Please could someone answer three questions about limiters and compressers to the dumbass noob writing this?

[Chrisulrich]

Dear Anyone.

I know these are dead basic questions, I know I should understand this already, I know I'm going to feel an eedjit when I learn the answers...... Here goes!

Compressors - as I understand them - turn sounds DOWN when they reach a point you set. As they're turning sounds DOWN, how can you use them to make sounds LOUDER? Flat don't see it!

Limiters. Same questions as the compression question above - but with a twist. I've got a limiter with 2 buttons on it, Threshold and Output Gain. I can make a track louder, when I put it on the out buss, by moving EITHER knob. Apart from one control controlling the speed of the attack, those are the only two knobs my limiter has. When do I tweak which knob and why, considering tweaking EITHER of them makes the track louder/quieter?

3. Here goes, cos I don't even understand what I'm asking here! You've got your limiter on the outbuss and a compressor on the sound (let's make it a one-sound flute mythical track to keep this easy, but you can add mythical other sounds in the answer if it helps. Dunno if it will.) You want to make the solo flute - but you can add other sounds in the answer if you want - nice and clear and decently loud. When do you just add more volume (turn it up a bit!) when would you use a compressor and when would you just use a limiter on the out buss and not bother with a compressor. And if you were just using a limiter on the outbuss, would you have this flute intentionally a bit too loud so the limiter can 'hear' it and turn it down a bit? And is that how they make their instruments/sounds sound richer/fuller? Sorry for all the sub-questions, it's just I hear pieces that use exactly the same sounds I have, but they always sound richer/fuller as a piece and I've got zero idea how they achieve that, apart from ONE person in another forum saying 'it's to do with compression and limiting'!

Yours respectfully and very puzzledly,

Chris.

 

[VomitHatSteve]

"makeup gain" is the answer to most of this.

(A limiter is just a compressor with specific settings pre-set, so I'm not really going to mention it otherwise)

You are correct that what a compressor does is turn the loudest peaks in a signal down. Notably, they do not turn them down below their threshold. If your threshold is -6dB, everything quieter than that will be (more or less) unaffected. Everything above that will be compressed by the appropriate ratio. (e.g with a 2: compression, a signal at -4 dB would be compressed to -5; a signal at 0 would be compressed to -3 db*)

Then your chain applies makeup gain. So with 3 dB of makeup gain, your uncompressed -9 and -6 dB signals are now at -6 and -3 respectively. Your -5 and -3 signals are -2 and 0. Everything is louder, but the quietest sounds are turned up the most.

* My actual math is wrong, but that's the principle

 

[UltimateOutsidr]

First off, an illustration. In the following picture the top track (blue) is mixed to be loud and the bottom track (green) is not. Even if you normalized both files (which raises the volume of the audio file to its absolute highest without clipping/losing information), the blue track will sound louder when played at the same volume as the green one.



This is because the green track has these huge peaks which prevent the audio file from being raised by more than a db during normalization. The difference between the "normal" music content and the peaks is preventing the track from being perceived as loud.

Both compressing and limiting solve this problem by shaving the tops off the waveform, giving you more headroom, meaning you can now turn the rest of the content up to make it sound louder at the same output volume setting. Limiters are usually simpler in practice, while compressors tend to give you more control over the process. (Compressors can "shape" a sound with the right settings while limiters are mainly just for keeping peaks under control, without altering the sound.)

In your example of the flute, generally you would have a compressor on the flute's channel strip for the purpose of evening out the volume of that part. Many/most instruments (and voices) vary in dynamics over the course of a performance. Some of this can be intentional, but some of it is just a natural side effect of the signal source. So you'd apply a compressor to artfully make the volume of that part more consistent, which makes it easier to dial in its appropriate level in the mix.

For example, when I have tracks with vocals, there are ALWAYS compressors on the vocal tracks (or the vocal bus) to make the singing volume consistent. And then while mixing I only have to worry about adjusting the vocal bus's fader to find the right level relative to the mix.

In my projects, I always have a brickwall limiter on my drum and instrument buses, and I adjust it to remove about 2db worth of peaks. After the limiter I add a bus compressor for "glue", where I adjust it to apply between 2.5-4.5db of compression (with an equal amount of makeup gain). On my master channel, I first have a brickwall limiter (which again applies about a 2db cut on peaks), followed by a more traditional limiter whose sole purpose is to bring the overall track up to my desired loudness. Some limiters are better at this than others (some limiters audibly distort if they apply more than db of gain reduction, while others can be pushed pretty far without destroying the signal.)

So, at the end of the day, and in their own way, compressors and limiters make tracks quieter by taming louder passages/portions of the signal, which lets you make them louder in your mix by turning them up without clipping.

 

[RFR]

First off, an illustration. In the following picture the top track (blue) is mixed to be loud and the bottom track (green) is not. Even if you normalized both files (which raises the volume of the audio file to its absolute highest without clipping/losing information), the blue track will sound louder when played at the same volume as the green one.



This is because the green track has these huge peaks which prevent the audio file from being raised by more than a db during normalization. The difference between the "normal" music content and the peaks is preventing the track from being perceived as loud.

Both compressing and limiting solve this problem by shaving the tops off the waveform, giving you more headroom, meaning you can now turn the rest of the content up to make it sound louder at the same output volume setting. Limiters are usually simpler in practice, while compressors tend to give you more control over the process. (Compressors can "shape" a sound with the right settings while limiters are mainly just for keeping peaks under control, without altering the sound.)

In your example of the flute, generally you would have a compressor on the flute's channel strip for the purpose of evening out the volume of that part. Many/most instruments (and voices) vary in dynamics over the course of a performance. Some of this can be intentional, but some of it is just a natural side effect of the signal source. So you'd apply a compressor to artfully make the volume of that part more consistent, which makes it easier to dial in its appropriate level in the mix.

For example, when I have tracks with vocals, there are ALWAYS compressors on the vocal tracks (or the vocal bus) to make the singing volume consistent. And then while mixing I only have to worry about adjusting the vocal bus's fader to find the right level relative to the mix.

In my projects, I always have a brickwall limiter on my drum and instrument buses, and I adjust it to remove about 2db worth of peaks. After the limiter I add a bus compressor for "glue", where I adjust it to apply between 2.5-4.5db of compression (with an equal amount of makeup gain). On my master channel, I first have a brickwall limiter (which again applies about a 2db cut on peaks), followed by a more traditional limiter whose sole purpose is to bring the overall track up to my desired loudness. Some limiters are better at this than others (some limiters audibly distort if they apply more than db of gain reduction, while others can be pushed pretty far without destroying the signal.)

So, at the end of the day, and in their own way, compressors and limiters make tracks quieter by taming louder passages/portions of the signal, which lets you make them louder in your mix by turning them up without clipping.

Very good explanation

 

[ashcat_lt]

If you can’t get the quiet parts loud enough without the loud parts getting too loud, compressing the loud parts down might help.

If you can’t get the average level high enough without the instantaneous peaks going too high, limiting the peaks might help.

 

[Papanate]

Compressors - as I understand them - turn sounds DOWN when they reach a point you set. As they're turning sounds DOWN, how can you use them to make sounds LOUDER? Flat don't see it!

Compressors -You compress the volume by a Ratio amount - 1.5:1 or 3:1 means that whatever comes in gets compressed down by the first number - 3:1 means whatever comes in at 1db will be compressed by 3 db - the speed in which it compresses is set by the Attack (fast to slow) - how long it compresses is set by the Release (fast to slow) - Knee is how fast (Hard) or slow (Slow) the compressors reacts at the threshold - then you use Make Up to compensate for the loss of volume - which when you’re using low compression ratios it might be 1db - 3 db - but when smashing snare drums or something might be 10 db -

Limiters. Same questions as the compression question above - but with a twist. I've got a limiter with 2 buttons on it, Threshold and Output Gain. I can make a track louder, when I put it on the out buss, by moving EITHER knob. Apart from one control controlling the speed of the attack, those are the only two knobs my limiter has. When do I tweak which knob and why, considering tweaking EITHER of them makes the track louder/quieter?

A Limiter works by setting a maximum output level for an audio signal, preventing it from exceeding a specified threshold. When the audio signal peaks above this threshold, the limiter reduces the volume of those peaks, ensuring a more consistent sound without distortion - in your case Changing the Threshold simply puts the Volume back to wherever you had it before the Limiter kicked in - likewise the output level is again changing the overall volume of sound depending on where you set it - If you Output is set to unity - 0 gain - and you increase it by 5 db then it’s going to be 5 db louder on the output and you are defeating the purpose of the Limiter (unless your purpose is to have no dynamics) - going below or above 0 db will raise or lower the volume of the tracks or Master Buss.

3. Here goes, cos I don't even understand what I'm asking here! You've got your limiter on the outbuss and a compressor on the sound (let's make it a one-sound flute mythical track to keep this easy, but you can add mythical other sounds in the answer if it helps. Dunno if it will.) You want to make the solo flute - but you can add other sounds in the answer if you want - nice and clear and decently loud. When do you just add more volume (turn it up a bit!) when would you use a compressor and when would you just use a limiter on the out buss and not bother with a compressor. And if you were just using a limiter on the outbuss, would you have this flute intentionally a bit too loud so the limiter can 'hear' it and turn it down a bit? And is that how they make their instruments/sounds sound richer/fuller? Sorry for all the sub-questions, it's just I hear pieces that use exactly the same sounds I have, but they always sound richer/fuller as a piece and I've got zero idea how they achieve that, apart from ONE person in another forum saying 'it's to do with compression and limiting'!

A Limiter is useful on tracks which have a high peak level - a Snare or simliar - You can use compression but it is IMO not as effective - On your Mythical lFlute track I wouldn’t use compression or a limiter - I don’t like the sound of a compression on wind instruments - electric guitars etc… I would because in my opinion it helps the sound - That is depending what the guitar sound is - once you get a grip on the compression settings you can achieve the results you are looking for -for example as I get more experience mixing I use less compression - not as a rule but as type of sound I want to achieve.

 

[DrewPeterson7]

Compressors - as I understand them - turn sounds DOWN when they reach a point you set. As they're turning sounds DOWN, how can you use them to make sounds LOUDER? Flat don't see it!

To try to answer this in as simple terms as possible -

A compressor only SELECTIVELY turns things down, over that set threshold. What that does in practice is lowers the volume of the very loudest bits of your audio, without lowering the rest of it.*

So, when you turn up the volume on something, since it's the very loudest bits that are the first to clip... if those are quieter, you can turn it up that much louder before you clip. If a compressor takes 5db off the top of your peaks, then if you have the track turned up so the loudest parts are at 0db, you can turn the whole thing up 5db and have it 5db louder after compression.



*super technical discussion following, OP you should probably steer clear: - my original understanding of how all of this works was that a compressor identifies only the part of that peak that exceeds the threshold, and scales it down by your ratio. So, if you have a 4:1 ratio, a threshold of -10db, and a peak of -6db, then (with an instant attack, etc) your new peak will be -9db, 1db oer your threshold instead of 4. BUT, I figured the signal below the threshold wouldn't be modified ad it weas only the components of it with an amplitude exceeding -10db that would be. I had a conversation with this once, though, with a guy who's definitely not a recording engineer or electrical engineer, but is a really bright guy, whose understanding was the reverse - that it's not possible to identify only the part of your waveform exceeding the threshold, so what the compressor does in practice is just "ducks" the signal by - in this case - 3db, and is sort of like a very smart autofader more than it is a dynamic processing effect that treats different amplitudes of the signal differently.

I honestly have no ideas which is which, and its kind of interesting. I did some preliminary testing with background noise and transients that kind of suggests he might be right, but I've never actually gotten a satisfactory explanation of mechanically what happens when a compressor engages, and t's an interesting question.

 

[rob aylestone]

This is exactly why compressors are so difficult to explain. The listener does not have the knowledge to be able to hear what it does, without explanation and explanations of compression mean nothing without hearing it. So many things in real life are compressed naturally, we often don't notice. Like the sun going up in the morning and setting in the evening. We notice the swiftness the sunrise turns from night to day and again when it goes from day to night, yet from before noon to after noon, the level is shooting up hugely, but we don't notice and get burnt. Every guitarist with any size amp knows that over 7, the volume knob doesn't do very much, bat change the sound, rather than volume.

Our ears and eyes just don't tell us the real facts of what is going on. It is as said, all to do with changes in level, so we look at the process on a graph. input level going in one direction and output level in the other - and with zero compression, as one increases, the other one does at the same rate - so it's a diagonal line. If your input goes from zero to 1 Volt, so does the output - that is unity - 1:1 expressed as a ratio. If you twiddle a knob and change the slope of the line so the output goes from zero to half a Volt, then that is 2:1 - increases in level are only half as extreme. 3:1 would be an even flatter slope, but still a straight line, yet we'd end up as a third of a Volt, output wise.

