MIDI DIN Jitter?

I am encountering jitter (to some, negligible, but to me, rather unacceptable), when multiple notes are playing on the same channel via MIDI DIN.

This isn’t new to me & my setup, but I was wondering, is it known?

I’ve guessed you mean on Pyramid (given your other posts are about the pyramid) , rather than Hermod.

what is your setup? are you using pyramid as master? where are the notes coming from?
what’s the jitter variation in mSec? how are you measuring it?

really without a lot of details its hard to know.

jitter and latency always exist in digital systems, its all a matter of ‘degree’ , esp over din which is a serial protocol :wink:

perhaps search the forum for others experience,

generally, Id say, 'many’experienced jitter when getting midi over usb from a computer, and discussed on this forum. this is ‘normal’ (unfortunately) due to the way general purpose computers process serial usb (as is used for midi) … again many topics on that subject, so won’t repeat that here.

apart from that, don’t remember seeing much discussion of jitter/latency over midi din.

as you say, perhaps it comes down to acceptability…

if you have specific examples and details (as detailed above, more is needed about your setup, and how you are timing) , perhaps worth reaching out to squarp via the contact form they could tell you if what you are experiencing is within their expectations ( * )


( * ) they are the only ones that know the code, so for example know polling times, and so things like quantisations effects.

2 Likes

MIDI is slow… 3 notes a millisecond or something like that.

Nothing pyramid can do about that except if you’re controlling a turbo supporting Elektron device or using USB

are you sure you’ve only got three notes? how are the synths configured? what else are you doing on the pyramid that you haven’t told us about? Does your salt have iodine in it? who are you man? who?

but really, my point is more like - ‘yea there’s a lot of factors when it comes to midi clock issues’ so, go deep. And the tip on reading other forum posts is solid.

This was posted in the Pyramid category. So obviously it’s not about Hermod.

My setup has my Pyramid’s DIN A going into my intellijel case where it is passively split into Expert Sleepers FH2 and Disting EX. Channels 2-11 control the FH2, channel 12 controls the Disting. FH2 does not take anything from channel 12, so while it does receive notes, it drops them. The Disting is where the jitter is most noticeable. I play drum samples through it, so when transients don’t line up, it’s quite blunt.

If the dev team needs measurements, I’ll sit down and get them. But it’s audible, noticeable, so it’s more than 1-2MS.

it was in “General Discussion” , I moved it into Pyramid … nice try though :wink:
(if you look at the pencil icon next to your post you’ll see my edit)


when you say transients don’t line up … so are we talking about some flam-like effect from multiple samples being played ‘at the same time’
is the disting playing all of these? does it not ‘flam’ if you trigger them manually i.e is it only via midi ?
(I’ll assume, you not aligning with something being driven off a sync’d clock, as thats a whole different can of worms)

its hard to say, as I don’t have the fh-2 or disting, so don’t know if these introduce any factors.

I have triggered the Rample from the pyramid quite a lot, and have not really noticed that having obvious flam - and, as above, Im sure lots of other users sequence drums/percussion from the pyramid, so you’d have thought there would be many complaints if it was creating audible delays.

just to be clear, Im not doubting you… I just don’t really have an explanation on why you are seeing more than (it would appear) others… as seems like a pretty common use-case… and people do care about ‘audible artefacts’.

if you have a scope, then I guess what Id do is look at trig outputs from your fh-2 & disting and see if they really are misaligned, and by how much.
then contact squarp on contact form, the more information/data you can supply to them , the more likely they can help you.

1 Like

okay, just to get the baseline. You’ve got 11 channels (2 through 12) coming out a single bus. Did you size up the amount of traffic you have on those 12 channels?

For example, if you’re sending clock, a whole bunch of CC automation, and note data on those 11 channels – there’s various possibilities for having too much data for either the pyramid or for the device (i’ll explain what I mean). With the pyramid, we know what the theoretical event threshold is for 4.0 variant pyramids, I imagine you already ruled that out.

Now when it comes to a bus of data and a device that only responds to one channel. I don’t know this device, but in the world of midi I’ve seen some really odd behaviors over the years when things get busy. It’s entirely possible there’s ‘too much’ for all of your midi citizens to play nicely on the A bus.

As a simply workaround or just something to try to get your setup to work better, have you considered running the device that’s skipping off of the other hardware midi port on the pyramid?

– if you want to know why I’m suggetsing this, a story –
I used to have everything on Din A as I was building my dawless rig. This included things like my delays, I have two of them. This was also way back in the pyramid 2.0 or maybe even 1.0 days lord, i’m old lol. Anyways, I noticed my delays would have weird jitter issues, the clock packets in this case were sometimes not making it ‘on time enough’ and this was causing them to temporarily assume other bpms and yea, my delays would sound weird. I moved them to a din b based bus, to isolate them from most of my note data. This seemed to me to yield a more stable clock. I don’t mean to assume that 4.0 has the same problems I had back in the day, but if you’re having timing of delivery issues, isolating a device to a bus with less stuff going on might help.

1 Like

Apologies, I was actually certain I posted it in the Pyramid category. The ‘nice try’ didn’t feel too good though. Consider that as a community manager. And please don’t make this conversation about that silly mistake. Let’s continue to the subject.

To your first paragraph after that, yes, it sounds kind of flammy. I have a hard time testing it manually since I seem to struggle to line the notes up myself, maybe it’s a result of that problem too, I don’t know.

Regarding trig outputs, I’m using MIDI to trigger the Disting, so I don’t think testing those will matter.

After testing: There is definitely a delay when using DIN MIDI as compared to USB midi. The delay is approx. 15ms between notes when I play 4 notes at once. In USB MIDI the delay is much shorter than I can test, tbh it seems instantaneous to me.

I tested two input methods (sending MIDI directly to the Disting bypassing the FH2, and sending gate signals from the FH2 into the disting via MIDI into FH2) and for both of those, DIN MIDI had the delay mentioned above. This means the problem stems from the Pyramid’s DIN MIDI and not from the Disting or FH2.

Is this expectable?

Thanks for the reply. While I never (ever in my setup) actually use ALL 12 channels at once, at the very most I use 9, and the jitter was found when I tested with only 4 simultaneous notes (see above), I’ll try rerouting using B only for the drums, and see if that helps.