hi,
i have a problem with a faulty sequencer which sends out duplicate messages. can i use midihub to setup a filter to delete ‘repeat’ notes and commands? thank you
This is tricky. I don’t see how MidiHub would have a way of knowing what input you want to keep and what you want to discard. You can filter types of data but then you are blocking all messages of that data type.
You could drop inputs that come in before a certain interval and after the first message is received, but then your input would have to be relatively quantized and consistent to act reliably.
Honestly more specific information of what you are trying to do is needed, like what messages are offending, how much of it do you want to drop? Are you playing back predefined sequences/arps or are you playing live? Is this a modular device that’s generating midi? Are your messages doubled and on the same tick or being sent milliseconds after the first?
It’s better to solve the issue at the source if possible. Why is your sequencer sending out unexpected messages? Are you using an arp that’s layering data in overdub mode?
Are you having contact issues with your pad or key sensors? Bad cables? Feedback Loops? If your device is not working as intended, even if you can craft a band aid now, what happens when that device further degrades and behavior changes?
Whatever the scenario, it’s probably prudent to ID the cause before trying to craft a fix for it.
1 Like
Agreed.
Seems a strange fault for a sequencer to have:
Ask around on forums/reddit for that sequencer?
If it does need to be solved ‘outside the box’, it sounds more like the sort of thing you’d write a computer/rPi script for (e.g, python.mido)
it is a strange fault. i was thinking out of the box to try and solve it but i’m not sure, with consideration it is possible with midi filtering. Because even if i solve the ‘output’ issue the data will be wrong inside the sequencer.
The fault i have tracked down to a UART chip with a mask rom, on that mask rom you can specify the ‘stop length bit’ someone at Roland got it wrong and an entire batch, if not the whole batch will be wrong. I cannot trace anyone mentioning this problem before but this is the MV8800 and its user base may not typically use keyboards.
I need to obtain the original firmware for the mask rom and then come up with alternative. The uart is based upon the 65C02 but it has a mask rom area - it is the Mitsubishi M38881M2