An interesting glitch in using only one channel

FreezeFlame(Alchemy) · November 01, 2014, 20:22:17

I don't know why, but if only one channel is in use, then it will cancel the note at row 15,31,47,63 in OpenMPT (in XMPlay, it cancels it out in row 31 and 63).

I made an test module to demostrate this 'voice cancel'.
In pattern 0, only one channel is in use (demontrating the little glitch).
But in pattern 1, another voice is included and the glitch dissapears (until you cancel out the sample/s being played on it).

Not an really important glitch, but worth mentioning.

Saga Musix · November 01, 2014, 21:22:48

QuoteNot an really important glitch, but worth mentioning.

Every playback bug is a bug worth mentioning, but I have no idea what you are talking about here. I hear all notes being played, so where's the bug supposed to be?
Channels never "talk" to each other on the code level in OpenMPT, so if there is something that sounds different if two channels are playing compared to one channel, it must be in the nature of the samples (e.g. phase cancellation) and no the code itself.

FreezeFlame(Alchemy) · November 01, 2014, 21:53:27

QuoteI hear all notes being played, so where's the bug supposed to be?

How can this be? It plays broke in XMPlay and OpenMPT 1.23.05.00.
But if its not to much to ask, does it not cancel the note at row 15,31,47,63 in you're current OpenMPT build?

But if it helps at all, i uploaded an OGG Vorbis file which shows, how the playback works on 1.23.05.00 on my PC on pattern 0.

Saga Musix · November 01, 2014, 22:22:25

Uhm, the OGG file just proves once more that every note is playing as expected? There's not a single second of silence in there. Do you need a picture to prove it?
Maybe acoustics are just playing a trick on you and your brain filters out notes that are actually there.

FreezeFlame(Alchemy) · November 01, 2014, 22:44:05

I think it is this phase cancellation that is causing this problem.
I have cut the part that cancels the note at row 15. Have a look at it yourself.
And i have checked the waveform (in stereo), thank you.

Saga Musix · November 01, 2014, 23:13:49

Err. I still don't get the point you are trying to make. Assuming that the content of that wav file is supposed to be row 15 and 16, then what is wrong with it?! It's a G-5 playing at half volume, and C-4 at quarter volume. No missing notes to be seen anywhere.
Besides, phase cancellation was an example of what could be happening with two samples playing at once. You cannot have phase cancellation with just one sample (unless it's of course part of that sample).

FreezeFlame(Alchemy) · November 01, 2014, 23:25:25

QuoteI still don't get the point you are trying to make. Assuming that the content of that wav file is supposed to be row 15 and 16, then what is wrong with it?!

That there is no sound to be heard. The waveform i uploaded captures how it goes from F5 at full volumen to G5 at half volumen and G5 gets quickly cancels out.

Saga Musix · November 02, 2014, 00:00:39

Yes, the sample's waveform actually decays to about half volume. Just zoom into it and look at it. So, nothing wrong.

FreezeFlame(Alchemy) · November 02, 2014, 01:08:56

I found the issue.
I tried to create an panbrello effect, which caused that G5 at row 15 gets quicly cancels out (in XMPlay, its in row 31).
But after i removed the S8x commands from the effect column, it works flawless.

Saga Musix · November 02, 2014, 01:17:18

...so you were only talking about the left channel in that example? Isn't it quite, err, logical that when you pan a note hard-right (S8F), it won't be heard on the left channel at all?

FreezeFlame(Alchemy) · November 02, 2014, 11:27:04

Quote...so you were only talking about the left channel in that example? Isn't it quite, err, logical that when you pan a note hard-right (S8F), it won't be heard on the left channel at all?

'sign'.
I think that this S8x command is the problem. I have replaced the supersaw with the snare and only leaved s8x commands on row 15,32,47,63 and it freaking cancels not only one note of an row, but everything on that channel until it gets to the next s8x command (in OpenMPT 1.23.05.00).

Saga Musix · November 02, 2014, 14:39:49

Okay, may I suggest you to check if your speakers are set up correctly? In any version of OpenMPT I've tried (OpenMPT 1.23 stable and 1.24 testing), this behaves as expected - 16 centered hits, 16 hits on the far right, 16 hits on the far left and then 16 hits on the far right again. If by "cancelling" you mean that you don't hear anything at all anymore, maybe one of your speakers is dead or are not wired up properly?

FreezeFlame(Alchemy) · November 02, 2014, 16:08:25

I do admit that i have an surround setup and could only wire one set (front or rear).
I'll redo my setup later.

FreezeFlame(Alchemy) · November 04, 2014, 18:46:38

I have tried it also on my old Netbook... and it worked without those nasty cancels.
Atleast i know that my surround set needs to be rewired.