FX Before or After Parallel Compression

Discussion in 'Tracking / Mixing / Editing' started by mrfye17, Aug 24, 2016.

  1. mrfye17

    mrfye17 Active Member

    Hi Peeps,

    I have a question and not sure what the approach of the pro-mixers are.

    If I have a track (any type - lets say Kick as an example) which I process with Eq, Gate, Basic Compression etc on the inserts. and I also send this to a Parallel Compression Aux track and now want to add FX (Reverb etc.)

    Do the pros route to the FX Aux tracks before or after the Parallel Compression?

    - mrfye17
    Brien Holcombe likes this.
  2. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    Generally my approach would be any sort of 'special effects' type things like reverb, delays, pitch shifting, stuff like that would be post the parelell compression. I veiw the parelell track as sort of a branch of the main track, and then any auxiliary type processing as an additional seperate stage.

    This should help keep phase anomlies to a minimum and keep the clearest signal path.
  3. Sean G

    Sean G Well-Known Member

    It really depends on the material...and if it needs EQ or the like prior to parallel processing IMO.

    When it comes to the parallel processed track, I like to EQ that with much top & bottom end then compress...then blend back in with the original track.

    But to acknowledge your original question, that really depends on the original track to start with...no two tracks or circumstances will be the same from one mix to the next.

    IMHO of course. ;)
  4. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    Yes it depends on the material like the guys said.
    But lets talk about the mechanics:
    • Parallel compression =
    • with 1 track + 1 buss, the signal can be sent pre or post fader (Question to all : does pre mean it is prior the effect bin or just the fader and pan ???)
    • Assuming post is post everything, you won't do parallel compression anymore, it will be compression of buss on top of the compression of the chanel.
    • To make a special effect, you could send the channel to a buss effect and send the effect buss to the parallele compression buss
    • You can apply effects on the channel and on the buss
    • You can send only the channel to the effects to get a more natural result or only to the buss
    • In the end instead of a buss, you can copy the track and have completly different settings and buss sends
  5. bouldersound

    bouldersound Real guitars are for old people. Well-Known Member

    Putting conventional eq on one path of a parallel process will produce phase interactions of some sort, though they may not necessarily be audible. That's a good place to use FIR filters instead.
  6. Boswell

    Boswell Moderator Distinguished Member

    ...as long as your DAW can compensate for delay on a per-channel basis rather than only at the track level!
  7. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    Never had problems of that sort with Sonar, but it's good to know it can happen !

    I often make a copy of a track and EQ and comp differently without any issue...
  8. Boswell

    Boswell Moderator Distinguished Member

    Sure. If you do it by making a copy of a track first, then the DAW can automatically compensate at the track level. It's when you source two processes from the same track that can you get problems caused by the difference in time delay between the two processes. It's not dissimilar from using an external hardware EQ or effects unit inserted into a track - some DAWs simply do not have the facilities to allow for the delays through them and correctly time-align those tracks in the mix.
    pcrecord likes this.
  9. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    Oh !! Brain freeze, I should have said when sending signal to Buss or multiple busses.. ;)
  10. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    I belive pre fader is pre fader/pan but post effects inserts. On an anlog console you'd likely have direct outs which would be a tap off the gain/trim knob, and before the insert jacks.

    Learned this the hard way in Dp7 :( we had to use a work around where we put a 'dummy effect' on the the channel like a trim pluggin, with no settings engaged, in order for the delay compensation to kick in. Sending just the dry signal with no effect via the aux/buss cashed phase anomalies.

    I send things like drums to their busses via the actual channel output, to the buss/subgroup. This way if things like the hi hat or snare bottom don't have effects I'm covered without wasting cpu or adding unecessary processing. They all hit the buss at the same time. And the bus will have some sort of effect of on it. If I'm doing things like kick and snare busses I feed them into the main drum buss for this reason. Since all the busses have an insert it avoids those issues. And if for some reason I need a dummy insert, it's only one per bus vs one per track/channel.

    I'm not sure specifically how this works in every daw, but it seems like a fairly safe practice.

