Please explain compression to me

Discussion in 'Effects, Pedals, Strings & Things' started by Dr. Lo, Feb 14, 2015.

  1. Dr. Lo

    Dr. Lo Supporting Member

    Messages:
    1,225
    Joined:
    Dec 31, 2006
    Location:
    Canada
    I'm feeling really dense. I have an Empress Compressor and would like to better understand what it's doing. Here are my questions:
    1. If the INPUT knob controls the amount of signal being fed into the compressor circuit, and compression reduces the gain of the signal, then why does my volume increase (as opposed to decrease) when I turn up the input knob?
    2. If compression reduces the gain of loud signals, then why does it also make softer signals louder? Does this not imply that there is some boosting going on somewhere? (this may be related to my first question)
    3. Does the MIX knob control the mix of the signal that was fed into the compressor circuit (as determined by the INPUT knob) and the signal that was not at all fed in to the compressor circuit?


    Cheers!
     
  2. Dr. Lo

    Dr. Lo Supporting Member

    Messages:
    1,225
    Joined:
    Dec 31, 2006
    Location:
    Canada
  3. Lolaviola

    Lolaviola Member

    Messages:
    8,044
    Joined:
    Oct 13, 2005
    Location:
    Eastern Standard Time
    I don't have that compressor. But don't overthink it. If you just think of it from a guitar pov, then you think about the envelope of your playing/strumming/picking. Most guitar compressors can modify the envelope of sound which is LOUD>soft to more guitar-friendly variations on this Typically it's Loud>Loud or even soft>LOUD.
    If you get my drift, the next thing to listen for is how it is affecting your treble on the overall signal. It may sound more natural boosting the threshold vs. the input.
     
  4. gearscrubs

    gearscrubs Supporting Member

    Messages:
    3,690
    Joined:
    Oct 10, 2013
    Compression, as its very primary function, reduces peaks in volume. So, when a signal level crosses a set threshold, it is attenuated according to the parameters of the compression circuit (ratio, knee, etc.). By increasing the input gain into the compressor, you are compressing more peaks, but also raising the mean/perceived volume. That's why it seems louder - it is louder, but with less difference between the loudest and softest parts.
     
  5. jonnytexas

    jonnytexas Supporting Member

    Messages:
    3,276
    Joined:
    Apr 23, 2014
    Location:
    Houston
    It squashes the signal....peaks lowered, valley are raised. When a note begins to decay naturally, the compressor boosts the signal, hence the sustain. Gearscrubs perfectly explained why it can be louder.
     
  6. chrisjw5

    chrisjw5 Supporting Member

    Messages:
    9,317
    Joined:
    May 22, 2003
    Try these 3 articles, they're good places to start.

    GW

    UG

    GOO
     
  7. jocfan1

    jocfan1 Member

    Messages:
    685
    Joined:
    Feb 3, 2012
    Location:
    New Philadelphia, OH
    The explanations above cover it pretty well. As a fellow Empress owner (and it's a spectacular compressor IMO), allow me to address your questions as they relate directly to this particular unit.

    Gearscrubs nailed this one, but you should also check the ratio setting. The Empress isn't a hard limiter, meaning once you hit the compression threshold rather than totally choking off the note it reduces it according to the ratio you've set. A 2:1 ratio means for every 2dB above the threshold, the output is reduced to 1dB of output; 4:1 means 4dBs get squished to 1, etc. If you feel like your output signal isn't getting compressed enough, increase the ratio. Conversely, if you want less squish, switch to a lower ratio. Personally I don't use more than 4:1 live, but higher ratios are common in studio work.

    They are indeed related questions. This is where your input knob comes into play. If you want your softs to stay softer, lower the input volume so it takes more to hit the threshold. The metering LEDs in the Empress are particularly helpful in this regard, as you can use them to see exactly how hard you have to play before the compression kicks in. This lets you really dial the response to exactly where you want it.

    I could be wrong, but I believe the mix knob is post-input. Otherwise yes, it's a wet/dry blend: fully clockwise is only the compressed signal, while fully counterclockwise is just the dry, uncompressed signal.
     
  8. DaveKS

    DaveKS Member

    Messages:
    15,854
    Joined:
    Feb 26, 2013
    Location:
    KS
    1. It's there so you can set the way it behaves with any certain guitar/pu, a compressor has a trigger, it behaves a certain way when a certain level of signal (high or low) is presented to it. It's a trim, old vintage Strat with low out put, you set it here, modern high output bridge bucker you set there.

