swamptooth
Are you using the phones out or the outs on the breakout cable?

Breakout cables can be prone to crosstalk if they terminate in a connector where the pins are really close together...

Anderton

swamptooth
Are you using the phones out or the outs on the breakout cable?

Breakout cables can be prone to crosstalk if they terminate in a connector where the pins are really close together...

That's what I was thinking as well.
At least it's not as bad as how my guitar kept picking up a spanish radio station a couple of weeks ago...

Getting bleed through the phones and the breakout cables...agreed that it doesn't appear to be SONAR-related.  Not sure what my next course of action is but I appreciate the research.

Anderton
As I mentioned previously, crosstalk specs of -55 to -60 dB are not uncommon. You can hear tape hiss easily at -60 dB and -55 dB is about the S/N of AM radio stations. With a hot click level, you could definitely hear it in the other channel if the levels were turned up...probably even if the crosstalk was -65 or maybe even -70 dB.

RME's specs for the analogue outs on the Babyface claim the same amount of channel seperation as their other USB/Firewire interfaces, i.e. greater then 110dB. I've just tried generating cross-talk in my UFX and I can't anything audible out of the right hand side with the left channel red-lined with white noise. maybe a few tiny flicks of the meter, that's it. The issue here could be a fault on the Babyface. Personally I'd be inclined to take it up with RME support to see what they say.

tlw
RME's specs for the analogue outs on the Babyface claim the same amount of channel separation as their other USB/Firewire interfaces, i.e. greater then 110dB.

It depends whether that's across the entire frequency range, at 1 kHz, etc. I've tested interfaces that specify X amount of crosstalk, and they meet that spec but only over a certain range of frequencies or at certain mic preamp levels. The only real way to determine crosstalk is to see a graph that shows the amount of crosstalk across all frequencies for a variety of input and output levels.
 
I'm not saying at all that RME is trying to be deceptive, just that a lot of times specs are "shorthand" because most consumers wouldn't know what to do with more technical info. Also, if 90% of interface manufacturers give best-case specs, you kind of have to go along. I know of one PA product (not Cerwin-Vega!) whose wattage specification is theoretically possible, but can be obtained only if you defeat all the thermal protection. Good luck with that after running it for 5 minutes 
 
Also, the channel separation within the interface may be tested without the breakout cable being connected, which could influence the results. Furthermore, the source signal is a click, so it's generating a lot of energy when it clicks. It would be interesting to know if this bleed tends toward high frequencies.
 
Bottom line is it's not SONAR, so I'm not ready to rule out crosstalk.

It seems obvious there's crosstalk getting in after Sonar, or at least I've never seen crosstalk in Sonar and, it being entirely digital, either it's not there or lots of people would have seen it. I just thought I'd see if it's an issue with RME generally, and as far as I can see it isn't, at least not noticeably so. Unless my interface is the odd one out. I thought of frequency issues which is why I thought white noise might be the way to go, but even as hot as I can get it nothing shows up.

Could be the breakout cable, as you say, or the preamp for the cable.

As for manufacturer specs, there should be an industry award for the one finding the most imaginative way to not quite misrepresent their products.

So correct me if I'm wrong, but the export from SONAR should be without crosstalk right?  Because the output device has nothing to do with the mixdown process.
 
Here's an export of one of the tracks in question:  http://kybernetos.com/03_Servants_CSm_v2.21.mp3
 
Can someone confirm/deny that the left channel is entirely devoid of click?  I can hear it faintly in the left channel played through a number of devices.

There is a click in there, however, it is buried in a noise floor and that fact alone is significant - it means that the click has become part of an audio stream somewhere.
 
Referring back to your original post, it's easy to show that panning a signal 100% to one side in SONAR does not produce any signal in the other side in the form of digital data. If you don't believe me, do this:
 
1. Insert mono audio into a track. Pan the track full left.
2. Bounce to a new track, with Main Outputs as the Source Category to make sure you're picking up the entire signal chain.
3. The new track will have audio only in the left channel.
4. Select the new track, then go Process > Apply Effect > Gain and turn down both left channels to -INF (0.0%). Click OK.
5. Now the bounced track appears to have no audio.
6. Select the bounced track and choose Process > Apply Effect > Normalize. If anything was in the right channel, this would bring its level up to maximum.
 
You will not find the cause of this issue by looking within SONAR's audio engine or mixer, because that is not where the problem lies. I can think of only two options:

• Something is misset inside SONAR, within a plug-in, or something that doesn't relate to panning per se.
• The click has "hitched a ride" on an audio signal, which I think is more likely. I don't know what you are recording, how you are recording it, how your patch bay is set up, if you use mics, whether you're picking up radiated leakage from headphones or speakers, or what could be the cause. It could perhaps be diagnosed by long distance but we would need a complete rundown of the audio routing and recording procedures you use.

The export I posted is a good example because it's a pretty straight-forward project as of now.
 
Track 1: an audio "guide" which is a Guitar Pro wav export, imported into SONAR.  Muted at mixdown time.
Track 2: a click track I created in Dimension Pro using the Piano Roll View and creating a groove clip
Tracks 3,4,5: Dimension pro string/brass/orchestral tracks.  I generally compose in Guitar Pro, export the midis from GP, import midis into SONAR for use with Dimension Pro
 
So no mic involvement, no actual "recording", just wav/midi manipulation.  I don't believe Dimension Pro or midi is the culprit.  As an even simpler example, the TestProj.zip I posted earlier is a single wav imported into SONAR and panned - I can export that and hear the same bleed.

scook
Are their any plug-ins in the signal path which may introduce crosstalk?

This can easily be the case. Some Mono/Stereo plug-ins may induce a little bit of cross-talk, which may be produced by the modeling techniques or just the stereo-to-mono conversion, if mis-matched. Disable all plug-ins with the FX button at the top and see if the cross-talk remains.