Parallel processing question.

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

27 Oct 2015

So one of the few mixing tricks I've wrapped my head around is the idea of parallel processing, i.e. making a parallel channel and compressing the shit out of a signal to mix it with the original signal.

I'm curious if anyone has any tips on how best to handle this in reason as far as what compressors you like to use for perhaps a bass vs. a lead or a drum track vs. a vocal track.

Last time I did it I just used pulveriser for all my parallel compression and I thought it sounded okay but my ears are about a million miles from golden when it comes to anything mixing related.

Thoughts? Suggestions? Any particular REs or settings you like to use?

User avatar
mcatalao
Competition Winner
Posts: 1824
Joined: 17 Jan 2015

27 Oct 2015

JNeffLind,

For hard compressing i'm mostly using the SSL Channel Compressor on the paralel channel, does not have any latency.

If you want to compress it with a non 0 latency device you have to take care of latency on the original channel, prefeareable with VMG-1.

I like the harsher settings of Dinamite too, with a little saturation from Saturation knob. But mind that you have to compensate latency and account for both devices.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

27 Oct 2015

It depends on why you are choosing parallel processing. If increasing loudness/intensity is your goal you will IMO need to use a compressor that has a very fast attack. If you do not, and use for example the SSL channel compressor, you will add transients/dynamic range to the signs rather than reduce it. FWIW, I don't find I need parallel processing to ADD transients, and find the SSL channel comp excellent on it's own for this job.

For adding loudness/intensity via parallel processing I find it key to use a fast compressor and often also saturation. The Softube FET and Saturation Knob are excellent for this job IMO. As mcatalao mentions you will need to compensate for latency with Normen Hansen's excellent VGM-01. Softube's Dyna-Mite is also a good choice, as is the MClass compressor IF followed by a saturation device such as Scream or Pulveriser.

For vocals I'm am of course biased, and use the Selig Leveler at 50%, which is the same as using a parallel channel (but without the need for latency compensation). The Leveler, when used as directed, gives you a true 50/50 blend at it's default 50% setting, unlike most other compressors with dry/wet controls. The reason I use the Leveler at 50% is that it gives far more gain control than other options. If I was using something like a LA-2a type compressor I don't find parallel processing to be necessary - and I'm 100% all about using as LITTLE processing as possible to achieve the desired goal.

All of this is personal choice, stating what I hope is the obvious. You may have to try a few options before deciding on a final choice for each job. But as I said, the first decision IMO is to ask the "why" question: what are you hoping to achieve or change about the sound in question? From there, make choices depending on your goals.

If you can mention more specific goals I'm sure you'll get more specific answers. :)
Selig Audio, LLC

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

27 Oct 2015

mcatalao wrote:JNeffLind,

For hard compressing i'm mostly using the SSL Channel Compressor on the paralel channel, does not have any latency.

If you want to compress it with a non 0 latency device you have to take care of latency on the original channel, prefeareable with VMG-1.

I like the harsher settings of Dinamite too, with a little saturation from Saturation knob. But mind that you have to compensate latency and account for both devices.
Thanks for the input man. Don't have the dinamite but it seems like the SSL may be the right choice. I'll have to try it out.

User avatar
gak
Posts: 2840
Joined: 05 Feb 2015

27 Oct 2015

I've found it utterly bizarre in reason. In fact, it's one of few real complaints I have. You should be able to set it up on one of the sends and control it that way, but that doesn't seem to work as well*. That is how I did it in every other host. Now the softube FET has it's own and that seems to work pretty well.

FWIW, having a combinator with a 14:2 seems to be the best route.

*The one exception being the MCDSP 670 which seems to be a beauty. The FRG/MOO, not so much. Strange to me.

User avatar
Marco Raaphorst
Posts: 2504
Joined: 22 Jan 2015
Location: The Hague, The Netherlands
Contact:

27 Oct 2015

Yes Pulveriser and Phase invert the signal. Love that on Parallel.

BTW using a Send effect can be nice too :) It's the same as Parallel Channel but without the channel strip stuff. But you can parallel process in any DAW that way.

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

27 Oct 2015

selig wrote:It depends on why you are choosing parallel processing. If increasing loudness/intensity is your goal you will IMO need to use a compressor that has a very fast attack. If you do not, and use for example the SSL channel compressor, you will add transients/dynamic range to the signs rather than reduce it. FWIW, I don't find I need parallel processing to ADD transients, and find the SSL channel comp excellent on it's own for this job.