    2. Your describing of a limiter, not a compressor. A compressor works on both ends of the volume range. Think of the range of your vol knob, it goes from 0-10 normally, with comp on now it only goes from 2-8. Soft picking is now louder, hard picking is now quiter, it limits your available dynamic range.

    3. Mix is exactly as it sounds, it blends effected signal with un-effected signal. Think of wet/dry mix on most delays, it's the same thing.
     
  9. colonoscopotamus

    colonoscopotamus Member

    Messages:
    589
    Joined:
    Sep 17, 2012
    Location:
    Birmingham, AL
    You always hear this, and it's what is misleading. A compressor (or pure compressor action) does NOT raise valleys. It squashes peaks by a set ratio. Any signal over the threshold is reduced by the ratio setting - so, for example, a peak that is six dB over the threshold with a 2:1 ratio will actually come out as 3 dB over the threshold. Higher ratios mean more squish, lower ones mean less.
    The signal gets louder when you turn up the input because, unless you've got an infinite ratio setting, peaks that were 3 dB over the threshold are now (for example) 6 dB over the threshold. With a 2:1 ratio, the initial peaks were squashed down to 1.5 dB over the threshold, and the new peaks are squashed down to 3 dB over the threshold. Still louder.
    So that's just pure compressor action. Most (probably all, I don't know of an example of one that doesn't) compressors have built in make-up gain. This is where the "raising the valleys" thing comes in. Assume your original signal peaked at 10 dB and the lowest your signal got was 0 dB. Now apply a compressor with a 2:1 ratio and a threshold at 6 dB. This means that your loudest peaks will now only reach (((10-6)/2)=2 dB over the threshold) 8 dB. Your quietest signals will still be at 0 dB, since they aren't loud enough to activate the compressor. Now, apply makeup gain so that your compressed signal gets as loud as your uncompressed signal - just add 2 dB to the compressed signal. Now your loudest peaks are at 10 dB (as before), but your lowest signals are at 2 dB - 2 dB louder than they were before compression.
    The above all assumes instantly acting , perfect compression. In the real world, compression can't start immediately as the signal crosses the threshold (unless the compressor can tell the future), and it can't stop immediately as the signal falls below the threshold (again, without predicting the future). You can get damn close, but you can't get immediately there. Further, by lengthening out the time it takes for the compressor to start working (attack control) or to stop working (release control), you can change the sound of your signal. An example is lengthening the attack so that the initial transient (pick attack on the strings) is not compressed, but the loudest part of the sustain of the note is. This makes the sound of picking a note stand out (because its uncompressed) in relation to the sustain (which is compressed) - it's where the "clicky" chicken-picking country compression comes from.
    So, since the compression curve is necessarily imperfect, and because the attack and release controls allow you to make the compression audible, you might find that the ideal compression amount adds just a little too much of some audible compression artifact to your signal. Hence the mix control, which allows you to mix your original signal back in with the compressed (and gained up) signal. This way, you can get the punchiness, increased sustain and / or attack of the compression, etc, while still maintaining the dynamic or tonal characteristics from your original tone that you want to keep.
     
  10. colonoscopotamus

    colonoscopotamus Member

    Messages:
    589
    Joined:
    Sep 17, 2012
    Location:
    Birmingham, AL
    Nah - there are multiple ways to generate a similar sounding signal. Transient information (on any instrument) is a lot closer to full-frequency than the sustain of the instrument - the sustain is where the character of the instrument really starts to appear. Guitar's character is generally defined in the very mid-frequencies you were likely cutting, while higher frequencies (like those you boosted) only appear in large amounts (or are most prevalent) during the initial attack. So, cutting those mids helped to reduce the prevalence / volume of the sustain of the note (as well as the initial attack, which also contained those frequencies), while increasing the highs helped to increase the prevalence/volume of the initial attack (which contains a lot of those frequencies) more than the sustain of the note (which contains far fewer of those frequencies to begin with).
    So, while compression alters the relative volume of the attack vs. sustain of the note by dynamically changing the volume as the note is played, the EQ does the same thing by statically reducing those frequencies prevalent in the sustain and increasing those prevalent in the attack.
     
  11. embrionic

    embrionic Supporting Member

    Messages:
    313
    Joined:
    Nov 18, 2009
    Lot's of good answers here, but i will take a pass too, to see if it helps explain what's going on.

