DIN vs USB MIDI timing on dedicated hardware


I am trying to work out how to best integrate my Pyramid and Hermod with a E-RM multiclock and Mio10 interface.

I’m using either my DAW -> E-RM or the E-RM internal clock as master (when no computer), which then goes into the Mio10 (via DIN) to distribute to the other devices.

It is more convenient to connect both Pyramid and Hermod via USB to the Mio, but I’m wondering if I should expect less accurate MIDI timing then.

I know that USB MIDI from a general purpose computer OS is usually not very reliable, but since all devices involved are hardware with dedicated firmware I am wondering if the same issues exist.

Has anyone experienced less accurate timing when using USB vs DIN on either the Pyramid or Hermod?


I made a quite long post about this before, so I wont repeat the details, you can find them here

main upshot is that with general purpose computers there can be issues with jitter, as they don’t prioritise usb serial traffic (for good reason)

BUT this does/should not apply to hardware devices using things like micro-controllers (like the pyramid/hermod) , in this hardware you have much more control over IO timing and prioritisation.

similarly there is no guarantee that DIN will be better than usb, it may or may not be, comes down to how it’s coded etc.

so imo, id say generally the issue should not exist…( assuming its been programmed well!)

in practice, Ive not noticed an issue (but only had by hermod for a short time), and generally i dont going looking for trouble (so whilst I know how I could measure / test it, im not going to :wink: )

the cool thing about your E-RM, is that you can add latency offsets… ALL midi devices introduce latency, and unfortunately neither hermod nor pyramid have any latency compensation features.
(its not bothered me at all, but for those after really tight sync its useful)


Thanks! I will stick to USB then until I run into noticeable problems :smile: