Additional bytes creeping in

Bug reports
1

I have some weird behaviour. Anyone knows what to look for?

I’m reverse engineering PWM Mantis’s Sysex Patch dump. Ben Supper of PWM was so kind to get me started (thank you, Ben).

The patch dump is always 116 bytes long, as seen on the Midi-Ox screenshot:

image
image482×459 10.9 KB

Hereby is “A03” the RTP MIDI port of the MioXL where the E1 sends to, and “A02” is the RTP MIDI port of the MioXL where the Mantis sends to, so I can capture both on the PC via MIDI-Ox. Within the MioXL , the E1 and the Mantis are interconnected.
Nothing gets sent out from the PC back to the MioXL.

If I then monitor the incoming SysEx on the E1, I get to see varying lengths of sometimes 117, sometimes 118 bytes. I was expecting 116 of course !

there’s this hex byte of F8h (see screenshot) that seems to sneek in on arbitrary places increasing the response value with 1 or 2 bytes. Needless to say it renders the parsing exercise tremendously difficult.

image
image722×856 34 KB

So the Mantis is not sending that byte, but the E1 seems to get it anyways…

Anyone?

2

Edit : F8h is the MIDI timing clock. Some device must be sending it and the E1 picks up about one or two of these clocks during the monitoring of each 116 byte SysEx message.

@Martin, it looks like the SysEx monitoring is also picking up these undesired MIDI controls signals apart from the real SysEx data and F0h or F7h.
For now I’ll filter out F8h, so I can continue parsing.

1 Like
3

I am almost done with the next firmware release. I will try to squeeze the fix for this issue in. I do think it was reported already in the past.

1 Like