It appears what you heard is what is referred to as Absolute Polarity or the Wood Effect. Clark Johnsen wrote the definitive book about it although Bob Katz covers it in his Audio Mastering book. The idea is that we hear sounds differently if the initial transient is positive going vs negative (e.g. blowing vs sucking on a horn). A search on the above terms will uncork lots of information about it. And controversy…

Many claim that what you heard is not real. Very disconcerting when you have experienced it first hand. For example, I participated in some listening tests at JBL/Harmon as part of AES. I discussed Absolute Polarity with their head of R&D, a much published audio PhD. He stated they looked into it and could not verify what you heard either on paper or by abx listening tests. But you did hear it very clearly as have countless others. As had all the frustrated AES members who were there listening!

While recording in Sonar, try zooming in on the initial major transient of each track, especially drums. Don’t be surprised to find that many go negative. Invert them so they all go positive and listen. You will find, as many have, that it makes a significant difference. But how then do tracks end up “upside down” in the first place when your signal chain is wired correctly and invariant? Because there is no standard for absolute polarity in individual components that get placed in and out of the signal chain (and AES won’t address it as the above mentioned author Clark Johnsen will tell you).

As a result many pieces of hardware, software, even guitars have an inadvertent initial negative going initial transient or inadvertently flip the signal during processing. Why? Because again there is little awareness or belief that it can make a difference. Engineers deal with the signal “in the wire or code” (where it doesn’t make a difference) and aren’t concerned whether an initial transient pushes your speaker cone in or out (ear pressure vs vacuum, the basis of the Wood Effect). The dbx Quantum Mastering processor used to flip the signal. When asked about it their engineers claimed it didn’t matter (it has since been fixed). I’ve used a Roland drum machine where the individual pad audio samples were essential random one direction or the other. But when you put them all positive going after recording each track separately, it’s like night and day. More punch and openess regardless of what the researchers say. I have a stock Rickenbacker 4003 bass that always records upside down and needs to be flipped.

Yes, this is one of those simple things that gets worse the more you look into it. But at least the fix is easy!

You have experienced something that affects all recording but is generally overlooked. Perhaps you will be intrigued enough to look into it further. For example, a simple Sonar track “initial transient polarity check” would be invaluable rather than having to zoom in and out of each track and manually adjust the polarity as required!

OK, enough from me. Great website and great software (I started with Cakewalk back in the DOS days). All the best.

Don