Pitch of single steps changes randomly depending on programmed pattern

Trialanderror

I have an issue with my Sonic Potions LXR I don't understand. Sometimes the pitch of single drum hits changes depending on the programmed pattern. As I change the pattern, the position of the wrong pitched step(s) changes too. Sorry for the bad video quality, I used my smartphone. But I think you can see and hear what I mean. 

The problem disappears if filter cutoff is set to 127. Only the filter changes pitch, not the OSC. Not only the kick is affected, this happens to other drums (and with other kits), too. Firmware is 0.32 official, also tried out the current Rudeog firmware as well as the official 0.23. What is happening here?

Can anyone reproduce this behaviour (most audible if click and overdrive are set to zero)? Is this normal?

I found this thread: http://forum.sonic-potions.com/discussion/247/random-click-on-drum-1-bassdrum - could it possibly still be the same glitch?

Used settings for the bass drum in this video:

OSC
coa 25
fin 0
wav Sin

AEG
atk 1

dec 8
slp 7

MOD
dec 27
slp 27
mod 41
DST mod
amt 0
vol off

FM
amt 0 
frq 24
wav Sin
mod fm

CLICK
wav Ofs
vol 0
frq 0

FILTER
frq 27
res 0
typ LP
drv 0

LFO
frq 0
snc off
mod 0
wav tri
rtg off
ofs 0
voi 2
dst off

MIX
vol 127
pan 0
srt 127
drv 0
out St1
len 16
ch 0
nte any

erwanone

Yeah i have the same thing very eareable when you set 1/5/11/12 steps but not at all on 1/5/9/15 . never heard that before so maybe its normal and thats the settings of the kick that are not logicall when you triggeer iit 2 two near each other. I try with other kick kit and sometimes you can hear it sometimes not, try change decay of mod and aeg maybe.

erwanone

And dont know if its filter realated as you can hear some click with the filter set to 127 and when filter off. But its more earable with LP low filter cutoff. really weird i never noticed that before need to check more on other drum. ANd yeah i dont remember the thread on the click i think its the same issue.

Trialanderror

Already tried turning down EG and MOD values. The problem remains.

Revok

Are you perhaps modulating it with the lfo of another voice?

pierre

Those are settings for the voice, I'm still new experimenting with this DM  but I think what takes care of the relative pitch between each step is the note parameter in the step editing If you do shift + hit on a step button, you see VEL then nte on the first page, if you want all the step the same pitch you need to make sure those settings are  the same for each of the steps that are playing. Hope that helps

pierre

And If it's relative to the cutoff, as you mention the Vel value is probably modulating the cutoff, but then it wouldn't change the pitch, but the tone, but for drum hits that are not really harmonically tuned that could be perceived as pitch change. The bottom line is for each step, if all the step edit parameters (2 pages) are exactly the same, each steps should sound the same, I think.

Trialanderror

Double checked all LFOs - none affects voice one. Checked all note parameters, too: everything identical on each step. The odd thing is that the position of the "wrong pitched hits" changes as I change the pattern. When I program a typical 4/4 (1, 5, 9, 13) the problem never occurs. 

erwanone

Yeah i checked all that too. Pierre if you try it too can you hear the same problem that is on the video? With this kick setting and a 1/5/11/12 pattern its pretty obvious. i always start with a blank kit/Pattern

pierre

Sorry I didn't watch the video first you're right it does that here too but I think I found what it is, it's not the pitch of the drum, it's the transient Wav, when you chose something else than OFS it act normally here, I don't know what ofs stand for but maybe it's meant to do just that, to get a bit more of a dynamic voice, I fought of checking there cause to me it doesn't sound like a pitch issue, it's more that the click is rather there or not. Let me know if you experience the same thing

pierre

Actually it's not that, sorry by turning to zero the transient i realized I was mistaken, The sample there at 127 was just masking the attack and wasn't hearing the issue, but unfortunately it's a bit more random...

Trialanderror

Thanks for trying to figure it out mates. Quite mysterious, is it? I really don't get what is going on there. 

pierre

Yeah in all honesty on my part I can live with that, I only really notice it on really round bass drums, and as mentioned before I think for one it's filter related. The manual states that it's a non linear phase filter so they'll always create some phasing and some stability issues. Implementing a Finite response filter chew up a lot of ressources, and probably just not possible here, especially if we want the degree of flexibility we have here...Aside from that, there is also a bit of a clicking issue, this on the other end should be fixable. One would need to really understand the sound engine but surely something to do with zero crossing or retriggering, but hey the beauty of this synth is that it's a work in progress and it will get better and better. When I have time I'll be happy to dig in the firmware and try to contribute as well but those who do are already doing a great job.

So I say, for now either a low sin type kick with no filtering, or the use of a transient sample to mask most of the clicking is good enough for me, for now

julian

I think i have to try to modify the filter algorithm.
I did not noticed this 'pitch change' before, but a variation in the transients of the sound (clicks on the beginning)
It also seems to be tempo related somehow, which amkes absolutely no sense [-O< :-?

Maybe I'll keep this one as 'animated filter' and add another more conservative filter algo who does not expose this behaviour. The current model is a zero-delay-feedback filter. so maybe it is a numerical problem related to the FPU on the cortex that can't compute in doubles.

ceasless

I think this is the issue that's been plaguing me! I love round bass kicks, so that explains why I see it so often.

julian

do you need the filter for those?
if you take a sine as waveform the filter won't have that much effect.
as a workaround the problem should vanish once you set the filter to off.

ceasless

I will test that, thanks Julian.