I was just trying some "loop audio recording" and found that if I have the metronome on with a two measure count in on record, each recorded take is offset from the original scratch track.
 
Here's the scenario:

• set the metronome to have a two measure "count-in on record"
• set the record button for "comping"
• create an audio track and arm it
• set the now time to measure 3
• press the record button to begin recording (the metronome should play a 2 measure count in before the track begins to roll)
• record a scratch track for a few measures and stop recording after a few phrases
• disarm the track you just recorded
• set the loop points to start at measure three and end at a logical point in the scratch track so it seamlessly loops to the beginning of the scratch track starting at measure 3
• create another audio track, select and arm it for record
• ensure the now time is back at the beginning of the loop (should be measure 3)
• press the record button to begin recording on the newly created audio track (the metronome should count in for 2 measures before the tracks begin to roll)
• play (in time) to the scratch track, recording to the newly created audio track for 4 loops and then stop recording
• show take lanes on the newly recorded track
• select the take lanes to play each one and note that each take lane is "out of sync" with the scratch track

If I turn off the metronome "count-in on record", set the count-in measures to 0 and perform the steps above, the recorded take lanes are in sync with the scratch track.
 
Due to the observed change in behavior when I turn off the metronome and count-in, it seems like this could be an issue while loop recording with the metronome on and count-in measures set.
 
I'm wondering though, could this be a problem with my setup or has anyone else had this issue?
 
Incidentally, I'm running Sonar Producer X3e on an Intel i7 4th Gen (2.7 GHz) under Windows 7 Professional 64bit with ASIO drivers connected to a Roland Octa-Capture at 48kHz/24bit.  My typical round trip latency is about 6ms (289 samples).
 
Thanks in for any input in advance,
Paul

Others have had this issue, including myself, but I don't recall anyone providing a recipe with sufficient detail to reproduce (nor has it happened to me consistently enough to figure it out). I will verify your steps, see if I can add any extra detail (like whether it still happens with a 1 measure count-in), and make sure the Bakers are aware of this thread. Thank you!

Thanks for the report, we'll look into this.  It's great when detailed recipes are provided.
 

Keith Albright [Cakewalk]
Thanks for the report, we'll look into this.  It's great when detailed recipes are provided.

Thank you Keith, and Paul, thanks again.

Anderton and Keith Albright [Cakewalk]:
 
Actually, thank you!  Your attention to my post has far exceeded my expectations!
 
Paul

Providing a reproducible recipe for an annoying bug is a great way to get attention 
 
Just an FYI for the future - entering something into the bug tracking system is a faster and more efficient way to log a bug, because there's no guarantee anyone in a position to do something about a bug will see a forum post. But I've been trying to find a way to reproduce this without success, so I immediately noticed your thread.
 
Everyone at Cakewalk is committed to producing as stable a piece of software as possible, so having the community's assistance is invaluable.