Sync issue when switching patterns

MIDIbox SEQ
1

Hi,

 

I have a weird issue with SONG and PHRASE mode on one song/session.

 

Basically this song is composed of 3 parts: Verse, Bridge and Break. 14 tracks are running, some with different dividers and lengths, but they all show-up nicely synchronized on the TPD display (the track lengths are set so that they all restart at step 1 at the same time without being synced to measure). The 4 groups of tracks have each 3 patterns named Verse, Bridge and Break.

 

I’ve set up the SONG so that the sequencer plays endlessly the 3 patterns one after the other. It works, but the sequencer “hicks up” when switching from 4 patterns to the next 4 patterns; it kinda “swallows” the last notes, and when switching from the third pattern (Break) to the first (Verse), certain tracks go out of sync.

 

The same thing happens if I’m in PHRASE mode and switching the patterns manually.

 

I tried with and without a guide track of 64 steps, with and without all tracks synced to measure, but still no luck.

 

The session settings are the following:

—Option 1 (Track Sync, Steps per Measure) = 64 steps.

—Option 2 (Pattern Change Sync, Change considered each) = 64 steps

—Option 3 (Pattern Change Sync) = enabled

—Option 5 (Restart all tracks on pattern change) = disabled (also tried enabled).

—All tracks not synced to measure (also tried in sync).

—All tracks have ReSt. off (also tried on).

 

I’ve been working on this track for a few days, and the weird thing is that at some point I know that the pattern switching behavior was working right, without any sequencer hick-ups, but I can’t find when those sync problems occured…

 

Any hint of what could be wrong, or may I have encountered a bug?

 

Thanks!

2

i am not 100% entirely sure, but i think a pattern switch reloads the pattern from SD-card. Thus the question: which SD card type are you using? Could you try with a different card that is formatted to FAT32? Recommend to try any bigger/smaller/faster/slower/SDHC/SDXC cards you may have lying around, to freshly format them and to transfer your session over and test it…

In general, if SD card access has some kind of latency, this could explain things getting out of sync.

Many greets and stay safe!
Peter

3

Hi Peter, thanks for the reply, I’m currently running the SEQ using a SanDisk SDHC 4GB card, but I’ll try with other SD cards and report back.

 

Greets,

 

Kevin

4

So I’ve tried four SD cards with no luck. But now I’m quite sure of what’s going on, but don’t see a workaround as it looks like a bug.

I slowed down the tempo of my song to 20bpm with Follow Mode ON and that allowed me to see more clearly what is happening. I ran several tests with just one track playing at a time, and they all share the same behavior, syncing the tracks to measure or not didn’t change the result.

Here’s a more detailed explanation:
—SONG mode
—Pos.A1 = x1 1:A1 (Verse) = 64 steps
—Pos.A2 = x1 1:A3 (Bridge) = 64 steps
—Pos.A3 = x1 1:A2 (Break) = 64 steps
—Pos.A4 = Jump to Pos.A1
—Steps per measure = 64 steps
—Change considered each = 64 steps
—Pattern change synchronization disabled or enabled (same result)

What’s happening is that when the SEQ reaches step 64 of Pos.A1, which is the last step of pattern A1, it plays instead the last step (step 64) of the next pattern (A3 in my case). And then it plays the first step of pattern A3 as expected. The problem is the same if I’m in PHRASE mode and switch manually from a pattern to another.

As said in my first post, I tried using guide tracks but that didn’t help. I’ll try to reproduce the problem on a simpler session and share it here, I can also make a quick video if needed.

All best,

Kevin

5

I just uploaded two videos showing the problem on my actual song. On the first video, pay attention to the TDP when the patterns are switched around 0:16 and 0:32:

 

And a second one with just the Organ lead sound, with the tempo slowed down and Follow Mode activated. At 0:42 we can see that the last step of the second pattern plays instead of the last step of the first pattern:

 

I managed to reproduce the problem in a super simple session which is attached in a zip file in this post. The issue seems to happen when two tracks of a same pattern don’t share the same length and/or division.

The session is composed of two tracks and two patterns.
Pattern A1:
—Track 1: Kick — 16 steps long, clockdivider 16 (normal)
—Track 2: Hat — 24 steps long, clockdivider 16T (triplet)
Pattern A2:
—Track 1: Kick —16 steps long, clockdivider 16 (normal)
—Track 2: Hat — 24 steps long, clockdivider 16T (triplet)

It’s a simple drum beat, and there’s only one open hi-hat note (D#3) at the very end of the pattern A2 (on the 24th last step). If you play the SONG (with a slow tempo and Follow Mode activated, which makes it easier to see), you will notice that this D#3 note plays instead of last step of pattern A1, so before the pattern A2 is supposed to be played.

Once again, I ran the test using both SONG and PHRASE mode, and results are the same.

Cheers,

Kevin

 

BUGREPRO.zip

6

good that you could reproduce and document it, hopefully could have a look at it - if not, you could repost the .zip and the video in the main SEQ v4/v4+ dev and feedback thread over here: ?

< base_url >/topic/13137-midibox-seq-v4-release-feedback/?do=embed

Many greets!
Peter

7

Sure thing, good idea Peter :slight_smile: