PC Engine waveform channel updates

I have recently taken interest in programming for the PC Engine/TG16. I've been meaning to do something interesting with the audio, so I tried regularly changing the waveforms to get more realistic instruments. https://www.youtube.com/watch?v=BI6so4w9mu8

Now based on what I read in the documentation, the waveform data can only be updated when the channel is disabled. But there is no mention whether the channel phase gets reset when you do that. If that actually does happen then the song in the video is actually impossible to do in an actual program. Could anyone confirm any of this to me? Thank you!

Cannot answer the question but this rearrangement wanted me to walk about town with this in my headphones. Great work!

Feel free to sign up to the PCEngine FX forums http://www.pcenginefx.com/forums/
and use the forum's search function for "waveform". There's been a fair amount of research into changing waveforms while channels are "active" and trying to avoid the popping and artifacts that go with it.

You must use timer interrupt for that,and it's definitively not trivial as it rely on the sound frequency .
The channel is off only when you are updating the new waveform, and switched ON after that,but you're limited on the frequency because it takes time to update the entire 32 bytes buffer, 209 cycles with TIN instruction, plus the others for interrupt and channel setting .

But it's doable in a game, it seems All the Fire Pro Wrestling games did it :
https://www.youtube.com/watch?v=V845pHyaePk

I did it for samples playback, and it works well only on the A rev. of the Hu6280.

Ah, I only intend to update the waveform once per field at the fastest (up to 1254 cycles if all 6 channels had to be updated), so to me the biggest concern is whether the waveform step is reset to 0 when I update the waveform.

You reset the waveform pointer like that :

When bit 6 goes from 1 to 0 the pointer is reseted and when you feed the buffer with the new waveform it start at 0 up to 31 and not at the current waveform pointer position .

I have since tried to do the same on the FDS, a system where I know that the channel has to be halted and reset to change the waveform, and the results are oddly clear with very little artifacts added. This is a promising next step.

BTW your PCE version of Journey to Silius is incredible