Is it possible to avoid Compressors?

[RobC]

Or more like substituting them?

Look, I don't like their sound, no matter what. It simply can't do what a human can do. It's still a program doing the work, judging by numbers and not by ear.

After all, "riding the fader" always was a thing. If the vocal still has quiet parts? You can quickly boost it manually.
Need de-essing? Extract the sibilance and set the levels manually. Today's tech can do what the fully hardware studios couldn't, that easily and fast.

One may say "yes, but what about the punch and the gluing effect?"

- Well, think of it this way: every sound should have its unique characteristics. They should range from soft to hard. Why? Because that's the key to getting a well balanced, dynamic mix. If everything is punchy, then after a few seconds, nothing will feel punchy anymore.

As for gluing a mix together? Honestly, I never felt the need for it. Though I am considering some multi-band saturation fun, transient-flipping effects, etc. And if there's something weak, and frequency cancellation is going on? Then I would do some quick multi-band gating trick, where the more important sound's fundamental frequency is let through, while the rest needs to shut up for that moment. That's a long story, but yeah, admittedly similar to compression, but still not the same!

As for what sound I go for? A balanced, professional one, but without the loudness war trends. My aim is to get the best possible sound at about 80 dB SPL. For which I'd do anything, obviously. : )

Do I believe it correctly, that manual "compression" sounds superior to an actual compressor?

[PhillipOrdonez]

I can't live without compression on the drums. I don't do 2bus compression and prefer riding the gain on vocals. But the drums? No, thanks, I'll have compression there all day long.

Do as you wish though.

80 dB SPL is very loud already, so watch your ears.

[avasopht]

You might want to at least put a limiter on the snare for its brief peaks. Sure, you could do that manually, but there is absolutely nothing to ... erm ... gain from that

The idea compression is all by the numbers is a bit of a misconception. You set its response. If you decide you want to drop the signal for such and such time, fader riding is no less a numbers game than setting a compressor on it. You're just inputting the numbers manually.

You don't have to apply compression to every channel, nor have the compression set the same throughout the song. If you want you can limit its application to select parts of the track.

Why not apply your manual inputs to the compressor settings?

Nothing inherently wrong with riding the fader. Try it out and see what results you get. Maybe for you, you get a better result.

It also doesn't have to be about the loudness wars. If you're riding the faders and your target just happens to achieve the same outcome, then wouldn't you still end up wit the exact same outcome (in terms of loudness wars, etc.)

[bxbrkrz]

IIRC Bruce Swedien wasn't a big fan of compressors either, so you are in good company.

[MrFigg]

Selig’s curve. Essential.

[huggermugger]

Sure it's possible to avoid Compressors. I just don't have the time.

[Chizmata]

Or more like substituting them?

Look, I don't like their sound, no matter what. It simply can't do what a human can do. It's still a program doing the work, judging by numbers and not by ear.

After all, "riding the fader" always was a thing. If the vocal still has quiet parts? You can quickly boost it manually.
Need de-essing? Extract the sibilance and set the levels manually. Today's tech can do what the fully hardware studios couldn't, that easily and fast.

One may say "yes, but what about the punch and the gluing effect?"

- Well, think of it this way: every sound should have its unique characteristics. They should range from soft to hard. Why? Because that's the key to getting a well balanced, dynamic mix. If everything is punchy, then after a few seconds, nothing will feel punchy anymore.

As for gluing a mix together? Honestly, I never felt the need for it. Though I am considering some multi-band saturation fun, transient-flipping effects, etc. And if there's something weak, and frequency cancellation is going on? Then I would do some quick multi-band gating trick, where the more important sound's fundamental frequency is let through, while the rest needs to shut up for that moment. That's a long story, but yeah, admittedly similar to compression, but still not the same!

As for what sound I go for? A balanced, professional one, but without the loudness war trends. My aim is to get the best possible sound at about 80 dB SPL. For which I'd do anything, obviously. : )

Do I believe it correctly, that manual "compression" sounds superior to an actual compressor?

as manual dynamics editing can adress the changes you want way more freely and precisely than a compressor: yes, of course it will sound better when properly done, its not even a question. its just tedious AF.

some things i do:
- when a track is purely digitally made, without any recordings, there might be no need to regulate the dynamics on anything because everything is stable and precisely designed.
- take the compressor as what it is and try to use it as a musical effect, not just something that "has to be there somehow".

