I have just begun paying attention to STATISTICS under the AUDIO menu, and after some investigation of current projects I have discovered this:
Original recorded audio files are normal - no DC offset and at 24-bit resolution. What appears to be happening is if I split the full-length event into much smaller events and move them around (for example, rearranging a recorded take of hand percussion by moving single hits and or short rhythmic phrases to new locations), then those small relocated events may have developed either or both DC offset and reduced bit resolution, according to Statistics.
The DC offset is usually less than 1% and can be either a positive or negative value, but is sometimes more than 1%. The resolution may or may not be reduced, to 23 or 22 bits.
My research is telling me that DC offset is introduced through hardware issues. But the affected audio is being processed only within Cubase - just sliced and moved, not rendered, frozen, exported and imported, bounced through a UAD card or anything like that. It has only gone through the Events to Part process.
Anyone know what’s going on? This does tend to shake one’s confidence the vaunted Cubase audio engine.
“Don’t worry about” it is certainly nice to hear. I got started on this because I found DC offset in a “final” stereo mixdown track, and then determined that DC offset in a project track causes it in the mixdown track. When a track with DC offset goes to mastering, I am told, the the asymmetry of the wave form causes it to clip earlier than it otherwise would, by the amount of the offset.
No, the offset and bit reduction are not audible to me. After years of working in Cubase, I just happened to notice Statistics on the Audio menu, and checked it out…and lo and behold, there was DC offset on that mixdown track, and that got my attention. I do in general want to understand what’s going on in my recordings, and Cubase doing this to the audio seems pretty odd and undesirable.
What I was actually looking for when I checked out Statistics was an analysis utility that would flag the highest peaks on a track. I moved to DAW-based recording in the mid-90s, using MTU/Microsound - a pretty great proprietary hardware and software system now long fallen by the wayside. The software had this peak-flagging utility. Run it on a stereo mixdown, get those flags, then revisit the mix and then make your own choices about how to ameliorate those max-peaks with manual gain reduction on one or more individual tracks in just those particular spots.