For adding loudness/intensity via parallel processing I find it key to use a fast compressor and often also saturation. The Softube FET and Saturation Knob are excellent for this job IMO. As mcatalao mentions you will need to compensate for latency with Normen Hansen's excellent VGM-01. Softube's Dyna-Mite is also a good choice, as is the MClass compressor IF followed by a saturation device such as Scream or Pulveriser.

For vocals I'm am of course biased, and use the Selig Leveler at 50%, which is the same as using a parallel channel (but without the need for latency compensation). The Leveler, when used as directed, gives you a true 50/50 blend at it's default 50% setting, unlike most other compressors with dry/wet controls. The reason I use the Leveler at 50% is that it gives far more gain control than other options. If I was using something like a LA-2a type compressor I don't find parallel processing to be necessary - and I'm 100% all about using as LITTLE processing as possible to achieve the desired goal.

All of this is personal choice, stating what I hope is the obvious. You may have to try a few options before deciding on a final choice for each job. But as I said, the first decision IMO is to ask the "why" question: what are you hoping to achieve or change about the sound in question? From there, make choices depending on your goals.

If you can mention more specific goals I'm sure you'll get more specific answers. :)
Thanks for the input man. As for my goals, I just want to make my stuff sound as "professional" as possible, whatever that means. I'm trying to advance from simply applying cut and paste techniques willy-nilly but my ears don't seem to be improving the way I'd like. Honestly, on my last few tracks I just applied Pulverizer on the basic pulverisation setting (the default patch) as parallels on everything and thought it sounded better, perhaps only because everything got louder though it did also seem to make individual parts stand out more. That's my most basic goal at this point, to make all the parts discernible and clear instead of just kind of mushing together. I'm guessing EQing is more important than parallel compression for this goal but I have (literally) 92 songs finished/produced/arranged that need only to have vocals recorded and be mixed. I don't have time or energy to make them all perfect, I just want the mixing to be a cut above the dude who plays on FL studio in between bongers which is about where my innate mixing skills reside. Whether or not I "make it" will balance on my songwriting (my strength) and I'm just hoping to get my mixing to the level that it doesn't obscure the musicality I do have. I know I should take my time, but I can't stop writing. I'm compulsively hypergraphic. Wrote two songs today actually (though they're just first drafts and still need lots of hours before I worry about any mixing on those, haha).

I do have the leveler and love it and use it on all my vocals. I was wondering if there is any special magic behind this "New York compression" technique that relied on downward compression instead of upwards as the leveler does. If I were to use the leveler on stuff besides vocals (like basses, drums, etc.) at 100% on a parallel channel and then apply some saturation via the Saturation knob would you expect a positive result? I know it is situational and there's no cut and paste answer that will always be right, but I guess I'm asking if the same principles apply for upward or downward compression when it comes to parallel processing. I've got lots of drums that I'd particularly like to even the dynamics out on and was intending to try Leveler for this. Or do you think I'd be better off going with the FET?

Thanks again for your input, and to the peanut gallery who might be reading this. BUY THE LEVELER IMMEDIATELY AND READ THE VERY STRAIGHTFORWARD MANUAL. You'll be glad you did. Great RE.

User avatar
Raveshaper
Posts: 1089
Joined: 16 Jan 2015

28 Oct 2015

The biggest difficulty I have had when it comes to making my work sound "professional" is overcoming the placebo effect involved with the fact that I made a song as opposed to a professional. I made it, so it must not sound as good as that other person. That said, there is a lot more to parallel processing than just adding in a heavily compressed signal.

I work pretty much exclusively in the electronic arena of things and use samples or synths for everything. But even with recorded instruments and vocals, you can layer things in interesting ways to get a more "pro" result.

I would look into creating parallels and applying EQ to them to carve out frequencies you want to emphasize less, thereby strengthening "sweet spots" in common between original and parallel channels. Another trick is to throw some D-11 foldback in front of an RV7000 and enable EQ with the lows pulled down and gentle boost around 1k of 2dB or less. Using that as a send effect, adjusting the foldback to taste, can really fill the background with satisfying reverb.

Remember, less is more. If you can get the sound live, do it live. Best quality sounds first, minimal treatment after. That's the real difference between pro and everyone else as far as I can tell. It's why even if a sound is from a synthesizer, the "pro" guy has the real hardware in his studio. He's getting the sound direct from the source, not pitching a sample to make do.

That's all I got. I am probably incorrect about some of this stuff, but there ya go.
:reason: :ignition: :re: :refillpacker: Enhanced by DataBridge v5

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

29 Oct 2015

Raveshaper wrote:The biggest difficulty I have had when it comes to making my work sound "professional" is overcoming the placebo effect involved with the fact that I made a song as opposed to a professional. I made it, so it must not sound as good as that other person. That said, there is a lot more to parallel processing than just adding in a heavily compressed signal.