[Timmy Crowne]

I think of the sounds produced by compressors as colors. There’s nothing right or wrong about red, and I can paint an entire piece without red. But I might have a really tough time conveying a red balloon if I don’t use it.

It’s similar with any effect. If I’m making some heavy stuff that “requires” a slammed drum sound or some crunchy guitar or synth part, and I want to fulfill that requirement, I may have a tougher time working without compression. I think it boils down to envisioning the sound I’m after and then determining what tool I want to use to get there. Often, it’s compression for me.

Compression doesn’t make a sound better; it makes a sound. It’s our choice whether we want that sound.

[RobC]

I can't live without compression on the drums. I don't do 2bus compression and prefer riding the gain on vocals. But the drums? No, thanks, I'll have compression there all day long.

Do as you wish though.

80 dB SPL is very loud already, so watch your ears.

I actually saw a Dan Worrall video regarding why to do the gluing on drums, which made me reconsider doing things as purely as possible. With live drums, I'd give in to compressor usage - otherwise too much work. But with my own synthesized drums, I'd rather do some manual engineering and make them work together when they hit at the same time.

Well, it's more the Bob Katz thing of setting a pink noise to 80 dB SPL as a "loudness" guide - what I meant.

You might want to at least put a limiter on the snare for its brief peaks. Sure, you could do that manually, but there is absolutely nothing to ... erm ... gain from that

The idea compression is all by the numbers is a bit of a misconception. You set its response. If you decide you want to drop the signal for such and such time, fader riding is no less a numbers game than setting a compressor on it. You're just inputting the numbers manually.

You don't have to apply compression to every channel, nor have the compression set the same throughout the song. If you want you can limit its application to select parts of the track.

Why not apply your manual inputs to the compressor settings?

Nothing inherently wrong with riding the fader. Try it out and see what results you get. Maybe for you, you get a better result.

It also doesn't have to be about the loudness wars. If you're riding the faders and your target just happens to achieve the same outcome, then wouldn't you still end up wit the exact same outcome (in terms of loudness wars, etc.)

I would rather go for a multi-band foldback-distortion type saturation to keep the original curves in some way - or first, set the limiter so I have a guide as to how much I want to tame the peaks; then I'd extract those peaks with a VST like Maths, and pull them back manually. After that, in Reason, it's easy to do a bit of group cut and paste, then crossfade them all. Yes, that's kind of like a compressor, but it's more flexible; it's not constantly on, and doesn't cause any distortion.
If it sounds good, I'd also try a "peak flipping" - if it goes above a certain threshold, invert the amplitude (so 75% would be 25%; 100% would be 0%) - this creates a "cliff" in the waveform, which we hear as a click. So even though we remove a transient, we actually may add an even cooler sounding one, while freeing up dynamic space.
That said, nothing is carved into stone, so I'm not saying I would entirely avoid a limiter - especially if I won't notice any difference. Then of course it's more efficient.

True regarding the number thing; but I rather meant that in the human vs. AI sense. Problem is, most compressors are made for live processing. In fact, I never saw a compressor that would properly analyze dynamics and would have the kind of flexibility we have with manual processing.

And again, all true with automating the compressor settings, but the truth is, I want to touch dynamics as little as possible. A compressor will go up and down no matter what. It creates deformation and distortion.

Now you got to an important point, because a compressor indeed can control micro dynamics like no human can, or has the patience to do. Not to mention that some micro, or rather nano dynamics are so short, that I will most likely never be able to tell the difference.
But the most important thing ~ I'm kind of hoping for getting a result that sounds better for everyone. It could be that compressors will win.

There's no way we'd get manually to the insane - 6 LUFS levels. And the fader riding and whatnot would rather be about just balancing dynamics, so nothing gets too quiet or too loud. So no such thing like 'how could I fit the waveform between a dynamic range of 75% and 100%.

IIRC Bruce Swedien wasn't a big fan of compressors either, so you are in good company.

It's comforting to know that I'm not entirely crazy then. x )

Selig’s curve. Essential.

I gotta look into that one ~

Sure it's possible to avoid Compressors. I just don't have the time.

