Compression suggestions please

I have a heavily strummed guitar which I like but I want to even the volume without attacking the peaks too much and I want to preserve the natural sound.

With compression I have tried a fast attack and medium release, threshold just barely touching it but it's not doing the job. I've also tried a slow attack and faster release but it sounds too processed.
Any tips please?
Thanks

Acnew by Mike Spot | Free Listening on SoundCloud
 
Long RMS time. Lookahead/PreComp half as long as RMS. Ratio probably doesn't want to be as much as 3:1. Set the threshold to just start catching the loudest part. Then set the knee so it's almost always doing something. Then adjust the ratio again until it actually works. This is a time I'll turn on automatic gain compensation. If your compressor doesn't have all of those knobs, it's not the right one for this job. ReaComp has them. It's maximum PreComp is 250, so I set RMS at 500 and it works for almost everything.

Add whatever other compression you want to adjust the envelope after this. It'll be nice and leveled out, and now you can deal with the peaks more consistently.
 
The first track is pretty compressed already, so I guess I'd like to hear the original to know where it's starting from. There's a couple spots where it seems like the mic placement moved noticeably in that track and I'd work on cleaning that up a bit because the difference in EQ will make that harder to place, regardless of the compression. If you do stick with it, I'd say it could still use both EQ and volume automation - not to address the transients, which seem to be bothering you most, but the variances in tone and general level (the latter is not something I'd generally try to do with just a compressor).

And, I think using your ears is the key, and it's nearly always in the context of the mix that you want to try and manage a track like this for all of those things. The nits I notice about the mic placement/EQ shifts might disappear if this track is way back and off to the side somewhere in those parts, so it's going to be a waste of time tackling them now. Even the transients might end up being something you want more of in some places - you just never know.

For yucks, I (re-?)compressed your first take to just knock about 1.5dB off the peaks without compressing more or changing the loudness. It's using one of the Logic built-in comps based on the SSL comp (which lack the knobs [MENTION=142548]ashcat_it[/MENTION] says you must have!).

View attachment acnewC3kr.mp3

P.S. FWIW, there's a comment in another thread about vocal tracking & comps IIRC and it mentions using 2 compressors - that's something that might work for something like this, too, where you just hit the transients, and only those, pretty hard, then use the more gentle stuff of that output.
 
Hi Keith,
The original track is not compressed at all.
You are correct about the eq on the first bit - I decided it needed to be more dynamic at the beginning so replayed it and I probably moved off the mic slightly.
I can hardly hear the compression you did but it sounds good. I may try one of the levelling comps.
Cheers.
 
How do I get around the issue of the precomp pushing the track out of time? Track compensation?
IDK, I never have that problem because PDC takes care of it. If you've disabled that for some reason...??? The track isn't record armed or part of a feedback loop? Those things defeat PDC.
 
(which lack the knobs [MENTION=142548]ashcat_it[/MENTION] says you must have!).
I said you need those knobs if you want to do the long slow lookahead leveling trick. It's basically just a way to "cheat" at volume automation.

P.S. FWIW, there's a comment in another thread about vocal tracking & comps IIRC and it mentions using 2 compressors - that's something that might work for something like this, too, where you just hit the transients, and only those, pretty hard, then use the more gentle stuff of that output.
There are plenty of comments on all kinds of threads about that. I've commented on several of those that it doesn't make sense to me to do it in this order. If the point is consistency between hits, then changing envelope of different hits based on their relative volume before leveling is going the wrong direction. If you read my comment, I suggested level first, then doing whatever envelope compression after so that second compressor does closer to the same thing on every hit.
 
As a starting point for learning compression I would hit the thing pretty hard and play with it. There are no magic settings for compression. It depends on what you have to start with and what you want it to sound like. You have to set it by ear.

The amount of work a compressor is doing is measured in gain reduction. Most of them have metering to show you that. Start with a fast attack and release. A lower threshold setting will make the compressor kick in at lower signal levels. This will make the compressor act on more of the entire signal, not just transient peaks. It will also increase gain reduction somewhat. A higher ratio setting will increase gain reduction. I'd start with around 10 to 15 dB gain reduction to make it fairly obvious to hear.

