Is it a good idea to use two compressors on the master?

How to do this, that and the other. Share, learn, teach. How did X do that? How can I sound like Y?
RELATED
PRODUCTS

Post

I work with ableton 9 and I've been using a custom rack to master my songs after mixing them. I make sure to leave enough headroom so I never go past -6.00 DBs when mixing, then I use my custom rack and boost the sound.
With one compressor having the threshold at -20 DBs, and the other at -14, I use the gain knob on the first compressor to boost the sound which will then be compressed by the second compressor, but I also use the gain knob on this one to make sure I reach -0.50 DBs or 0 even.

So, my question is:
Does this make sense? Is it a good idea?

Post

Sounds like you are pulling some pretty heavy compression with those thresholds. If it sounds good, go for it, I say :)

(link us some before/after sound files)

Post

Here is a song I worked on. I would like to point out that the only reason I rendered the full song is because it contains many elements that feel more alive; like the snare at the beginning, for example.

Rendered without using the rack

Mastered using the rack

As you may have noticed, the song has a busy low end. That is intended.
Tell me, what do you think of it?

Post

Dude it sounds gorgeous. Really punchy and sparkly - well balanced.

More dynamic range in the pre-comp version obviously but I like what the heavy compression is bringing to the track. What comps are you using may I ask?

Post

MogwaiBoy wrote:Dude it sounds gorgeous. Really punchy and sparkly - well balanced.

More dynamic range in the pre-comp version obviously but I like what the heavy compression is bringing to the track. What comps are you using may I ask?
I'm using ableton's own Glue Compressor, which is actually a watered down version of Cytomics' Glue Compressor, which costs around 90 currencies :hihi:

I was thinking of sharing the Effects rack with people, but I see that my two samples have eleven plays each, which means people did saw it, and are not interested.

P.S:
I forgot to mention I also use Thrillseeker's XTC Blue before the compression.

Post

If it works and if you / tune needs it, go for it.
Soft Knees - Live 12, Diva, Omnisphere, Slate Digital VSX, TDR, Kush Audio, U-He, PA, Valhalla, Fuse, Pulsar, NI, OekSound etc. on Win11Pro R7950X & RME AiO Pro
https://www.youtube.com/@softknees/videos Music & Demoscene

Post

I'm responding to the PM request.

Took some time to listen to the A/B version, which are sadly not great to A/B if you used Thrillseeker XTC to boost the low frequencies. But I get where your concept is coming from. You want an answer to the question "are two compressors in series usable for mastering".

The short answer is: "it depends on the production", some might even say "if it works for you, go for it".


The long answer is:
If we only talk loudness raising, then you might get better away with a compressor -> limiter array. The first compressor would "compact" the signal, then boost the output of the compressor which is then driven into the limiter. The limiter then takes care of the maximum signals peaks only. Depends on how much you want to push it.

What you're doing is "double compression", and honestly... without a screenshot or settings, it's hard to tell what you're doing. If one compressor is handling peaks only, the other the average signal strength, or both doing the same, etc.

What I can easily tell from the demo however, is that I find the compression a bit too much (too low threshold for the first compressor - especially if you don't utilize a reference signal level during mixing like the infamous -18dBFS). It pumps in certain section (beginning and fairly quiet synth part). Unless that was a desired effect. I would not go that route as this changes the listening "experience" so to speak. But then again, I'm special on that behalf - I'm not "rooting" for the Loudness War/Race.


So what is the correct way? There is no absolute correct or absolute wrong way - there many "possibilities" in between. And many people will have many different opinons. From experienced users, to veteran mix engineers. I can only speak for myself. Though generally I'd not compress the sh*t out of a production during mastering, even if it's EDM (unless stuff is all over the place - which it isn't here). I'd rather automate the sections in question before hitting the compressor and merely do 2-3dB of gain reduction with the "broad compressor", then merely raise the loudness to the desired level and let a suitable limiter array cut of the highest peaks.

Again, YMMV and all that...



Else - interesting track. Definitely something I'd enjoy listening to while playing SciFi racing games.
[ Mix Challenge ] | [ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ]

Post

It sounds overcompressed to me. The synth drop in the middle sounds too loud because of it, as do the snares at the beginning. I prefer the first version.

Post

\@Compyfox
Thanks for replying, I was looking for an opinion from experts like you. That said, I am not rooting for the loudness war either, but I got no choice. The first version is mixed below -6.00 Dbfs. I created this to give it a boost to my songs (yes, to make them sound louder, basically). I did not use Thrillseeker to boost the low end, I made two bass instruments play at the same time on purpose; had I boosted the low end, it would have sounded too bad. I'll take a screenshot so you can see the rack.
I did boost the middle part and the high end, however
As for the else part; I'm happy to know my objective has been accomplished.
/

Everyone says that it's too compressed. OK, maybe you are right, I agree; I should increase the threshold on the first compressor. I was mostly concerned with the loss of dynamic range. I opened the project again to take a screenshot, I decided I should check how it sounds without Thrillseeker turned on. Turns out, this one device was making my song sound more bland, and the sounds lose their dynamics. I'm not saying they haven't lost any, I'm saying they lost less.

EDIT:
Here are the links:
Screenshot
The Rack

Post

You can, if you want.

But there are several things to consider :

- First, you have to know that with 2 compressors in serial, ratios are "multiplied", and not "added". So, if you set up a 4:1 ratio on each of them, you end up with a 16:1 ratio (4x4), which is a huge lot, and not a 4+4=8:1 as one may think.

- Then, ratio and threshold is one thing, but with 2 compressors in serial, time constants and compression "mode" are more important to me. You can use one fast compressor (vca, fet) to tame the peaks, and one slower compressor (opto) with longer release times to "glue" things. If your compressor allows for it, you can also play with peak / rms modes, which can make a rather big difference.
I think the point is avoiding to have two compressors doing the same thing (in such a case, better have one good than two with crap settings). You can use one for color, and the other one for more transparent duties, etc.

You can try the free Kotelnikov compressor, which features a quite interesting topology, with one single compressor handling two release times. It's not really the same than having 2 compressors in a row, but nonetheles an interesting experience.

Post

There is a approach with compressor stacking to use one compressor to grab the peaks and a second compressor more for the overall dynamic/loudness with longer release times.
So yes, it can be a good idea and if the result sounds good why not.

I personally don't like compression and so I try to use it only sparingly. I use only one compressor on the master.

Post

All right, a couple of things here.

  • You boost your signal by 14,5dB alltogether (not taking into account, that the GR of the first compressor could be 5dB - so it's about 10dB of overall boost)
  • I think the attack time for the first "glue" is a bit fast - maybe try higher attack times of 20ms or so, and a gain reduction of 2-3dB max. You want to "glue" the track, not swash it
  • due to the fast attack time, your "Bass" get's boosted intensely
  • if you want the second limiter as your "peak savety network"; then the threshold is way too low. And even at 10:1, the "SSL Bus Comp" clone does not limit (it let's rogue peaks through). Use a suitable (brickwall) limiter instead
  • if you're not into the Loudness Race/War, but still want hot tracks (since you think that -6dBFS maximum peak/-20dB RMS avg is too low) maybe try to work with -12dB RMS avg - this is a good balance between a moderate to high dynamic range, and non-squashed master - I have plenty of topics about metering tools and reference levels in my KVRMarks
  • Thrillseeker's MOJO mode can be turned off, this turns it into a regular parallel EQ that doesn't smear transients and in turn still boosts your mids and highs
I can't further comment on Ableton Live, as this is a host that I'm not familiar with. But it should give you a small insight of how you could "fix" things.


