The problem is that controller data is truncated under certain starting conditions in DP 9.1. When Breath Control is truncated by the bug in DP 9.1, the notes are clipped like a pizzicato effect, which makes it a good test, because it's very obvious when it is failing. It's clear that this is not related to sustain pedals or settings, but is about controllers in general. Since it affects the beginning of any playback, if the conditions are there, it's a very serious bug. I haven't tested yet whether it affects bouncing to disk, but it may well do so since it occurs nearly 100% of the time when the conditions are there, and those conditions may exist at the beginning of disk bounces. We'll explore that later.
The reason this is a difficult bug to figure out is because it involves (at least) FOUR conditions:
• Pre-Existing Data The track contains data, either pre-existing or newly recorded. This never happens in an empty track. It can happen either on playback or when recording OVER data, playing along WITH data, or when overdubbing. There may be more slicing and dicing of this condition, so don't get hung up about it.
• Starting Location (cursor near or within MIDI controller data, vs. a certain distance before it)
• Pre-Gen (On/Off)
• Event Chasing (On/Off)
In addition, when conditions are right to cause truncation of controller data response, the data may intermittently play correctly. It's never a 100% fail situation, as far as I can tell.
When does it play abnormally, and when does it play normally?
- Starting condition: cursor set near or within the MIDI data.
- Pre-Gen ON — Event Chasing ON •••• resulting notes are clipped off probably 98% of the time.
- Pre-Gen ON — Event Chasing OFF •••• resulting notes are clipped off probably 85% of the time.
- Pre-Gen OFF — Event Chasing ON •••• resulting notes play correctly probably 100% of the time
- Pre-Gen OFF — Event Chasing OFF •••• resulting notes play correctly probably 100% of the time
- Starting condition: cursor set some distance before the MIDI data begins.
- Pre-Gen ON — Event Chasing ON •••• resulting notes play correctly probably 100% of the time
- Pre-Gen ON — Event Chasing OFF •••• resulting notes play correctly probably 100% of the time
- Pre-Gen OFF — Event Chasing ON •••• resulting notes play correctly probably 100% of the time
- Pre-Gen OFF — Event Chasing OFF •••• resulting notes play correctly probably 100% of the time
Pre-Gen is the primary trigger of this malfunction.
Starting Condition is a primary trigger
Event Chasing is involved, but it's a secondary trigger.
Existing Data is required to see any of this occur.
I used auto-rewind set to the beginning of the data to test this, and I had it rewinding to the starting point of the MIDI data or a bar before, depending on the test.
Workarounds: Always leave an empty bar before your starting data. When possible, start playback a certain distance before the data that is critical to hear correctly.
This is a test in-progress. There may be more (or less) to it than I've listed here. I've just got to record what I have and stop for now. Maybe a movie would help; I'll consider that after sleeping.
Shooshie
No, I haven't tried that. I have gone from being a smug and self-satisfied "I can't reproduce it" skeptic, to figuring out certain key details to produce the issue, to working with MOTU to get them to confirm and reproduce the bug, all in less than a day. This has been a hectic and tedious 24 hours. I apologize if I have sometimes not been as careful as I would like with all my comments. 
]