Normally audio compressors are not this blunt. We can pick a point on that straight line where the compression starts, and the line goes from 45 degrees at that point to a shallower angle. Compression might be 3:1, starting at .5V.

That's the science - this is how a basic compressor works - BUT - most people will not hear it because they do not know what to listen for. They still hear drums, or a voice. They hear pitch, they hear maybe an annoying lisp from a gappy tooth - but they do not hear compression ................ until the Eureka moment when they suddenly do! Then they realise they were listening for something totally different. All the jargon that comes with compressors - the slope, the cutoff, the ratio, the gain makeup, the threshold and even attack and release make no sense until the magic moment when the light goes on in your brain.

In the UK education system, kids are being questioned on compressors for a qualification. They draw graphs, they use the words and I bet less than 10% have actually heard it and used in creatively. When we hear submissions in our own mp3 clinic here, we often here recordings that sound 'commercial' and those that sound 'home recorded' - very often the commercial sound is just compressed in a sort of standard way. All the clever mics, audio treatment, careful mic positioning pointless if the squashing is done badly, or perhaps not at all. Maybe that is why mastering as a skill is getting lost. Maybe real mastering engineers just knew how to work compression really well?

The maths and science allow you to explain compression. Nothing anyone remotely can do let's you hear it. I started recording in 73. I think I had my compression Eureka moment in 1996.

 

[ecc83]

Compressors don't make quiet parts louder, you need an Expander for that. Piccy is the one in Samplitude ProX 6. I am sure there will be one in Reaper and probably Audacity.
In hardware terms there is really no difference between a compressor and an expander, it just really depends what you "tell" the gain changing element* to do!
Put them both together and you have a "Compander" beloved in the (awful!) days of tape recording because of its pitifully poor dynamic range (sorry Mr Beats!) Now we can have 100dB for less than a nifty, companders are rarely seen. The DBX company made their name with the most common one but they have moved on to other, better things.

Rob, I have never had a problem understanding compression but then the electronics background helps. even built a couple of simple ones. I am sure a few oldies here will know of the 2N3819 FET?

*Can be a FET, diodes, VCA, variable 'mu' valve. HowTF they do it digitally I will never know!

Dave.

 

[rob aylestone]

Dave - I think plenty of people understand the theory, but I suspect they may not actually hear it. I knew the theory for a very long time and routinely talked about it - then one day, I heard it!

 

[ecc83]

Dave - I think plenty of people understand the theory, but I suspect they may not actually hear it. I knew the theory for a very long time and routinely talked about it - then one day, I heard it!

Oh! Right! Well that's MY theory and I'm sticking to it!

Dave.

 

[bouldersoundguy]

Compressors -You compress the volume by a Ratio amount - 1.5:1 or 3:1 means that whatever comes in gets compressed down by the first number - 3:1 means whatever comes in at 1db will be compressed by 3 db

Not quite. The way it works is that the first number represents the input level over the threshold and the second number represents the resulting (target) output level over the threshold. With a 3:1 ratio, if the signal is 9 dB over the threshold, the target output will be 3 dB over. I say target because the attack and release will affect the actual output as the gain changes will lag according to those settings.

 

[bouldersoundguy]

People often misunderstand how attack and release work. They don't relate to the threshold, they relate to the target output level vs. the actual output level at any particular moment. The milliseconds marked on the attack and release controls aren't absolute times, they're the time it takes to decrease or increase the level by some amount (typically 10 dB, I think). So when the signal level rises above the threshold, the detector circuit generates a running calculation of the gain reduction needed to produce a given target level. The gain module begins applying that gain, but at a rate controlled by the attack value, until the target level and actual level match. As the input level starts to fall, the actual output level will drop below the target output level and the amount of gain reduction will decrease at a rate controlled by the release value, which can and usually does happen well above the threshold (which is why the release speed doesn't relate to the threshold). If the release value is slow enough, it may still be happening after the signal falls below the threshold, but it's still not the threshold that is triggering the release.

 

[ashcat_lt]

Compressors don't make quiet parts louder, you need an Expander for that.