The attack time will control how fast the compressor kicks in. A very fast attack with heavy gain reduction should kill the transient. This should make it sound fairly mushy and dead. If you increase the attack time slowly you should find a point where it's slow enough to let the initial transient through before it clamps the signal. This will bring back some of the natural sound.

The release time is critical for shaping what the signal is going to do. If it's too fast you'll get pumping sounds. It can make the sound sort of backwards and unnatural. If you increase the release time slowly it should start to make things sound more natural. It's going to be dependant on the tempo and feel of the track, as well as all the other settings. If the release time is too long it won't allow the compressor to "let go" before the next transient, so it will kill the attack again. Specific numbers are useless. You have to set these by ear.

If you're still smashing the signal too hard for it to sound natural you can lower the ratio to the point where you're only getting a few dB's of gain reduction. This will make it a lot less obvious that there's a compressor. The "output" control, or Trim or Makeup Gain or whatever it's called on the unit you're working with is used to match the level of the output of the compressor to the input signal. If you flip the compressor on and off, these levels should be as close to each other as possible, by ear. Don't worry about the meters. If the levels are the same when you flip the thing in and out of bypass, you can decide whether or not the signal is better with or without the compressor. If you don't match these levels, our ears tend to play tricks on us. We will prefer the louder signal whether it's better or not.

As Keith suggested, it's not uncommon to chain separate instances of a compressor plugin. Using a fast attack and release, higher threshold and higher ratio will set the compressor up to act more as a limiter. A dB or two of gain reduction like this will tame the transients somewhat. If this is followed by another instance running a lower threshold, lower ratio and slower attack and release doing a couple more dB's gain reduction, this can help to increase the sustain and thickness of the track while still allowing it to sound natural. You might also want to use an EQ after (and/or before) depending on how the compression might change the balances that way.

If you have individual phrases that jump out, something like "Baa da da da, Baa da da da" and you want it to sound more like "baa da da da, baa da da da", it might be better to use a volume automation envelope to reduce the B before you reach for a compressor.

Also note that different compressor plugins behave differently and might have some, none, all or a few of the controls typical to a compressor. It doesn't matter. Depending on what platform you're on, there might be a bunch of free stuff you can play with to learn what it does. You might find something that can work well with a given track depending on what you want to do. Might be worthwhile, if nothing more than a learning experience.
 
I have a heavily strummed guitar which I like but I want to even the volume without attacking the peaks too much and I want to preserve the natural sound.

With compression I have tried a fast attack and medium release, threshold just barely touching it but it's not doing the job. I've also tried a slow attack and faster release but it sounds too processed.
Any tips please?
...
So, I re-read your initial post, and have to say it's a little confusing. When you say "volume" it can mean different things, though I think most of us think of that as the average (RMS) level, even that can be measured in different ways. And, the average is always affected by the peaks, so changing the average without doing much to the peaks is going to be a challenge.
Hi Keith,
The original track is not compressed at all.
You are correct about the eq on the first bit - I decided it needed to be more dynamic at the beginning so replayed it and I probably moved off the mic slightly.
I can hardly hear the compression you did but it sounds good. I may try one of the levelling comps.
Ok, it's not compressed, but it doesn't have a lot of dynamic range overall, TBH. What it does have (first attachment) is a fair amount of variance in where the dynamics are, and you can see there's some at the beginning and a little more toward the end. So, now the question is whether this is what you wanted, or whether it's a problem with the performance. If it's the latter, then you can do as others have suggested, and that's to use some automation, or what I'd probably try is "clip gain" on the separate parts and compression applied only on the ends, perhaps to even out dynamics. However, that conflicts with your wanting to preserve peaks.

So, really, this is one of those places where compression might be an answer for some part of the track, but if you're trying to even out a performance that is not done right or recorded right from the start, it's not the best place to spend time. I'd just re-record it. *Then* you might want to use EQ and compression to make it sit in the mix. Starting with a track (like this, i.e., a rhythm, acoustic guitar) that needs something to make it sound the way you want it to, outside of the mix, is something I do when I'm mixing a live recording and there's no other recourse.

Original MP3
Screen Shot 2018-12-30 at 4.24.47 PM.png

SSL (plugin modeled) compressed MP3
Screen Shot 2018-12-31 at 8.28.44 AM.png
 
Oh yeah :)
I have a previous acoustic guitar take which is very non dynamic, more of a pad.
I'm going to do another more dynamic take on top of the other and see where it leads me.
 
