Patch with MIDI problems (midiout not working?!)

The fact that that note counter could be much simplified is not so important,
but stoping seq recording, before last Note Off gets recorded should be fixed...

Not that it has to do with loss of midi output, which could be anything else

Thank you for your optimized patch.

Are you able to reproduce my problem?

might have something to do with no midiflush after your seq object.
also, if you are just recording midi note events, try using notein (and noteout) rather than midiin /midiout-- there might be program change information being recorded by seq that affects sound-engine output

Thank you! I will edit my patch accordingly.

Does it make a difference for this patch, whether midiflush comes before or after seq?
The second patch in this topic has it inserted before seq and your suggestion is to put it after seq.

purpose of midiflush is to send note off for all
hanging midi notes.
It is absolutely necessary to include it before
seq object.
If You stop seq playback before all notes get turned off,
then You have to insert it also at the seq output,
and bang it when using stop message to seq.

seq expects raw midi messages, so using anything else than
midiin, like suggested notein, would mean prepending
midi note status byte before sending pitch/velocity to seq.
And if You want to keep channel on which notes were sent,
it would get more complicated, one would have to map ch 1
to 144, ch 2 to 145 and so on,
and then packing it into message, respecting order.
note in outputs as most max objects data in right to left order,
status byte is first in the list, so to simplify it, stick with midiin.

P.S. if You suspect that something gets recorded into seq object
that breaks the midi device reception,
just capture the playback into capture object,
and look what's in there, or You could print
first 15 events into max console, or even save
recorded data as text file.

Hi sourceaudio, everything you say makes sense--I haven't used seq in a long time so apologies to readers for the misinformation-- but in the seq help file there is a midiflush after the seq object, not before, so maybe that's a factor.
I think I was confusing seq with detonate-- so Timo you may want to explore that as an alternative to seq, if you're just storing note events

here's an adaptation of the detonate snippet file:

it might be worth testing this against your other patch at least for diagnostics

It is allways difficult to suggest best solution to a problem,
if one does not know the details.
I am wondering - why this 5 notes limit ?
Maybe something else than seq would be a much simpler solution,
but one can't think so far without knowing what really is the goal of the whole
exercise.
Beside possible problems that could arrise due to misuse of seq, midi data etc, there were problems reported with midi flow and system based virtual synths,
on both platforms, both mac and win, so one has to at least eliminate
the part with capturing and playback of midi data, before looking into other things.