Umm...no? A downward expander makes quieter parts quieter. An upward expander makes louder parts louder. Both have the effect of making quiet parts more quiet than loud parts. There actually isn't a good dedicated expander in vanilla Reaper. A couple things that can be made to do it, but it's the one thing we're really missing. I made a JS expander that works reasonably well a while back, and I think there are a couple others user-made JS ones out there.

The thing with attack and release is where things get really confused. A lot of what "normal" people do with a compressor involves like shaping the envelope of individual hits - whether they are drums hits or bass notes or whatever. Very often this involves setting the attack control such that it misses the initial transient and instead clamps down on the sustained portion. Since the transient is so much louder than the sustain, this amounts to turning down the quiet parts and sort of makes the "problem" of peak levels even "worse". It certainly doesn't help us to turn up that kick drum without clipping.

 

[rob aylestone]

I hate it when we start to talk of compression (or expansion) making things quieter or louder. All they do is reduce or increase dynamic range. Loudness is perceived as a result of dynamic range changes. I totally get it that very often compression is used alongside level changes to make a mix louder, but compression can also make a sound quieter - hence the make up gain function.

We need to make sure we use precise language, because what dynamics changes do is context related. We all, I think are on the same track when we talk about pitch, EQ, distortion, harmonics, noise and other stuff - we ALL understand posts using these terms correctly. With dynamics vocalbulary, understanding is extremely variable. Dave and Ashcat are both correct, which is impossible - but with compression and expansion it just goes wrong without going into the details.

 

[ecc83]

I hate it when we start to talk of compression (or expansion) making things quieter or louder. All they do is reduce or increase dynamic range. Loudness is perceived as a result of dynamic range changes. I totally get it that very often compression is used alongside level changes to make a mix louder, but compression can also make a sound quieter - hence the make up gain function.

We need to make sure we use precise language, because what dynamics changes do is context related. We all, I think are on the same track when we talk about pitch, EQ, distortion, harmonics, noise and other stuff - we ALL understand posts using these terms correctly. With dynamics vocalbulary, understanding is extremely variable. Dave and Ashcat are both correct, which is impossible - but with compression and expansion it just goes wrong without going into the details.

Yes, I am thinking of "expansion" in the terms of what R3 sound balancers did to smash Beet 5 through AM radio! Pulling up the quiet passages and puling down the loud. "Compression" can be applied to low levels to make them more audible or the top end to prevent overload or both

Possibly THE most famous companding system was Dolby A but let's no get into that!

Dave.

 

[ashcat_lt]

We’re aware that compansion works (to the extent that it does) specifically because compression and expansion are opposite processes, no? A compressor reduces the dynamic range - usually so that it can survive a transmission and/or storage medium that won’t support the full dynamic range - and then an expander increases that dynamic range back to something closer to the original for playback.

You can use your own colloquialisms if you want, but if you actually grab an expander - hardware or software - it’s only ever going to increase the dynamic range by making the quieter parts of the signal more quiet in relation to the louder parts. Like, you’re welcome to call a duck a dog, but if you try to tell somebody who obviously doesn’t know any better that all dogs fly and quack, I’m afraid I might have to call bullshit.

 

[rob aylestone]

This is exactly why compression has to be the most mangled subject we have in recording. Logic, science and experience all competing with each other. A guitarist with a stomp box compressor might experience an increase in volume when they hit the button, and a DAW user might get a volume reduction when they turn on a compressor. Both results happen. When you have experienced both results its easier to understand how misunderstanding happens. Its very simple to explain the technical things happening to the signal, but very, very hard to explain in words what it sounds like. The only bullshit is trying to explain what one individual can hear to another.

 

[ecc83]

I was probably wrong to call a device that brings up low level in formation an "expander" but then I don't know quite what to call it?

A compressor can increase volume if make up gain automatically tracks with gain reduction but since "limiting" is merely an extreme form of compression, the latter on its own will reduce level?

As Rob says, peeps need to play with the devices and learn their foibles and with dynamics software this has never been easier or cheaper.

Dave.

 

[bouldersoundguy]

Many compressors, especially "vintage" style ones, operate by turning up an input gain control into a fixed threshold. That does increase the level, and an output gain control is typically provided to trim the level back down. Whether you're increasing or decreasing level is a matter of how it's operated.