PDA

View Full Version : DX/VST latency not compensated



Cary B. Cornett
07-31-2008, 09:20 AM
After not needing it for quite a while, I found I needed to use parallel compression in a mix. I assigned a track to an additional buss, and tried to put an eq and compressor pre-fade in the buss. I have used these particular plugs this way before in SSL with no problem, but now, when I put them in, I get the dreaded "flange" effect that indicates uncompensated latency. This happened with both DX and VST versions of the same plugins. :(

I also tried this trick using the SS Graphic EQ and Levelizer, and it works perfectly... no flange. :cool:

Is it possible that latency compensation got "broken" in one of the more recent updates??

Bob L
07-31-2008, 10:42 AM
Not that I am aware of... but... if you are using parallel bussing with latency plugs in one or both of the separate mix paths, it is possible that the input side of the mix is being forced to loop around at a different rate than the parallel bus which is then looping back around over again to deliver the final mix signal... you could be causing phase issues...

The loop to assemble the mix gets very complex when latency plugins are used and with the idea of processing the same signal thru varying busses using latency compensation within one or both of those busses... anything could happen.

Latency plugs are very constrictive to the freedom of the mix, in my opinion, and I have generally opted to stay away from them. You might have success with your particular mix, by first rendering one of the signal paths thru any latency plugs... then remove the plug and finish the mix with only one active latency plug in the parallel path.

Bob L

Cary B. Cornett
07-31-2008, 12:15 PM
Not that I am aware of... but... if you are using parallel bussing with latency plugs in one or both of the separate mix paths, it is possible that the input side of the mix is being forced to loop around at a different rate than the parallel bus which is then looping back around over again to deliver the final mix signal... you could be causing phase issues... The only place where I put latency-causing plugs was in the parallel path. The "direct" path had no such plugins.

In this situation I had 2 piano tracks (a stereo pair), which were fed both to the Master buss and another buss labeled "Piano sqz". That second buss was, of course, routed back to the Master. I put the processing plugins, pre-fade, in that second buss. I tried an EQ feeding a compressor (these are all the Sonitus-fx R3 plugins), and got the flanging problem. I then removed these and replaced them with the Sonitus Multiband Compressor... same problem.

In time past I have used the Multiband compressor in a similar fashion in mixes and did not have this problem.

Bob L
07-31-2008, 12:21 PM
Well... you can tell if any plugin is changing buffer sizes (and therefore possibly adding latency) by looking in the top left corner of the main titlebar during playback for the Buffer Size Change icon to popup.

You can click in that icon to see the name and track of the first plugin causing the change.

I also just thought of something... if you just moved up to Full... make sure you installed in its own fresh folder and not overtop the old Lite folder... the preference files will not be compatible and perhaps you are running with corrupt options and maybe a disabled latency compensation.

Bob L

Cary B. Cornett
07-31-2008, 12:37 PM
Well... you can tell if any plugin is changing buffer sizes (and therefore possibly adding latency) by looking in the top left corner of the main titlebar during playback for the Buffer Size Change icon to popup.

You can click in that icon to see the name and track of the first plugin causing the change.

I also just thought of something... if you just moved up to Full... make sure you installed in its own fresh folder and not overtop the old Lite folder... the preference files will not be compatible and perhaps you are running with corrupt options and maybe a disabled latency compensation.

Bob L I'll check out that buffer size change thing.

This install of Full was in fact on a fresh XP install that had never had Lite installed on it. I did copy over the Fkey files, but not, AFAIK, the Preference file, especially since I know some settings I liked in Lite have not yet been made in SS Full.

Dave Labrecque
07-31-2008, 04:10 PM
I'll check out that buffer size change thing.

This install of Full was in fact on a fresh XP install that had never had Lite installed on it. I did copy over the Fkey files, but not, AFAIK, the Preference file, especially since I know some settings I liked in Lite have not yet been made in SS Full.

Cary,

Work-around idea: patch the same plug-in causing the latency on the direct channel as well, but keep the ratio or threshold "out of the way", so that it doesn't actually impact the signal, other than giving it the same latency. Might not work, since, as Bob points out, the different bus loops may be doing different things, but might be worth a try.

Alternatively, you could try to assess the latency difference (if it's constant) and use JMS' latency compensator plug-in thingy to match 'em up.

Cary B. Cornett
08-01-2008, 05:29 AM
Cary,

Work-around idea: patch the same plug-in causing the latency on the direct channel as well, but keep the ratio or threshold "out of the way", so that it doesn't actually impact the signal, other than giving it the same latency. Actually, that is exactly what I did on a couple of occasions before Bob implemented auto latency compensation.

An odd thing, BTW: On a different mix, I was using the PLParEQ3, which claims a latency of 8192 samples, and it did not pull the track I used it on out of sync. I was not, however, running it on a "parallel path", just directly inline.

Dave Labrecque
08-01-2008, 09:50 AM
An odd thing, BTW: On a different mix, I was using the PLParEQ3, which claims a latency of 8192 samples, and it did not pull the track I used it on out of sync. I was not, however, running it on a "parallel path", just directly inline.

That would be the latency compensation working correctly, then, no?

Cary B. Cornett
08-01-2008, 01:39 PM
That would be the latency compensation working correctly, then, no? Yes, at least in that context. Still, for parallel processing things are not quite right... except for Saw Native plugins. Too bad I can't afford more of those right now. :(