I guess both solutions will be the ideal:
- The player should give a option to choosing no sound at all, music and sfx simultaneouly and no music but sfx, no sfx but music.
- If music and sfx are selected, 1 channel could be used at sound routine will (more channels could be possible, but do not is the ideal). Anyway, should be a preference of the author of the sound fx and music how many channels will be used for each sound fx and music.

My preference:
- Music in 3 channels. Only 1 channel with drums and percussion (using noises).
- Sound fx in 1 channel. The exactly percussion channel should be used to sound fx. Remember ou learn thats is only one noise generator.

I will use (probably) SEE and PT3 for my sweet Cosmic Battle. If I could obtain the integration of both, I will post the code here in MRC.

In SEE there is a "prio" byte that will determine the priority of a requested SFX. (a explosion will prolly be more important than just your gun making some noise)
Same principle can be used in PT3. (some status bit for each channel or sumfp.)

see newsfact:
http://www.msx.org/newspost2984.html

thx (b)ro ^_^

Channel masking for sfx over music produces "low quality" sound. Konami's don't mask sounds, they just produce 2 channel melodies and then reserve the third channel for sfx. Do you dare say its music is bad? ;-)

mostly masking (also on konami tracks) is done over channels with lower prio. like a drumtrack, voice double/echo etc.

In Nemesis saga my ears tells a thing somehow differently. I did'nt debug the games. Seems to me that some channel is used for music tones/noises and sfx tones/noises. Is it in this way?

All the players are called at the IRQ rate (1/50sec).

Do you have examples of players that can be safely
called at lower rates (2/50sec) ? (this is usefull
when you run out of time machine)

even harder now

Does anyone have tryed to interlace two players
the first plaing the music, the second something else?

each one would have an execution rate of rates 2/50sec,
can anyone figure out how the result sounds like ?

If I stay at the DSP theory, 1/50 sec of sampling rate
cannot be used for interlacing sounds (I get huge aliasing).
Nevertheless PSG keeps its programming from a sample
to the next, thus I cannot speack of real down sampling ...
The spectrum (frequency not the ZX) should be more or less preserved...

Any guess on the subject?

Quote:

Quote: PT3.LHZ is version 1.0 while PT3OOLS is version 1.1 has no ASM codes!! That's not good! I have send a new package that includes the .asm's to snout to replace the current package.

Could you send me your PT3player code version 1.1 ?

Artrag: it's --> RIGHT HERE <--

It's really not that hard! You use the first 18 wave channels of the OPL4 for music, and the remaining 6 for SFX!

Alex Wulms should release SME4.0, supporting PSG, MSX-MUSIC, 5 SCCs, 2 MSX AUDIOs and 1 or 2 Moonsounds. That should get the job done ^_^.

What is SME4.0 ?
Who is Alex Wulms?

No engineers or mathematicians to solve by question about sub-sampling ?

Half Timing is already done in most replayers. Depending on the "speed"/"tempo" of the track. you get in trouble tho when using soft tremolo/vibration etc etc

Oracle for example; every interrupt is used to update volume/cut-off/frequency/arpeggio and standardPlus-instrument change. So that'll be a problem.

Some less complex trackers/musicdrivers might just do some notesettings depending on trackspeed. that replayer will be perfect for downsampling.


so in theory (and prolly in practise) you can double the track speed and cut the timing to half sampling. but you'll end up with an un-even interrupt length... don't know if ya want that.

No problem if you reduce rate of the player.

If I half the "tempo" and half the call rate to the player
the final effect is not noticeable.

As the HW keeps producing the sound
between two calls to the player,
the problem is only in the speed of updating the
tones (e.g. for vibrato effects as you say).

Different is if you interlace the execution of two different effects.
I mean that you can call a routine that play music on even frames,
and another that plays effects on the odd frames.

What ugly result should I expect from such a mess?