) not supported

error 3009: undefined symbol <_LinkerDB> at CODE+0x81c, reloc type 1

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_Enable + 0xffffffc0) not supported

error 3009: undefined symbol <_LinkerDB> at CODE+0x8aa, reloc type 1

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_chan1 + 0xffffe000) not supported

Anyone any ideas whats wrong here and how I can fix it??

[Tigerskunk]

Originally Posted by Steril707

I tried to include this in my code, but get lots of errors when compiling (./vasmm68k_mot -Fhunkexe -o e -nosym Innviya/main.asm):

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_Enable + 0xff800000) not supported

error 3009: undefined symbol <_LinkerDB> at CODE+0x81c, reloc type 1

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_Enable + 0xffffffc0) not supported

error 3009: undefined symbol <_LinkerDB> at CODE+0x8aa, reloc type 1

error 3005: reloc type 10, size 16, mask 0xffffffffffffffff (symbol mt_chan1 + 0xffffe000) not supported

Anyone any ideas whats wrong here and how I can fix it??

Works with ptplayer_noa4.asm though, so maybe I will just use that one..

edit: but now my program hangs... damn...

[phx]

There have been some improvements to the ptplayer 2.0 release on Aminet, years ago. Besides some bug fixes and enhancements there is a more flexible function for playing sound effects, which was already used in games like Power Glove and Tiger Claw. From the readme:

The sound fx system gives you the possibility to play samples on a channel of your choice or on the channel which the player thinks is the best one.

It may be a channel which is currently not replaying music and/or has the longest period of silence ahead. This has the effect that the replayed song is often not disturbed at all.

Up to four sound effects can be played at the same time and any of them has its own priority, which is especially useful when trying to play several sounds on the same channel. You may for example define that a shooting-sound has a higher priority than a jumping-sound.

For automatic channel selection you can additionally reserve specific channels for music only, or define the maximum number of channels which may be used for sound effects at once.

The master volume is always applied to the music, but does not affect external sound effects at all.

The new sound-fx function interface:

_mt_playfx(a6=CUSTOM, a0=SfxStructurePointer)
  Request playing of a prioritized external sound effect, either on a
  fixed channel or on the most unused one.
  Structure layout of SfxStructure:
    APTR sfx_ptr (pointer to sample start in Chip RAM, even address)
    WORD sfx_len (sample length in words)
    WORD sfx_per (hardware replay period for sample)                
    WORD sfx_vol (volume 0..64, is unaffected by the song's master volume)
    BYTE sfx_cha (0..3 selected replay channel, -1 selects best channel)  
    BYTE sfx_pri (unsigned priority, must be non-zero)
  When multiple samples are assigned to the same channel the lower
  priority sample will be replaced. When priorities are the same, then
  the older sample is replaced.
  The chosen channel is blocked for music until the effect has
  completely been replayed.

To make the ptplayer 4.0 easier to maintain for me I integrated the small-data and the single-section variants into a single source text, using BASEREG by default. But I never released it, because I was not happy about the fact that BASEREG is only supported by vasm and AsmOne.

Now a reworked 5.0 can be found on Aminet (http://aminet.net/mus/play/ptplayer.lha), which is portable enough to work on most assemblers. Tested are: Devpac, vasm, PhxAss, Basm, SNMA, AsmOne, AsmPro. Additionally I added a leading underscore to all exported function- and variable-names, so they can be directly accessed from C source texts as well.

[BSzili]

Nice! Have you considered adding small stubs that take care of the stack/register parameters, and get the VBR and the chipset address? That would make it much easier to use it from C.