Bonus: Download Compression Cheat Sheet
Compression works wonders on vocals. It allows you to bring them to the front of the mix while not having to worry about their 'power note' making everyone in the room cover their ears.
No longer do you have to ride the fader up and down so that the peaks and valleys of the vocal fit in the mix. The compressor will do this for you, automatically.
Vocals have a wide dynamic range - they can sing really soft or really belt it out. A compressor takes this wide range of volume and makes it as narrow as you want it to be.
But, keep in mind, dynamic range on a vocal is good, so you do not want to cut it too much, you simply want to make it manageable.
When compression is set correctly, you reap all the benefits without any of the drawbacks.
So, let’s go through the settings on a compressor one at a time and how to properly apply them to vocals.
The threshold sets the height of the loudness ceiling. In other words, it decides how loud the signal has to be to activate the compressor.
So, if the threshold is set to -10dB, it will compress anything louder than -10dB. If the signal is quieter than -10dB (i.e. -20dB), it will remain untouched.
For example, let's say that the threshold is set to -10dB. If the signal from the vocal hits -6dB, the compressor will 'activate' and push the signal down according to the ratio. This is much like using the fader on the soundboard to bring the volume down on a singer's big note but then bringing it back up when they begin to sing at a normal volume again.
I can't give you an exact number for setting the threshold because every situation is different. However, I recommend setting the threshold so that the gain reduction meter rarely reads more than 6dB.
The gain reduction meter shows you how much the compressor is compressing the signal. On vocals, I have found that they begin to sound lifeless and over-compressed when compressing more than 6dB.
The ratio setting on a compressor decides how aggressive the compressor will respond. A higher setting (i.e. 12:1) will compress much more aggressively than a lower setting (i.e. 3:1).
When talking about a ratio of 3:1, you may be wondering what the figure actually means. Basically, for every 3dB the signal exceeds the threshold, the compressor will allow a tolerance of 1dB to pass through.
For example, let's say that the threshold is set to -6dB and the ratio is 3:1. If the signal from the vocal hits -3dB, the compressor will allow up to -5dB to pass through because of the 1dB tolerance. If the signal from the vocal hits +3dB, it is now 9dB over our threshold of -6dB. Since our 3:1 ratio allows a 1dB tolerance for every 3dB over the threshold, there is now a 3dB tolerance. So, up to -3dB will be allowed to pass through the compressor.
I could have just started by giving you a ratio setting that works well for vocals, but it is important that you understand what is actually going on. So, if the previous two paragraphs just went over your head, head back up there and read them again until it 'clicks'.
Now, I have found that a good ratio for vocals is 3:1. If you go lower than that, you may not get the compression needed. If you go higher, the vocal will most likely end up sounding smashed.
The attack setting on a compressor decides how much time it takes for the compressor to compress at full force. So, a lower number (i.e. 5ms) will cause the compressor to fully activate much more quickly than a higher number (i.e. 100ms).
For vocals, you generally want this number to be low (between 5ms and 20ms). If the number is too high, the beginning syllables may come out too strong and sound a bit odd.
The release setting on a compressor decides how much time it takes for the compressor to stop compressing. In other words, it is the amount of time the compressor stays active after being triggered.
For a vocal, if the setting is too low, it will bounce back erratically and sound unnatural. If the setting is too high, the vocal will sound lifeless and over-compressed. You will most likely not want to set this longer than .5sec.
If the option is available, I have found that setting attack and release to auto works well for vocals. I recommend going this route with attack and release times to simplify the compression setting process.
Keep in mind, you will only want to use the auto setting on vocals. If compressing a snare or some other instrument, you will want to set these manually to get the desired effect.
The output setting on a compressor does what you'd think: adjusts the volume of the output. So, if the gain reduction meter shows that you are compressing the signal by 6dB on average, you will want to set the output to +6dB to make up the difference.
This brings the signal back to where you started but with a smaller dynamic range. In other words, now the quieter moments are louder and the louder moments are quieter.
If you are using a digital sound board, compression is most likely built in for each channel. So, use what you have learned in this article to dial it in just right.
You might also enjoy: How to Use Your Digital Mixer’s Compressor
If you are using an analog sound board, you will need to buy a compressor for each channel you want to compress. I recommend the Behringer MDX2600 Composer Pro-XL.
If you know anything about Behringer, their products are hit and miss. Some are good and some aren't. This one is a good one. I have purchased four of these and we use them every Sunday on every vocal channel.
Great post! I've read several articles on compressor including the manual for the Behringer you mentioned - yours made my bulb burned the brightest. I have many Behringer products, 5 of that model compressor plus 5 of its predecessor, the 2200 for instruments.
The manual alludes to using the side chain for Voice over controlled EQ - but doesn't even hint how to do it.
Please reply w/an explanation & tips for the expander part of the 2200/2600 if that can helps w/vocals or instrument solos
Hey Tony. I have not used the Behringer compressor for "Voice Over" compression (aka "Ducking"), so cannot speak on how to set it up. However, I don't know that there is really a good application for this feature with church sound. What were you planning on using it for?
