Sounds like a very interesting development!

Quote:

Ops! Sorry for any incomvenience... By the way, I've been coding hard these last two days, and I have done a sample editor to be included in a MSX tracker. It allows to edit 32 step samples, adjusting envelope, noise frequency, mask for both effects and tune modulation. Sounds quite well for a simple PSG thinggy. It's somewhat based upon SoundTracker for ZX Spectrum. In a few days I will put it on-line in my own page. Kind regards, Ed Robsy

I had no idea the ZX Sound Tracker could use samples but that ideed would be a very good thing to have such a tracker ported to MSX. PSG rules

Keep us updated on your progress.

Well, indeed they are not real "samples". They are a combination of 3 series of data: amplitude envelope, noise frequency and tune modulation (signed offset).
So, a sample is played for every note using these 3 modifications affecting it. The replaying speed for a simple sample is 50 Hz (PAL) or 60 Hz (NTSC). And it can be up to 32 steps long.
Old good Gremlin Graphics games used this feature for its impresive intro songs (remember MASK2 or AVENGER?). And I think that also MATRA's excelent songs for the PSG use this feature.
A lot of interesting effects could be produced enabling/disabling tune and/or noise channels, and combining all the different effects. Detunning can also produce interestings sounds. And using all this is possible to simulate vibratto, tremolo, fake different bass sounds, etc.
As soon as I finish with the SAMPLE EDITOR I will start with the ORNAMENT EDITOR. With these two base programs I will mix them and add an editor to produce a complete tracker for the MSX system. PSG rulez! AY-3-8910 forever!

Kind regards,

Ed Robsy

Quote:

AY-3-8910 forever!

amen

Quote:

I also heard about pseudo-8bit samples playing (you can check it in some Bandwagon demos). Atari users use the PSG (they called it simply 'Yamaha' chip) to make that kind of sounds. Perhaps yo should ask Atari ST coders/musicians for this sampling technology on PSG.

Hey,

That's true and for the last demos (more speficially Vesikauhu), the samples were played realtime. The thing worked really simple. We had a sample playing at 8000 Hz and an effect refreshing at 50. So I'd

1. Play 160 bytes of the sample (=8000/50).
2. Dump one byte to the VRAM.

And so on, until the screen is finished. It was surprisingly easy even on the MSX1. 3.57 MHz is a lot of CPU power, really, when you hog the whole CPU to yourself. Too bad no-one seemed to notice our little trick.