I work pretty much exclusively in the electronic arena of things and use samples or synths for everything. But even with recorded instruments and vocals, you can layer things in interesting ways to get a more "pro" result.

I would look into creating parallels and applying EQ to them to carve out frequencies you want to emphasize less, thereby strengthening "sweet spots" in common between original and parallel channels. Another trick is to throw some D-11 foldback in front of an RV7000 and enable EQ with the lows pulled down and gentle boost around 1k of 2dB or less. Using that as a send effect, adjusting the foldback to taste, can really fill the background with satisfying reverb.

Remember, less is more. If you can get the sound live, do it live. Best quality sounds first, minimal treatment after. That's the real difference between pro and everyone else as far as I can tell. It's why even if a sound is from a synthesizer, the "pro" guy has the real hardware in his studio. He's getting the sound direct from the source, not pitching a sample to make do.

That's all I got. I am probably incorrect about some of this stuff, but there ya go.
Thanks for the input man. That all makes sense. I'm still trying to wrap my head around EQ (I understand the basic concept, just can't seem to develop any intuition for it and have little confidence in the choices I make, often reversing them). I believe I've chosen good sounds so maybe I'm worrying more than I should. Like I alluded to earlier, my ears are molded from tin foil.

sleeper0013
Posts: 18
Joined: 20 Feb 2015

30 Oct 2015

parallel is the best way to get that get that professional sound.


Squashing the hell out of reverb with fast attack and long release, while shaping the it with EQ rolling off the base frequencys, can realy bring out presence clarity. adding reverb to source channel really limits your abity to shape the reverb how you need it with out changing the source as well.

using parallels to boost, is also great. you leave your source alone, draw up a parallel, filter out a scoop you want to boost, add some subtle effect that compliments the scoop. for base the pulveriser or flange and phase is good, distortions or chorus. i play with it a bit if, then dial it in for targeted dynamic that doesnt distorting general sound shape. in the end you get this shaped dynamic boost that not only pokes throught mix is sounds full and rich.

i generaly use a rule that the SSL is for coloring and not potent FX. except for reverb.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

30 Oct 2015

Raveshaper wrote:If you can get the sound live, do it live.


;)

User avatar
gak
Posts: 2840
Joined: 05 Feb 2015

31 Oct 2015

I'm not clicking on THAT vid. Though not a "dem" I have much anger towards him! :lol:

User avatar
RFX
Posts: 30
Joined: 06 Apr 2015
Location: Germany
Contact:

31 Oct 2015

Why didn't anyone mention Scream 4 yet? I really like using it with the tape setting, to give drums some more punch. It doesn't really sound good when there are hi hats playing, but I usually have those on a different bus from my kicks and snares anyways. Oh, flip the phase, so you compensate for the phase flip Scream 4 does (can someone explain why, btw?).

But in general, the idea of parallel processing isn't only about compression. Especially when designing synth sounds, I like to do stuff like parallel distortion (even though I agree that shouldn't really be done when mixing). But then again, I don't know what kind of music you guys make. My ideas of using parallel processing is probably a bit different to you guys' idea, since I probably produce entirely different music.

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

04 Nov 2015

sleeper0013 wrote:parallel is the best way to get that get that professional sound.


Squashing the hell out of reverb with fast attack and long release, while shaping the it with EQ rolling off the base frequencys, can realy bring out presence clarity. adding reverb to source channel really limits your abity to shape the reverb how you need it with out changing the source as well.

using parallels to boost, is also great. you leave your source alone, draw up a parallel, filter out a scoop you want to boost, add some subtle effect that compliments the scoop. for base the pulveriser or flange and phase is good, distortions or chorus. i play with it a bit if, then dial it in for targeted dynamic that doesnt distorting general sound shape. in the end you get this shaped dynamic boost that not only pokes throught mix is sounds full and rich.

i generaly use a rule that the SSL is for coloring and not potent FX. except for reverb.
Good tips man. Thanks for the input. I'll have to try that stuff out.

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

04 Nov 2015

RFX wrote:Why didn't anyone mention Scream 4 yet? I really like using it with the tape setting, to give drums some more punch. It doesn't really sound good when there are hi hats playing, but I usually have those on a different bus from my kicks and snares anyways. Oh, flip the phase, so you compensate for the phase flip Scream 4 does (can someone explain why, btw?).

