I realize this is a difficult thing to explain, so I first want to make sure I’m being clear about the behavior I’m describing. Here’s the exact use case that has inspired me to finally post about this. I’m adding “Stretchin’ Out” by Gayle Adams to my DJ library and because the BPM is all over the place on this old record, I am manually beat gridding it. I’m about 3/4 done. So let’s say I insert an anchor on bar 127. Moving forward with the remainder of the track, the BPM is set to 112.03. I get to bar 131 and insert a new anchor point. It turns out the actual BPM of this four bar stretch was 111.65.
Here’s the issue: adding this anchor point has now changed Denon’s BPM for the remainder of the song to 112.05. In other words, telling the software that the previous section of the song was slower than expected is causing the software to move the BPM further in the wrong direction.
I can sorta begin to imagine what kinda coding would lead to this behavior, but this is a clear bug in my view that makes beat gridding more difficult for obvious reasons. I think the ideal behavior would be for the software to assume that the BPM implied by the most recent anchor points carries over for the rest of the track.