MIDI - fixedInputVelocity applied to every event

[gi0e5b06]

MidiPort.cpp lines 131-148
The following code should be moved inside the test just above. FixedInputVelocity should only be applied to relevant events.

if( fixedInputVelocity() >= 0 && inEvent.velocity() > 0 )
{ inEvent.setVelocity( fixedInputVelocity() ); }

[zonkmachine]

Hi!

Well spotted! This was actually fixed in stable-1.2 in this commit less than a week ago: #3678
We're working on merging back commits from stables-1.2 to master so you'll see it in master in a day or two.

PS. If you press/select the line number on code in github, you can copy/paste the link and send people directly to the issue. Here's what the fixed code looks like: https://github.com/LMMS/lmms/blob/d65e1a361a60498f2371accf3b50b32ff834ab2f/src/core/midi/MidiPort.cpp#L117:L141

[gi0e5b06]

Actually I think the changes on velocity shouldn't be implemented there, as it modifies midi messages that you want to receive unchanged. Also, in the same function, there is a test that discards Note* messages if the key is not standard (0-88). But on a lot of hardware, buttons are keys with a higher number. I had to remove that test to use the buttons on my midimix.
Velocity control is needed somewhere, just not here.

void MidiPort::processInEvent( const MidiEvent& event, const MidiTime& time )
{
	// mask event
	if( isInputEnabled() &&
		( inputChannel() == 0 || inputChannel()-1 == event.channel() ) )
	{
		m_midiEventProcessor->processInEvent( event, time );
	}
}

[PhysSong]

I had to remove that test to use the buttons on my midimix.

Are there any problems with the test? Could you please explain more?

[gi0e5b06]

@PhysSong The buttons of (all) my hardware midi contollers generate NoteOn/NoteOff events. These events were discarded by the test so they never reach the controllers in LMMS (mute for example).