In general I'd still do this:
  • if your production is (say) at about -18dB as reference level (0VU or -18dBFS RMS realtime), you can use the first compressor at -20dB (threshold) to glue together your mix a bit (read: infamous "SSL bus comp glue trick")
  • if you want a more consistent/flattened (loudness wise) mix, use automation for the sections
  • want to use "mid range" and "high range" boost/"sparkle" that is not affected by the compressor, put it post compressor but pre the loudness raise array (Gain Module -> Limiter)
  • boost the signal post compressor with a dedicated (clean) gain plugin
  • use a suitable brickwall limiter as final instance to catch rogue peaks (I'd select -0,5dB as ceiling to clip off peaks bar minimum!)
Hope that helps


4damind wrote:There is a approach with compressor stacking to use one compressor to grab the peaks and a second compressor more for the overall dynamic/loudness with longer release times.
So yes, it can be a good idea and if the result sounds good why not.

I personally don't like compression and so I try to use it only sparingly. I use only one compressor on the master.
This trick is for example great for percussion or vocals that need more hard squashing (the infamous 1176/LA2A in row trick).

I'd not use it on a full mix myself. Here the Compressor (handling the average signal strength -> Gain module -> Limiter does pretty much the same
Last edited by Compyfox on Thu Aug 04, 2016 3:46 am, edited 1 time in total.
[ Mix Challenge ] | [ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ]

Post

I think I'll try that compressor, sounds interesting.

Post

[quote="Compyfox"]All right, a couple of things here.

  • You boost your signal by 14,5dB alltogether (not taking into account, that the GR of the first compressor could be 5dB - so it's about 10dB of overall boost)
  • I think the attack time for the first "glue" is a bit fast - maybe try higher attack times of 20ms or so, and a gain reduction of 2-3dB max. You want to "glue" the track, not swash it
  • due to the fast attack time, your "Bass" get's boosted intensely
  • if you want the second limiter as your "peak savety network"; then the threshold is way too low. And even at 10:1, the "SSL Bus Comp" clone does not limit (it let's rogue peaks through). Use a suitable (brickwall) limiter instead
  • if you're not into the Loudness Race/War, but still want hot tracks (since you think that -6dBFS maximum peak/-20dB RMS avg is too low) maybe try to work with -12dB RMS avg - this is a good balance between a moderate to high dynamic range, and non-squashed master - I have plenty of topics about metering tools and reference levels in my KVRMarks
  • Thrillseeker's MOJO mode can be turned off, this turns it into a regular parallel EQ that doesn't smear transients and in turn still boosts your mids and highs
I can't further comment on Ableton Live, as this is a host that I'm not familiar with. But it should give you a small insight of how you could "fix" things.


In general I'd still do this:
  • if your production is (say) at about -18dB as reference level (0VU or -18dBFS RMS realtime), you can use the first compressor at -20dB (threshold) to glue together your mix a bit (read: infamous "SSL bus comp glue trick")
  • if you want a more consistent/flattened (loudness wise) mix, use automation for the sections
  • want to use "mid range" and "high range" boost/"sparkle" that is not affected by the compressor, put it post compressor but pre the loudness raise array (Gain Module -> Limiter)
  • boost the signal post compressor with a dedicated (clean) gain plugin
  • use a suitable brickwall limiter as final instance to catch rogue peaks (I'd select -0,5dB as ceiling to clip off peaks bar minimum!)
Hope that helps


Thanks a lot for your help and the feedback of everyone else. If anyone else wants to try the rack for themselves, the link is pasted above.

Post

You mean an 1176 or LA2A? They're just other "vintage type" compressors, like "The Glue", which is basically based upon the SSL Console Bus Compressor. So your Ableton compressor is "based upon" a "based upon" concept.


Maybe a compressor that makes more sense for your desired task, is Tokyo Dawn Labs Kotelnikov (as mentioned above). It is a so called "full spectrum compressor" (or "single band compressor", same thing - there are no multiple bands to setup), but it has some tricks up it's sleeve that are really nice (can treat both peak and average signal levels at the same time). At least to me as Audio Engineer. Beginners might be a bit overwhelmed (and yes, your "Glue" version is still more than suitable at this stage).

And as limiter, basically use anything that does "brickwall limiting" (a compressor at 10:1 or 20:1 just doesn't cut it, unless the attack is 1 sample)


Now it's only down to "learning by doing".
[ Mix Challenge ] | [ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ]

Post Reply

Return to “Production Techniques”