Yeah, that's one of my biggest fears. x )

as manual dynamics editing can adress the changes you want way more freely and precisely than a compressor: yes, of course it will sound better when properly done, its not even a question. its just tedious AF.

some things i do:
- when a track is purely digitally made, without any recordings, there might be no need to regulate the dynamics on anything because everything is stable and precisely designed.
- take the compressor as what it is and try to use it as a musical effect, not just something that "has to be there somehow".

I imagine all the work, especially with vocals. But maybe when extracting sibilance and alike; and processing the rest by extracting bits every few decibels, then leveling... hmm, but if I do that, then it will be just approximate, like compressors. Nope, no easy way out.

Exactly! I used to do multi-band parallel compression for everything, when it only made things... not necessarily much worse, but odd. ~ That was only for loudness war sacrifice though. And unnecessary.

Effect-type compressor usage - that I understand. It definitely can do things that I manually can't.

I think of the sounds produced by compressors as colors. There’s nothing right or wrong about red, and I can paint an entire piece without red. But I might have a really tough time conveying a red balloon if I don’t use it.

It’s similar with any effect. If I’m making some heavy stuff that “requires” a slammed drum sound or some crunchy guitar or synth part, and I want to fulfill that requirement, I may have a tougher time working without compression. I think it boils down to envisioning the sound I’m after and then determining what tool I want to use to get there. Often, it’s compression for me.

Compression doesn’t make a sound better; it makes a sound. It’s our choice whether we want that sound.

Personally I prefer to color with distortion tricks. I usually saw compressors as a tool so simply even out volume levels when something was overly dynamic. But I do get now why they are popular.

---

Thanks everyone for the informations so far!

[selig]

You're speaking mainly about macro dynamics, which humans can indeed to if you don't mind the time involved. I was lucky to be able to watch some amazing engineers 'perform' automation/fader moves and learn from that experience over time. So I'm in the camp that there is nothing like a well-ridden fader for overall level control - BUT, like anything, you have to practice to get good at at, and if you do it poorly it makes things worse than using a traditional downwards compressor IMO. I should also mention I'm a HUGE fan of the 'sound' of compression, as most of the records I grew up listening to featured it prominently. But even then, I developed Leveler using upwards compression for cases where I didn't want the sound of downwards compressors but still wanted a bit of automated level control. But I digress

But micro dynamics are another thing altogether. Doing things like reshaping envelopes or controlling the all important crest factor of various tracks is pretty much impossible to do well 'by hand', either in real time or by drawing in the timeline. For one, most automation is just not responsive enough for micro dynamic control. For another, the time factor involved vs using a compressor cannot be ignored.
As for glue, I also learned to mix on SSL and add the bus compressor at the end of the project, though I never thought of it as 'glue' per se.

As for punchy, why is it an all or nothing thing? Punchy is a spectrum, from not punchy at all to TOO punchy. IMO, if you consider this another way to say 'crest factor', which is also a spectrum, then I would say it is absolutely important to be able to control individual tracks. Sometimes you are increasing 'punch' (aka crest factor), sometimes you are diminishing it. The key is to get all tracks with the right amount of punch for the mix you have in mind, and without being able to control punch you are out of luck. To date, there is no better way to control crest factor is with dynamics devices, either classic compression, downwards compression, expansion, transient control, and of course even saturation/clipping etc.

[motuscott]

One can also avoid sex before marriage but I find the reasons less than compelling.

[integerpoet]

Even before Photoshop, all photos were lies.

Even before compressors, all sound recordings were lies.

It's art. The only thing that matters is whether you like it.

[guitfnky]

if you don't like their sound, then that's that. if it's for purist reasons, that would be odd, considering the whole point of electronic effects is to do things that can't be done 'manually' (you can't manually cause things to distort in the way they do through an amp, you can't manually make your bedroom mic sound like you're in a cathedral, etc.). it would be strange to be so adamant about managing volume without compression yet continuing to use all the other tools available to do other things.

I think most of us probably do a combination of both compression and 'manual' moves when needed via automation. whether one sounds better than the other is up to the listener.

that said, whatever the rationale, anyone who's not interested in fighting in the loudness war is alright by me.

