Loomer

gentleclockdivider · Post by **gentleclockdivider** » Fri Jul 05, 2019 3:03 pm

I've noticed that when I record the output of the midi swing module there is difference when compared to the internal swing module of a step sequencer Most noticeable when the last notes are tied
Both swing factors are set to 50 % , and swing base to 16 th
While playing they sound exactly the same
When recording the output into my daw the swing module has some gaps at the end , because tied notes are not tied

colin@loomer · Post by **colin@loomer** » Fri Jul 05, 2019 3:12 pm

It should be the same, as you suggest, so this may be a bug. Swing is based on the absolute position of the transport, so one difference with recording could be that you're playing in time with yourself, but not with the transport. Ideally, if you could send me a patch and MIDI files, I'll try to reproduce the issue and fix it. Thanks.

gentleclockdivider · Post by **gentleclockdivider** » Fri Jul 05, 2019 3:13 pm

First picure , while is might not be visible from the screenshot , the notes are not tied when recording the output ( swing module used )

Second screenshot , the notes are tied correctly when internal swing is used ( and recorded )

gentleclockdivider · Post by **gentleclockdivider** » Fri Jul 05, 2019 3:25 pm

here it is

p.s. when routing the stepsequencer trough the swinger , make sure the stepsequencer swing is set to zero

colin@loomer · Post by **colin@loomer** » Fri Jul 05, 2019 3:29 pm

Great, thanks for uploading these. I'll check them at the tail end of the weekend as I'm currently travelling til Sunday.

gentleclockdivider · Post by **gentleclockdivider** » Fri Jul 05, 2019 3:50 pm

Here's an audio file to
https://app.box.com/s/zkcxzh15tewoqtii81au2fuy3tnl1cst
First you 'll hear 8 beats of internal sequencer swing , sending midi to charlatan vst , at the same time recording the midi output
Then you will hear the recorded midi from the daw

Each example is playing for 2 bars

1. Architect internal monoseq. swing , directly to vst
2.recorded architect midi output ,

3.Architect . swing module directly to vst
4.recorded architect midi output

gentleclockdivider · Post by **gentleclockdivider** » Tue Jul 09, 2019 12:05 am

Any news ?

colin@loomer · Post by **colin@loomer** » Tue Jul 09, 2019 8:05 am

Still looking into it. I hope to have shortly gotten to the bottom of what you're seeing.

colin@loomer · Post by **colin@loomer** » Tue Jul 09, 2019 8:40 am

Right, I've found the problem. It is, as you implied, because of the tie and the way that it interacts with swing. Should be an easy fix. (Oops, I've said the magic words that guarantee it'll take me three days!) Thanks for the example preset and write-up.

gentleclockdivider · Post by **gentleclockdivider** » Tue Jul 09, 2019 10:41 am

NO need to rush it
Keep up the good work

colin@loomer · Post by **colin@loomer** » Tue Jul 09, 2019 10:45 am

OK, I've improved it, but I will leave the issue open in the bug-tracker because I'm not 100% happy with the solution. Essentially, what was happening was that the swinging algorithm was placing two events - the end of the first note and the beginning of the second - at the same tick. This was despite the fact that the end of first note should occur one tick after the beginning of the second, the overlap being what gives it the legato effect.

I've changing the biasing in the MIDI swinger, which does fix yours and many other cases, but isn't a proper solution. The correct way to do this is to rewrite the MIDI swinger module so that it tracks which notes overlapped on input, then aim to preserve this overlap on output. But that way is a lot of work, so I've scheduled it in for a later release. (And the reason why the sequencer swing performed differently is because it does exactly that, being simpler in that case because it know when notes are meant to be tied.)