I messed around with it some more and I found out a few specific settings you need to replicate the bug:

The release slider must be set exactly on notch #5 (counting upward from the bottom). Setting it any higher or lower alleviates the bug and the voices start stealing normally. But having it exactly on notch #5 makes it start bugging out. This is the thing I found oddest of all.

The attack slider starts the bug around notch #5, but starts getting worse as you move up to notch #6 or higher. Having the attack lower than notch #5 alleviates the bug.

The envelope destination FEG/AEG slider must be set at notch #4 or lower. pushing it above notch #4 and making the envelope start modulating the cutoff alleviates the bug. The bug it at its worst when this slider is all the way down to the 100% AEG position, where the envelope is only modulating the amp.

The bug starts happening when I have the looper tempo slider about halfway between notches 8 and 9, with the bug getting worse as I increase the tempo even further.

None of the other sliders seem to have any effect, not even the decay or sustain. it all depends on the attack, release, envelope destination balance, and looper tempo.

Attack, on the Amplitude Envelope Generator, is a parameter that determines how a sound begins.
If the Attack (A) setting from the EG section is too high you will not hear an immediate attack. It does not matter that you continue to trigger notes, ATTACK is a Time paramaeter. As you raise Attack the amount of time between pressing or triggering a Key and the sound beginning is lengthened.

This is a synth basic.

Release is the parameter that determine how a sound disappears after Key-Off. If the Release (R) setting from the EG section is set too high, you will be unable to retrigger the Tone Generator until its TIME value is reached.

Attack and Release work together to determine how a sound behaves. Keep experimenting... you’ll find their is no bug, this is normal operation for partially any synthesizer on the planet.

Turn OFF the Looper so you can control when a Note is triggered manually.
Set Attack all the way up ... press a note and wait for the response. The higher you raise it the longer TIME it takes for the amplitude (loudness) to occur. This is how you slow the Attack time. You will ultimate reach a point where 8 note-ons are triggered before all 8 notes get to sound...that is called ‘being out of polyphony’. A natural occurrence you can avoid by recognizing the Polyphony limit.

Now work with just the Release... as you raise the Release the time it takes for the notes to disappear is lengthened. If you set this so that it hold multiple notes, it will run the instrument out of polyphony when the maximum number of notes is reached. Again a natural occurrence you can avoid by understanding how it works.

Again basic synth stuff. Thanks for the report but we think there is no issue — certainly, no Bug.

Thank you for your reply, but that doesn't seem correct.

As you can hear in my youtube video of the bug, the synth starts dropping random notes altogether, while still sounding others.

https://www.youtube.com/watch?v=lsGGl42H6RY

If it were just a matter of the attack or release setting, all the notes in the sequence would be equally effected, but in this case, only some notes get effected, while others don't, and which notes seem to get dropped from the sequence seems to be selected by the synth at random? New notes in the sequence do not seem to be triggering consistently or properly.

If the Release (R) setting from the EG section is set too high, you will be unable to retrigger the Tone Generator until its TIME value is reached.

this is not true on the Reface CS. if the polyphony mode is set to mono, the voice will always retrigger upon hitting the note a second time, even if the release has not been given time to fully finish. on polyphonic mode, hitting the note again also immediately triggers it (using up a new voice in order to do so). if the release is set high enough so that 8 notes end up on their release phase at the same time, the next note pressed will mute and "voice steal" from the oldest currently-sounding voice.

Also if this were just an issue of the release slider being set too high, increasing the release slider above notch #5 would make this bug worse, not better, like it seems to do on the Reface CS?

You will ultimate reach a point where 8 note-ons are triggered before all 8 notes get to sound...that is called ‘being out of polyphony’

again, not the case here. there are only, at most, two notes being simultaneously pressed in this looped sequence (for a very brief moment as one note switches to the next). this is not an issue of the CS running out of polyphony due to more than 8 notes attempting to be simultaneously held.

hi steve. Did this bug ever get fixed or do you still have it?