Sometimes more dynamics could be a good thing.

Compression gives you less dynamics. It squashes the dynamics. That's why if the sound and dynamics are okay but the volume fluctuates, that's when I'd do volume automation. Volume automation doesn't do anything about dynamics, just volume. There's a really cool Reaper video by Kenny Gioia that shows a few tricks that way.

Compression is also a big part of the sound of most stuff that isn't Jazz or Classical music. Whether we realize it or not, there's been gobs of it on everything for decades. An acoustic guitar and a vocal on its own might not even need or benefit from compression before the mastering stage. Acoustic guitar and vocal in a bigger mix with distorted guitars and drums might need substantial compression to get it to even sit in the mix.

Plus there's just certain compressors that work easier for specific stuff. It's not uncommon to have a few different ones to go to for whatever character they might have.
 
Sometimes more dynamics could be a good thing.

Compression gives you less dynamics. It squashes the dynamics. That's why if the sound and dynamics are okay but the volume fluctuates, that's when I'd do volume automation. Volume automation doesn't do anything about dynamics, just volume. There's a really cool Reaper video by Kenny Gioia that shows a few tricks that way.

Compression is also a big part of the sound of most stuff that isn't Jazz or Classical music. Whether we realize it or not, there's been gobs of it on everything for decades. An acoustic guitar and a vocal on its own might not even need or benefit from compression before the mastering stage. Acoustic guitar and vocal in a bigger mix with distorted guitars and drums might need substantial compression to get it to even sit in the mix.

Plus there's just certain compressors that work easier for specific stuff. It's not uncommon to have a few different ones to go to for whatever character they might have.

Yes. Old tricks are sometimes the only tricks.
 
Of course "dynamics" means almost exactly "changes in volume". It can actually refer to any number of things depending on the time scale that you're looking at. You can have kind of "meta" dynamics where different whole sections of a piece are at different levels - like the soft verse loud chorus thing. Then you have "macro" dynamics within each of those sections where each note or hit might be different - in this case some uneven strumming. Then there's the "micro" dynamics - what I usually call the "envelope" - the way that individual notes or hits develop - Attack, decay, sustain, release... Certain strategies are more appropriate for dealing with each of these.

When you're trying to adjust two or three of these things at the same time, it's important to think about what order you want to do it in. It seems most people try to do it micro>macro>meta, but for me it usually works better to do it the opposite direction. The compressor you strap on to control the micro dynamics will always be affected by the macro dynamics. If one hit is quieter than the next, it won't get squashed as hard, and it's micro dynamics won't be changed as much. If you bring them both closer in level first (macro), then they should come out the compression closer to the same.


Edit to put it a slightly different way - Leveling before compression is essentially like having a more consistent performance to begin with. Nobody would ever argue that a more consistent performance wouldn't just help and make any peak compression more consistent and natural. Volume leveling may not be ideal, but it can have same effect, but only if it comes before the compressor. (And of course my first reply uses a compressor to affect volume leveling because we can do that with digital tools in ways that analog comps can't)
 
It's also helpful and not uncommon in a complicated mix to be able to ride the volume of a given track both pre and post FX chain. Before, as Ash mentioned, to give compression something more consistant to squish. Once the sound is what you want (post effects) a second volume envelope to balance the level of the track in the mix to what it needs to be at any given moment.

Must have been interesting in the days before console automation when the mix engineer would solicit the help of another 4 or so pairs of hands to do rides in real time while printing the mix.
 
ashcat, I used reacomp and it does sound better.

Acnewcomped by Mike Spot | Free Listening on SoundCloud

How do I get around the issue of the precomp pushing the track out of time? Track compensation?

alterman, yes I could use volume automation but it's time consuming and I really should get me head around compressors, cheers.
Cheers

Try a bunch of different compressors. They're not all the same. Some are distinctly more transparent than others when placed on the exact same settings.

I didn't read much of the previous threads, but you can time the attack and release to the bpm of the song. Convert the ms to bpm and sometimes this helps keep the groove intact.
 
Maybe instead of compressing the track you can use parallel compression keeping the dynamics of the original track and adding energy with the parallel one
 
Back
Top