[RobC]

You're speaking mainly about macro dynamics, which humans can indeed to if you don't mind the time involved. I was lucky to be able to watch some amazing engineers 'perform' automation/fader moves and learn from that experience over time. So I'm in the camp that there is nothing like a well-ridden fader for overall level control - BUT, like anything, you have to practice to get good at at, and if you do it poorly it makes things worse than using a traditional downwards compressor IMO. I should also mention I'm a HUGE fan of the 'sound' of compression, as most of the records I grew up listening to featured it prominently. But even then, I developed Leveler using upwards compression for cases where I didn't want the sound of downwards compressors but still wanted a bit of automated level control. But I digress

But micro dynamics are another thing altogether. Doing things like reshaping envelopes or controlling the all important crest factor of various tracks is pretty much impossible to do well 'by hand', either in real time or by drawing in the timeline. For one, most automation is just not responsive enough for micro dynamic control. For another, the time factor involved vs using a compressor cannot be ignored.
As for glue, I also learned to mix on SSL and add the bus compressor at the end of the project, though I never thought of it as 'glue' per se.

As for punchy, why is it an all or nothing thing? Punchy is a spectrum, from not punchy at all to TOO punchy. IMO, if you consider this another way to say 'crest factor', which is also a spectrum, then I would say it is absolutely important to be able to control individual tracks. Sometimes you are increasing 'punch' (aka crest factor), sometimes you are diminishing it. The key is to get all tracks with the right amount of punch for the mix you have in mind, and without being able to control punch you are out of luck. To date, there is no better way to control crest factor is with dynamics devices, either classic compression, downwards compression, expansion, transient control, and of course even saturation/clipping etc.

I would probably just pick out bits of mostly vocals - maybe just every time there's a noticeable stop - and level that. At least that should work for sure without being too much work, and can be done in a fool-proof way.
Compressors can't really judge though the way we do. Though I guess things get rather apparent when the compression is pushed to extremes.
I myself am thinking about extracting the lower volumes and expanding them.

All true on micro dynamics and envelope shaping, too.
But I don't understand that gluing. Don't people worry in that case that it breaks the mix?

Crest factor it is, then. That kind of will bring me to another topic, cause there indeed is a thing like too punchy.
I'm also kind of looking for a distinguishable range, cause I didn't find much info about that. Just the obvious that the human hearing is capable of hearing really quiet and really loud sounds. Although I once thought of a 0 dB -6 dB and -12 dB range when designing sounds as hard, medium and soft hits.
My guess would be, when I design a sound, and I level it by ear to the same volume as an 80 dB SPL pink noise - for loudness reference - but I have to stop because a transient starts getting too loud, then I need to take care of it. And if the sound is too weak and flat, it's time to add some punch to it.

Luckily, single-shot sounds can be easily edited manually. Something as colorful as a vocal I can agree and understand now that one way or another, for the best results, compression has to come in.

One thing though: in Reason, I think it's pretty easy to extract parts of a recording with the help of Maths VST for example. After editing/leveling, it can be rejoined with the rest and crossfaded - meaning that we kind of can create our own flexible compressor.

One can also avoid sex before marriage but I find the reasons less than compelling.

You misunderstood, cause this isn't for traditional reasons.

Even before Photoshop, all photos were lies.

Even before compressors, all sound recordings were lies.

It's art. The only thing that matters is whether you like it.

Yes, a matter of preference is pretty much what I meant. : )

if you don't like their sound, then that's that. if it's for purist reasons, that would be odd, considering the whole point of electronic effects is to do things that can't be done 'manually' (you can't manually cause things to distort in the way they do through an amp, you can't manually make your bedroom mic sound like you're in a cathedral, etc.). it would be strange to be so adamant about managing volume without compression yet continuing to use all the other tools available to do other things.

I think most of us probably do a combination of both compression and 'manual' moves when needed via automation. whether one sounds better than the other is up to the listener.

that said, whatever the rationale, anyone who's not interested in fighting in the loudness war is alright by me.

I'm all for tech! I like to step back and see/hear if there might be another, better way.

I just thought I might have better control and a more transparent sound. Of course, there are problems, such as with bass in voice for example, when at one point it's too loud, then too quiet. That most likely will call for band compression. Although I'm not a big fan of that it makes things sound a bit unnatural.

Yeah, the loudness war is just silly.