Hi,

In general, is there any latency introduced by parallel compressing using a bus?  Or does Sonar know how to compensate for this kind of thing?
 
Sonic

Any plug-in requiring PDC will affect the project regardless of its location.

I better clarify further....  I wasn't referring to real time latency per-se, but rather wondering if Sonar is smart enough to compensate so it doesn't get things out of alignment internally.   Here's a specific example:  Let's say we route a track to Aux1 bus which has a plug-in with some latency, say a compressor.  So parallel compression is happening (the main dry track and Aux 1 playing in parallel).  Is Sonar smart enough to compensate for any latency caused by the routing through Aux 1 so that the resulting master bus output doesn't have latency/phasing issues?
 
Sonic

SonicExplorer
I better clarify further....  I wasn't referring to real time latency per-se, but rather wondering if Sonar is smart enough to compensate so it doesn't get things out of alignment internally.   Here's a specific example:  Let's say we route a track to Aux1 bus which has a plug-in with some latency, say a compressor.  So parallel compression is happening (the main dry track and Aux 1 playing in parallel).  Is Sonar smart enough to compensate for any latency caused by the routing through Aux 1 so that the resulting master bus output doesn't have latency/phasing issues?
 
Sonic

Yes, any plug-in requiring PDC will affect the project regardless of its location.

How then does someone determine what latency may be involved, and how does one compensate for it so two parallel tracks (one dry one wet) don't end up out of sync at the master output bus ?
 
Sonic

You're overthinking it. It just gets taken care of automatically.

Good question though
 
Following onto to ops question - never been able work out under what circumstances would you use the PDC toggle??

PDC depends on the plugin itself to report back what its internal latency is, so that Cakewalk can slow everything else down to match it. 99.9% of the time, the scheme works perfectly. It's one of the details that CW worked out long ago, before most other DAWs did.
 
Once in a while, though, a plugin isn't accurate in its reported latency and phase cancellations can ensue. Sometimes external hardware can also be an issue because there is no mechanism for it to report its latency back to the DAW.
 
It's easy enough to test a plugin by sending a sharp transient to it and zooming in on the waveform to make sure the main and parallel paths line up.
 
The only time you're going to routinely run into latency problems is with plugins that feature minimum-phase filters, meaning any equalizer or any effect that incorporates an equalizer. For that reason, you want to be careful applying EQ to parallel paths and avoid steep EQ boosts. Or use linear phase equalizers. These issues do not normally apply to compressors, even if the compressor incorporates a sidechain filter.
 
As for when you'd want to intentionally defeat PDC, I can think of only one scenario: when you have to track something new on a nearly-finished project, such as a last-minute BGV. In that case, it's acceptable to temporarily kill PDC for the sake of synchronization while tracking.

"

bitflipper 
The only time you're going to routinely run into latency problems is with plugins that feature minimum-phase filters, meaning any equalizer or any effect that incorporates an equalizer. 
 

 
bitflipper, do you mean it is only a potential concern for some types of EQ plugs which do NOT report PDC?  Or does your statement apply regardless of whether or not PDC is reported? 
 
I'm using parallel compression and some EQ plugs to bring out drum transients. So this paradigm we are discussing could apply in my case (that's one reason I created the thread....I was hearing tiny random differences between the hits of the snare and hi-hats, but turns out it must be the "humanize" feature of the drum VSTi).  Anyway, it got me thinking about potential latency issues.
 
Thanks in advance to everybody for helping me understand this potential latency anomaly. 
 
Sonic

There is “potential latency” but it’s very rare indeed for it to be an issue with parallel processing.

If it were all kinds of things would demonstrate problems - as well as linear phase eqs being notorious for high latency while they do their sums there’s anything that uses a “look ahead” algorithm, which many compressors and limiters do.

Some plugins take enough time to do their thing to need the audio buffer increasing to the point of noticeable latency if tracking and monitoring through the DAW using the track echo function. There are even a few around which require over a second on a pretty hefty PC/Mac, but even that gets compensated automatically so everything lines up.

9,999 times out of 10,000 all the modern DAWs can compensate for this without issues unless the plugin is very badly written. Which is very rare.

The DAW knows, from the answers to the sums it does, where on the timeline audio data should be placed, so all it needs to do is work out which plugins requires however much time to do their thing then shift the processed data around the time-line accordingly.

There’s really no need to worry over this, at least not unless you plan on coding your own DAW.