Print

[LigH]

Hi.

A long, long time ago ... when there were landline mailboxes instead of the Web, and I still used plain DOS without Windows 3.x or 95 even, I already listened to modules by different authors, including the "Members of Mayday" (Silicon Records), and one of their modules was a "HardCore VibeS (RmX)" by Spike.

I noticed that it sounds differently, depending on the player used. The sounds I liked were those produced by an older version of the Cubic player (up to legacy version 2.0α) and by MOD4WIN, playing the module with fast repetitive Jungle break beats. Other players instead (like all more or less modern players today: ModPlug, foobar2000, Winamp, Sonique, VLC, XMPlay, even PPPlay which was presented here a while ago) played and still play this song with boring occasional single percussion hits.

Using the (Open)ModPlug Tracker, I discovered that the patterns in this song contain a lot of Oxy "Set sample offset" effects, but mostly without playing a new note (see attached screenshot crop).

So I read the documentations about the S3M format I could find. In the documentation provided with ScreamTracker 3.21, no exception was mentioned for applying an effect to a continuing note. Other media Wikis were even briefer. The OpenMPT manual instead mentions a remark:

This effect does not do anything if there is no note in the same pattern cell.

Now, which behaviour is correct?

a) "Millions of flies can't be wrong", and even the final ScreamTracker 3.21 (in DosBox) behaves this way
b) but the author will have had intentions, or he could have omitted these effects

Maybe an early ScreamTracker 3.0 beta behaved differently? ... The truth may be somewhere in between.

The other question is how to "fix" this module to make it sound like it may have been intended to sound, in my humble opinion.

[Saga Musix]

b) but the author will have had intentions, or he could have omitted these effects

Believing that tracker music is an exact science where every effect is placed very carefully with maximum precision is just wrong - you will find tons of similar commands that do nothing in many modules, and the reason is quite rational: convenience. In ScreamTracker 3, the author could simply hold down the "O" key for a few seconds and all notes had the intended offset effect. Most of those O commands didn't do anything, neither something good nor something bad, so the author simply didn't care to remove them. Without knowing that there are buggy players out there which will mis-interpret those commands, the author is unlikely to do anything about them.


Another "proof" that this behaviour is most likely the intended one is that the sample sample plays on channel 4 and 5 of that module - one of them with the offset effect to create a stereo impression. For this stereo impression to work on all note-less rows, the note would have to be repeated on all rows of channel 4 where there is a note-less offset effecto n channel 5. But this is not the case, so you only get machine-gun snare drum hits on the left channel and occasional snare hits on the right channel.

Maybe an early ScreamTracker 3.0 beta behaved differently? ... The truth may be somewhere in between.

The song was saved with the latest version of ScreamTracker, not with 3.0 beta.

I know that sometimes player bugs make a song sound more exciting to some people and they firmly believe that the buggy behaviour is the correct one - but I can assure you that all the "million flies" are correct in this particular case, and mod4win/cubic play the song incorrectly. Unlike e.g. OpenMPT, their playback routines could of course not mature in the last twenty years, so there are many playback bugs in their routines.

The other question is how to "fix" this module to make it sound like it may have been intended to sound, in my humble opinion.

To "fix" the module to sound the way you want it to sound, you need to place the previous note (D#6 in this case) next to those note-less offset commands.