    This is also why I've become a big fan of compressors w a mix knob like the waves h-comp. it may now be identical to parelell processing but does the gyst of it good enough for me, and eliminates tons of routing and under the hood things.
  11. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    This would mean to have real parallel processing we need to copy the track, because with busses, we are actually chainning the effects and not doing parallel. unless we send the signal to 2 seperate buss and comp there...
    I've been doing this for ages by instinct and now I know why... Interesting !! :)

    The compressor I use from fabfilter has that feature but I've yet to use it in that manner. Next mix maybe !
    kmetal likes this.
  12. bouldersound

    bouldersound Real guitars are for old people. Well-Known Member

    I'm not sure I get what you're saying. When one channel gets to the main bus by two paths it's parallel. If one or both paths have separate processing it's parallel processing. You don't necessarily need to create two buses to make it parallel, one bus plus a direct path to the main mix is sufficient for it to be parallel. I don't know about other software but in Sony Vegas any processing delays are compensated for: you can totally route a channel through a bus and direct to the main mix at the same time, with processing on the bus, and it the two paths will not be delayed relative to each other or any other tracks.
    kmetal likes this.
  13. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    I think Boulder said it about right. But it depends on the daw like bos said. The old PT LE had no delay compensation to add to its already phasey sound propriety hardware.

    Interestingly I was looking at the waves page and certain plugins have there own inherent latency and others have none. Req has no latency where others had 256 samples of latency.

    In dp7 if you bussed a dry signal to a buss you had to add a pluggin to the dry to compensate for latency.

    If your parelell processing it's still parelell if there's insert effects because your sending the same signal to two diff places at once like Boulder said.

    In the case of compression even if there's a compressor on the source track your still parelell compressing if your bussing it to another compressor.

    If I'm doing this sort of thing I'm gonna take care of things like HPF for instance at the source track on the insert before it gets bussed to a compressor or whatever.

    Duplicating a track can avoid some need for delay compensation but also cause the need to add duplicate plgigins like HPF or whatever to maintain good phase. This can due to pluggin latency (brand new consideration to me), or just general frequent masking. It could also add cpu load.

    There's no perfect solution. In the case of auto delay compensation all your doing is adding a delay so the tracks line up w the slowest / highest latency track. This keeps things in sync at the expense of realtime performance.
  14. Sean G

    Sean G Well-Known Member

    Thats generally how I do it...I believe Studio One compensates for any delay with processing
    pcrecord likes this.
  15. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    I might be wrong but to me parallel compression is having the same raw signal entering 2 different compresor which maybe adjusted differently. Commonly one heavily compressed and the other lightly.
    But if I put a compresor on a track and the result of that compression goes to a buss for more compression, it's not parallel. It's a chain like if I would put 2 compressor on the same effect bin..
    Am I wrong ??
  16. mrfye17

    mrfye17 Active Member

    For Kick and Snare I usually Compress lightly on the original track (max 3db GR) then use an Aux for a heavy Compression (15 - 20 GR, fast release / fast attack) for thickness and is blended into the original.
    Both tracks then go to a Sub Grp and if I need more point (transient attack) I may either use
    a Comp (med attack / slow release) or a Transient designer.

    ....and that's when I get stuck as to where the FX Sends should come from lol!
  17. bouldersound

    bouldersound Real guitars are for old people. Well-Known Member

    Not really wrong, just more specific. The channel compression and the bus compression may be in series, but that's beside the point. It's parallel compression if the signal gets to the main mix by two paths and at least one path is compressed, which includes your scenario of two buses with different compression. One or both paths may be compressed, and compression may be applied to a submixed group of channels or only used on individual channels.
    pcrecord likes this.
  18. bouldersound

    bouldersound Real guitars are for old people. Well-Known Member

    Normally I try to get by with one shared reverb but that's not always practical. In this case you might try just taking it from the channel sends. If that doesn't give the density you want maybe copy the reverb to the parallel compression bus and use the built in wet/dry blend controls to add some there.
  19. kmetal

    kmetal Kyle P. Gushue Well-Known Member

    I'm not exactly sure. I tend to think what Boulder said, but I'm not sure what's technically correct. Interesting.

    I'd tend to feed a little of both into an aux for things like reverb or delay. Probably a little more heavy on the send for the less compressed track. Sending a super compressed signal could muddy things up as far as reverb goes, so I kinda apply the same general thinking where the heavily processed / parelell track is there for reinforcement not as the primary sound.
  20. pcrecord

    pcrecord Don't you want the best recording like I do ? Well-Known Member

    I'm sure you guys are right about the term parallel.
    The most important thing is to be aware of what we are doing. If we want to compress the raw track twice or use 2 layers of compression. (copy track or buss)

    Those are the best discussions, when we get to re-evaluate the way we do things ;)
    kmetal likes this.

Share This Page