    The manual for the empress does a good job explaining some of the basics of compression.

    1. The empress compressor uses a design with a "fixed threshold". The threshold is the point at which the compressor will act upon the signal. If the input signal goes above the threshold, signal is reduced according to the ratio you have selected. Another way to think about a compressor pedal is even if you stomp on the footswitch and the led goes on, the pedal might not be "on" at all! The "on" switch for a compressor is really the "threshold". Below the threshold... no compression (even with the pedal switched on). Above threshold... compression. Anyway... as noted above, if you raise the level of your input so some or all of your playing goes above the threshold , then you will get compression. Some compressors have an input and threshold control, but on the empress, you raise the level of the incoming signal until it exceeds the fixed threshold point (not unlike many legendary compressors such as the UREI 1176). So even though there is no knob or button for "threshold" on the empress, it is there and the end result is the same.

    2. The more you raise the input, the more compression you get. The ratio tells you what goes in (the first number) and what comes out (the second number) but only for signals that exceed the threshold. If you raise the input so you are above the threshold by 10db, and you have the ratio set to 10:1, the compressor will "squash" or "turn down" 9 of those db's so that only 1db comes out on the other side. Keep in mind that now, anything below the threshold is also now boosted 10db, but since the compressor did not act (below the threshold) on those "quieter" signals indeed the softer stuff now seems 10 db louder compared to with the pedal off. The higher the ratio the more audible the compression is since it is reducing the signal in a more extreme fashion. So yes, there is "some boosting going on" because you turned them up with the input knob.

    3. The mix knob on the empress lets you blend back in totally unaffected uncompressed signal with the compressor settings/sound you have created.

    Not sure what you are seeking in compression, but if you want a really transparent compression setting, the Empress is pretty sweet and you can also meter gain reduction to help set that up. Move the switch to "gr" and raise the input gradually with a 4:1 ratio until you just see 3-6 db reduction for your loudest playing. That with the attack and release about in the middle (10 o'clockish) should make for a nice smooth setting.


    Compression can be tricky for those just learning about it. True story... my very first ever guitar pedal I bought ages ago was an EHX "Black Finger". I brought it home, excited to change the world with mind blowing effects, and thought the pedal was actually broken and did nothing. I brought it back and bought a deluxe electric mistress that I still have :)


    Good luck... great compressor!
     
  12. jordane93

    jordane93 Member

    Messages:
    8,903
    Joined:
    Mar 10, 2013
    Location:
    Long Island, NY
    Search JustNick's video on youtube about compression.
     
  13. ryandfl

    ryandfl Supporting Member

    Messages:
    1,542
    Joined:
    Jul 12, 2012
    Location:
    South FL
    I'm sure this has been covered, but you're not thinking about compensation / output gain.
     
  14. Dr. Lo

    Dr. Lo Supporting Member

    Messages:
    1,225
    Joined:
    Dec 31, 2006
    Location:
    Canada
    Thanks for the input folks! Based on what I've learned, the answers to my initial questions seem to be as follows:
    1. If the INPUT knob controls the amount of signal being fed into the compressor circuit, and compression reduces the gain of the signal, then why does my volume increase (as opposed to decrease) when I turn up the input knob? ANSWER: Increasing the INPUT control leads to greater volume simply because I am letting more of my signal (volume) into the pedal. To the extent that the OUTPUT control is set above unity, the result will be volume above unity.
    2. If compression reduces the gain of loud signals, then why does it also make softer signals louder? Does this not imply that there is some boosting going on somewhere? (this may be related to my first question) ANSWER: While compression certainly reduces peaks in volume, it does not increase valleys (low volume signal). However, as peaks are reduced, the dynamic range (dB difference) between peaks and valleys is reduced (squashed). Thus, as we turn up the pedal's OUTPUT control, we are increasing the volume of those valleys without increasing the volume of the peaks (as much). That's why we hear (what used to be) low volume sounds, like harmonics, more easily when compression is applied and pedal output is increased.
    3. Does the MIX knob control the mix of the signal that was fed into the compressor circuit (as determined by the INPUT knob) and the signal that was not at all fed in to the compressor circuit? ANSWER: Yes
     
  15. colonoscopotamus

    colonoscopotamus Member

    Messages:
    589
    Joined:
    Sep 17, 2012
    Location:
    Birmingham, AL
    By Jove, I think he's got it.
    Enjoy the empress, it's a great pedal.
     

Share This Page