Sonar Ipswich-Lexington Users – “Remove DC Offset During Record” - Head's Up

Edit 12/30/2015: Changed title to reflect problem still exists in Lexington (tested by WalkerTalker - I have not yet loaded Lexington). Also see post #9 for recipe.

Hello All!
 
I recently came across a bizarre problem in Sonar after buying and attempting to record a stereo pair of mic’s as dual mono tracks.  I was recording fingerstyle guitar and found an inexplicable unnatural bit of distortion/buzzing upon playback, but not while monitoring during recording process (using input echo for monitoring).  During my tedious three days troubleshooting I finally stumbled across a solution when finally, grasping at straws, I started changing settings that you’d think would have nothing to do with the issue.
 
The final answer turned out to be: turn off “Remove DC Offset During Record” in Preferences> Audio>Playback and Recording. 
 
During my troubleshooting I found that the issue only occurs when recording two (or more?) tracks simultaneously in Sonar Ipswich through Kingston (I happen to have Platinum).  Recording one track at a time in Ipswich through Kingston or dual mono tracks in Sonar X3 through Hopkinton work fine.
 
In ASIO mode the issue manifests itself as a strange distortion that’s sort of buzzy in a way.  In MME mode, it shows up as one or more decaying amplitude clicks after a louder transient in the material (even after you increase the audio buffer size to get rid of the “clicky things” that happen in MME with the buffer too small).
 
If you’ve experienced these manifestations, try turning off “Remove DC Offset During Record”.
 
I know at least one other person has experienced this issue, as I noticed a post on the forum yesterday that “sounded mighty familiar” and I suggested turning off the DC offset removal, and that fixed his problem.
 
I imagine other people may have fought and still are fighting this problem, and I thought I should post this save others from some grief and tedious troubleshooting.
 
Cheers,
Pete

Yep, sounds like something I've been battling with this weekend. A buzzy resonance particularly noticeable on percussive tracks, in projects where I've recorded eight tracks simultaneously. Except I'm on Jamaica?
 
Now I would welcome any suggestions for ameliorating this. Sometimes a severe low pass filter helps, fine for bass tracks, but not the others.
 
John

I have just added a Tascam US-16x08 to record drums (someday).
No opportunity at present to try reproduce this, but I would be very
interested in hearing more.

WalkerTalker
In the meantime, stuck with this noise on a couple of tracks, I wonder if you or any other forum member has found a way a way to remove/minimize it?

I found that for me at least, the "noise" was usually around 1KHz (the fundamental at least).  The noise looks like it's actually very similar to a sawtooth on top of the signal.  You can try a very steep (High Q) filter and sweep it around 1KHz to see if it helps, but I have a feeling that anything that lessens the already recorded noise will pretty much screw up the program material a bit too.  I think the only real option is to re-record with the "Remove DC Offset" disabled, though it may not be an option for you.

mudgel
I have followed the recipe strictly and can't duplicate your results.
There must be something else going on.

Thanks Mike!  I don't doubt there's more involved.  It's probably not very widespread either or we would have heard a lot of screaming.
 
If anyone else would like to try, here's the steps from CWBRN 42851.  Rather than recording guitar, which would have made the testing 6 days instead of 3, I found that you can quickly and easily hear what's going on by just tapping the body of the mic's with your finger.  You may not hear the artifacts tapping at low input gains, but if you adjust your record levels to get the peaks of the tapping in the -10dB to -5dB range, it should be obvious.  Make sure you don't clip, of course.

1. Make sure Preferences>Audio>Playback and Recording>Remove DC Offset During Record is set
2. Starting from a blank new project, insert two audio tracks
3. Route Track 1 input to Audio Interface In L and Track 2 Input to Audio Interface R
4. Route both track audio outputs directly to Audio Interface Stereo Out
5. Using a microphone on each Audio Interface Input, adjust input levels such that tapping the body of the mic's produces a high input, but NOT CLIPPING
6. Record enable both tracks, enable input echo, and monitor the recording process with headphones (set Audio Interface for NO direct output, Sonar output only)
7. Press record and gently using a finger, tap the body of one mic several times, and then the other (all that is heard is the "thud" you'd expect)
8. Stop recording
9. Play back the recording (in addition to the "thud" you will hear a roughly 1KHz ringing after each thud in ASIO mode, and one or more decreasing amplitude clicks in MME mode)
10. Repeat the recording, but with only one track enabled for recording
11. Play back (the results will be fine)
12. Repeat the recording with only Track 2 enabled for recording (results will be fine)
13. Repeat with BOTH tracks enabled for recording (the results will again contain the artifacts)

I have some jpgegs of the waveforms and clips of the artifacts.  I guess I'll have to figure out photobucket and soundcloud.
 
Also, I'd like to add that when I said everything works in X3 through Hopkinton, and the issue only occurs in Ipswich through Kingston, I forgot to say that statement refers to ASIO mode.  Using MME, only Kingston was tested - I don't know what happens in any release prior to Kingston when using MME mode.
 
Pete

John
Also one should only use this if there is a known problem with DC offset with your audio interface. Otherwise it should not be used. 

Yeah, that's a good point, John.  Your interface should not be putting out any DC component.  I always just had it on for "good measure" ... figuring what harm could it possibly do?  (I gotta stop figuring ...)

Hi Pete,
 
I was not able to reproduce this until I saw that the recipe involved recording both sides of a stereo input pair to mono tracks simultaneously. Bingo, I can reproduce that. I did not use mics or even an analog input; it's reproducible with ADAT input from a hardware synth. And, as you pointed out, it's not in the signal while it's being recorded; it appears to be getting generated by the DC Offset removal algorithm as the data are written to file(s).
 
I'm not sure why other's can't reproduce it since it seems like a DSP problem that's all "inside the box", but I think your detailed recipe should be good enough for the Bakers to get to the bottom of it. Good one.

I recently had the same problem too.  I never used the 'remove offset' switch in the past, until chasing a different issue when I found it again and thought it wouldn't hurt turning it on.  I recorded 16 simultaneous tracks of acoustic drums, and it sounded awful.  I heard a bad distortion on the two kick drum tracks (inside mic and subkick) that sounded an awful lot like digital clipping or maybe a sketchy cable or bad mic.  After switching snake channels, mic cables, channel strips and audio interfaces, I finally traced it down to Sonar.  In thinking back to the changes I recently made, I remembered turning that feature on.  For giggles I went back, unchecked it, and test recorded again.  Perfect!  FYI, more than just the audible distortion, it also affects the overall sound.  Kinda like having a compressor set completely wrong....it's hard to explain.  But it was really noticeable on the kick drum.  I'm running Platinum Kingston 64-bit and everything is back good again.   

WalkerTalker
Just tested this in Lexington. Still a problem, FYI.
Pete, any response from Cakewalk over on the bug report system?

The current status is "Submitted to Development", which happened on the 23rd of December. I haven't heard anything beyond that.