But in general, the idea of parallel processing isn't only about compression. Especially when designing synth sounds, I like to do stuff like parallel distortion (even though I agree that shouldn't really be done when mixing). But then again, I don't know what kind of music you guys make. My ideas of using parallel processing is probably a bit different to you guys' idea, since I probably produce entirely different music.
Thanks man. Scream 4 has kind of been forgotten but I agree that the tape setting (as well as tube imo) are very useful.

User avatar
JNeffLind
Posts: 976
Joined: 16 Jan 2015
Location: So. Illinois, USA
Contact:

04 Nov 2015

gak wrote:I'm not clicking on THAT vid. Though not a "dem" I have much anger towards him! :lol:
I used to feel the same way. Now I just see him as a sort of barometer of far right nonsense. He's a clown playing to a specific audience.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

04 Nov 2015

JNeffLind wrote:I used to feel the same way. Now I just see him as a sort of barometer of far right nonsense. He's a clown playing to a specific audience.
I'm not even from the US and thats obvious to me. He did some disgraceful talks in his studio nonetheless - I guess the people still taking this guy seriously don't deserver better ^^

Stranger.
Posts: 329
Joined: 25 Sep 2015

04 Nov 2015

ΣΣΣ
Last edited by Stranger. on 20 Jun 2016, edited 1 time in total.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

04 Nov 2015

Stranger. wrote:Sorry Normen,but you NEED to make that unit FULLY automatic per sample,to really make it usefull'.
Sadly 99,999 is too high a number to be controlled via one parameter in Reason. But you can automate it as it it per sample no problem, you'll just need more than one automation lane.

Stranger.
Posts: 329
Joined: 25 Sep 2015

04 Nov 2015

ΣΣΣ
Last edited by Stranger. on 20 Jun 2016, edited 1 time in total.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

04 Nov 2015

Stranger. wrote:Nice edit ;)
There would be less problems if all latency is dealt with internally,not via another plug that won't shift automatically.
Also,there is NO way of actually seeing or getting a valid readout where and when to automate..
How would you recommend your own solution in practice Normen? :roll:


To the OP- try this link to get a slightly deeper insight of some history,how,why and when to (although this is also just 1 perspective) -->http://ryanpeoples.com/masteringparallelcompression
I don't quite get your issue. The VMG is meant to compensate internal processing latency of plugins - no plugin I know of has variable processing latency. Some compressors allow enabling or disabling "lookahead" latency and you could enable or disable a plugin, in both cases you'd know exactly when you do either of these things. So in practice I'd suggest putting the change in latency exactly where you enable that other plugin in the chain you compensate for.

Stranger.
Posts: 329
Joined: 25 Sep 2015

04 Nov 2015

ΣΣΣ
Last edited by Stranger. on 20 Jun 2016, edited 1 time in total.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

04 Nov 2015

Stranger. wrote:How would you recommend your own solution in practice Normen? :roll:
In practice I'd suggest putting the change in latency exactly where you enable that other plugin in the chain you compensate for.. :geek:

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

04 Nov 2015

Stranger. wrote:
selig wrote:you will need to compensate for latency with Normen Hansen's excellent VGM-01.
TBTH-the VMG is flawed,quite horrendously--- it will NOT compensate for any 'live or automated' adjustments in a mix.
Sorry Normen,but you NEED to make that unit FULLY automatic per sample,to really make it usefull'.

IMHO the best processing is 0= for every gain or highlight that's made,there has to be a loss or gain elsewhere of some kind,which in audio's case,that's normally going to be DB,frequency,or phase.
The loudness war was over years ago----there was many casualties,but humans are learning. :cool:
All people are doing by parrallel is masking,adding,subtracting,multiplying or dividing,as i understand.

The same thing occurs with imaging,example shown is an addition blend mode(basically merging.) :ugeek:
Image
https://drive.google.com/open?id=0B5T_Y ... XA2d2VHVGM
I honestly have no idea what you're talking about here, as I've found the VGM-01 to be incredibly useful - there is literally no other way to accomplish what it can do in Reason! Either you're exaggerating for effect, or you're doing things I've never heard of doing before (wouldn't be the first time for either!).

Maybe you could back off the hyperbole for a bit and explain what it is you're trying to do when you run into this issue?
:)
Selig Audio, LLC

User avatar
gak
Posts: 2840
Joined: 05 Feb 2015

04 Nov 2015

I haven't tried it. I guess now I know why I've had some issues with some things. Was always wondering why I was getting funny results with stuff like this in some cases.

"dear god, bitwig tried and failed to make the perfect host, please bless the props with the perfection to fix every single thing in the universe and make it always work....close is not good enough! " :x

Post Reply
  • Information
  • Who is online

    Users browsing this forum: Ahrefs [Bot] and 14 guests