PlayStation Development Network

I made a small example that shows how to use the XMPlay library and tools to add music to your application.
Please read the README.TXT included to see how the example is used.

The .BAT files are commented to show what each action does.
I tried to comment the main.c as clear and simple as possible to show the workings of XMPlay

Have fun!

Small addition
I am using the cpe2psx tool here (its in the MAKE_AND_PLAY.BAT)
if you do not have this, please change this command to CPE2X
and remove the line "ren main.psx main.exe"
Sorry for not adding this in the readme

If I follow the instructions in the readme file, the .exe does not compile.
This is without changing the .xm. It comes up with a parsing error.
I did change the .bat file as you instructed in your post since I wanted to use cpe2x.
I think cpe2x can't compile the code properly or something.

I have had the XMPLAY.LIB file on my PC for ages, but I never got around to using it nor uploading it. I checked the MD5, and it matches yours so both versions are the same. I was hoping yours would be newer since I've been playing around with it, and a few XM's don't play correctly which lead me to the library being not that accurately coded. As an example, I made a song from scratch using Fast Tracker, and the timing is off, even with a fixed BPM. The library isn't perfect, but this is a fantastic way to play several minute long songs with everything running from SPU RAM which is why it's being implemented into PSIO's Menu System.

I also contacted Jason Page (since he wrote the library) and I asked if he still had the source code. Unfortunately he does not, but, he did give me the PlayStation 2 source code which can be downloaded below.

Does anyone know if the included SEQ players in the psyq, are faster than this library?.

I imagine the included MID/SEQ player is just another software program that reads midi files, so it will be as fast as any other program (like this XM player).

I read some psyq samples and docs, and it looks like there is no "hardware" midi player, some programs send samples to the SPU with a few parameters, so I think the CPU has to read and process the MIDI/SEQ.

Correct the CPU has to tell the SPU what to do, it can't play a sequence on it's own. I wish I had something to add, but I haven't yet messed around with sequenced music, only standard CD audio.

MottZilla wrote: ↑July 7th, 2023, 2:23 pm Correct the CPU has to tell the SPU what to do, it can't play a sequence on it's own. I wish I had something to add, but I haven't yet messed around with sequenced music, only standard CD audio.

Thanks!

Mills wrote: ↑June 21st, 2023, 10:10 pm Does anyone know if the included SEQ players in the psyq, are faster than this library?.

I don't know. But the SPU-ADPCM playback is done by hardware, and the CPU does merely need to send some parameters (for things like changing the sound frequency, sending key on messages). The CPU load should be less than 1%, so there shouldn't be much of a difference between faster or slower sound drivers (or completely disabling the music).

I have the same timing problem here. Not a question of CPU or performance. Even at slow BPM, notes are not played on time. It sounds like VSyncCallback(XMPlayer::vbl); was not stable in term of time interval.

The same timing problem than what? Nobody mentioned any timing problems.

Anyways, the timing problems I could think of:

Vblank accuracy is approx +/-0.001 minutes, but I doubt that even audiophiles could hear differences between 119.999 bpm and 120.001 bpm.
PAL vs NTSC have different vblank timings.
Don't block audio playback by doing things like disabling IRQs for longer periouds, or doing huge DMA transfers with hundreds of kilobytes.
That callback thing sounds as if you need to call it (exactly?) once per frame. Which would be up to you to do so accurately. It may be best to call it from within vblank interrupt handler. Or when manually polling vblanks: Be sure to do that polling often enough (eg. poll several times during slow initialization functions).

This looks interesting from the perspective of someone who would love to see hombrew sequenced music running on either the PS1 or PS2 (preferably the latter, which is why I'm glad to see the source code for this program for that console here). I'd personally prefer either MIDI or its PlayStation SDK equivalents in SEQ/VH/VB and SQ/BD/HD, but you've gotta start somewhere, right? I'm presuming that this all runs using IOP cycles; or in order words, not using the actual SPU/SPU2 channels besides playing the final output mixed on the IOP. I'm making said guess based on the fact that the PowerPoint slides say that reverb isn't implemented in at least whatever version they're referring to, which I would be relatively trivial and you have access to the registers through the sound drivers, isn't that true?