My apologies. I got suggested "ducking" settings from my church sound dude. You asked how I'd use ducking:
Sub Group one EQd and compressed to enhance solos, either voice or instrument.
The other SGs
three (Lead Instruments), &
four (rhythm sect)
each have tailored EQ curves. The "ducking" help those Solos stand out.
Makes sense. I have just never heard much about using ducking in a live setting - it is usually a recording technique.
Hi there Kade,
I was reading about the compressors and finally realised that my church has been missing this type of equipment.
I read your review on Behringer MDX2600 and was impressed with the outcome. However, I have 8 vocals. Does that mean I would have to get 4 units?
Or do you think 2 of MDX4600 would do thr trick?
If you go with the MDX2600, you would need 4 units for 8 vocals.
With the MDX4600, you would only need 2 units, but you would also loose a few features. The main thing is the desser, which is why I went with the MDX2600 in the first place. When you compress vocals, the S's can tend to come out really harsh, so the desser really comes in handy. But, if working with a tight budget, the MDX4600 is definitely better than not having a compressor at all.
For me, in a small setting like a coffeehouse, my goal has been to travel as light as possible without sacrificing a certain standard in sound quality. I have a small Fishman acoustic amp and I have a vocal pedal board with one of them being a vocal compressor pedal. All that just to say thanks because even though a pedal doesn't have all the knobs and dynamics of a mix board, the principle is the same and I definitely hear a quality improvement with your advice compared to how I was using it.
Just a question, understanding that "some" separate units are better for doing what is desired than just the features of whatever digital board you may have. Do you use these MDX2600 with your X-32? If so, what is the quality difference between the on board X-32 compressor and the MDX2600?
I used to use the MDX2600 with an analog board, but now use the X32. Based on my experience, I think the X32 offers better quality. I also enjoy everything being self contained and not having to worry about external equipment. Really, the only upside to the MDX2600 is the extra features it offers, like the desser and a few additional compression controls. But, the X32 has a desser as well, it just has to be used in one of the 8 FX racks and applied to a single channel or bus.
A useful article, however I must take you to task over the statement "The threshold setting on a compressor sets the height of the loudness ceiling. In other words, it decides the loudest signal allowed to pass through." This is not correct: what you are describing would be a hard limiter, which sets the maximum signal level. The Threshold of a compressor is simply a predetermined signal level, above which the louder signals are 'squashed' by the value of the Ratio. With a threshold of -10 dBs, a signal reaching the compressor whose level is -6 dB's will have the 4 dB's above the threshold reduced in amplitude by the value of the Ratio i.e. with a ratio of 2:1, the 'extra' 4 dB's would be output at 2 dB's, making the output level (without make-up gain) -8dB's.
Good catch, John. I have updated it to read, "The threshold sets the height of the loudness ceiling. In other words, it decides how loud the signal has to be to activate the compressor."
Please let me know the settings when auto is not available and the difference settings for a lead vocal from a bgv. I’m using a Midas 32r for live broadcasting.
Hey Jose - You can set the attack to 15ms and release to 40ms. This should work for vocals in most situations.
Two questions about compression. First of all, we have an analog A&H mixer with no compression available. For now, I'm holding off adding it because we are in the process of consideration of a major renovation in which case I'll be going to a digital board. For now, however, I have two specific situations where I thing compression could really help, that is, if I'm understanding it correctly!
Here goes—our senior pastor has a tendency to really "quiet down" as he approaches the end of his sermons. It's loud enough in the house for everyone to hear (although I coach my techs to be sure they're listening and ride the fader, if needed, so no one goes to sleep) but I find myself having to do some post-production editing for the podcast upload to the web. I do have the ability to apply compression (on just his channel) to that feed in our DSP so my question is am I on the right track for a solution? If so, what would you suggest for a starting point for settings—what I'm trying to do here is make sure his volume doesn't get BELOW a certain point rather than limit the upper end.
Similar question but on the other end—the audio feed that goes to our video stream is just fine 90-95% of the time. I only have problems when (in our traditional service) the organ opens up with choir going all out and maybe the brass is playing as well (you get the idea) or in our contemporary service the team goes super big so again I'm wondering about a starting point for compression which I have the ability to set in the streaming software?
Hey Steve - You are definitely on the right track with your pastor's mic. Keep in mind, compressing the loud moments is essentially closing the gap between the loud and soft moments, so it is the same as making sure the volume doesn't go below a certain point.
You'll find my recommended compression settings for a pastor's mic in this post.
When it comes to you live stream mix, you should just need some light compression there. Start with this and adjust from there:
- Ratio: 2:1
- Attack/Hold/Release: Auto
- Threshold: Set where loud moments are compressed about 3dB
Hey Kade.......Right from your explanation that "When you compress vocals, the S’s can tend to come out really harsh, so the de-esser really comes in handy. "........so then what's the suitable setting that can enable one to achieve the best results?
Hey Richard - I've never been able to get a great result using the X32's desser. It always sounds unnatural.
So, I've opted to back off the compressor to fix the problem instead. You can even experiment with the attack time on the compressor.
I was wondering what knee you use for compression.
I generally keep the knee setting around 2 or 3.