Normal compression VS Parallel compression ...a quick question

[leoaudio13]

Hi guys,
I'm currently studying about audio engineering and I'm in the middle of a research about Parallel compression in modern Rock. Would it be possible for some pros here give me some advice on what would be related to once we talk about parallel compression ? In my opinion, it's all about dynamic range, transition ( specially on drums). Would it be the best to make a test to compare b/w normal compression and parallel compression by using some plugins to measure the RMS and dynamic range of a Rock track ? Advices and suggestions would be deeply appreciated.

[dualflip]

Originally Posted by leoaudio13

Hi guys,
I'm currently studying about audio engineering and I'm in the middle of a research about Parallel compression in modern Rock. Would it be possible for some pros here give me some advice on what would be related to once we talk about parallel compression ? In my opinion, it's all about dynamic range, transition ( specially on drums). Would it be the best to make a test to compare b/w normal compression and parallel compression by using some plugins to measure the RMS and dynamic range of a Rock track ? Advices and suggestions would be deeply appreciated.

Its always about dynamics, but the thing is this, with parallel compression you get the compressed sound plus the uncompressed sound (or less compressed) which translates into a compressed and tight sound with less sacrifice over dynamics, since you are hearing the compressed sound plus the more natural sound. This also translates into like a psychoacoustic effect of hearing a sound with the advantages of compression but with a more natural feel, youll be like "it sounds very compressed but uncompressed at the same time " kinda thing, which is a very good way of enhancing an intrument(s) timbre and dynamics, with out making it feel all squared out and life-less

My 2 cents!

[leoaudio13]

Originally Posted by dualflip

Its always about dynamics, but the thing is this, with parallel compression you get the compressed sound plus the uncompressed sound (or less compressed) which translates into a compressed and tight sound with less sacrifice over dynamics, since you are hearing the compressed sound plus the more natural sound. This also translates into like a psychoacoustic effect of hearing a sound with the advantages of compression but with a more natural feel, youll be like "it sounds very compressed but uncompressed at the same time " kinda thing, which is a very good way of enhancing an intrument(s) timbre and dynamics, with out making it feel all squared out and life-less

My 2 cents!

Thank you for your reply. About the psychoacoustic effect, would it be possible that harmonic contents are related ? I think that dynamic is the most important thing in compression but would it be something else that might change when we apply paraleel comp./normal comp. technique to the original signal ?

[jinksdingo]

Yeah sure,
An ensemble of a drum kit pretty much covers the whole dynamic range.
An excellent player naturally hits a drum with varying velocities including and especially the kick. There include accents and if your present your likely to tap your foot or sway in response to its feel.
Now capture that ensemble with mics and you begin to move away from the effect of the ensemble compared live.
If you straight out compress most of the mics used you change the effect of the drums. The varying velocities disappear as louds get squashed and quiets gets louder. This was useful in the days of recording to tape but even a CDs dynamic range is much less then that of the live dynamic range.
Tape also had a self noise and compression became a way to increase the signal to noise ratio.
We have since become accustomed to the solutions to recording to a constrained dynamic range even to the point of dampening the drums themselves to control their resonances as resonances get louder under compression.
So modern rock drums go for whack and punch but to retain some of the musicality from its player no compression allows the dynamics of the playing thru.
Parallel compression is a mixture of uncompressed and compressed.
Use an aux send or more with an inserted compressor or more and mix this in with the uncompressed.
Some plugin compressors have mix control to allow a portion of non compression thru.

[jinksdingo]

Excellent question!
The harmonics would be effected.
Harmonics are under and overtones combined with the fundamentals which make up the tones you hear.

A compressor is a "voltage volume turn it downer".
You can determine when it begins to turn it down and when it resumes from turning it down, even how quickly it does it turns it down so you can push more voltage into it to be turned down.

As such harmonics below the compressor threshold and beyond the attack setting would be increased proportionally along with the fundamentals.

[Nordenstam]

Parallel compression under microscope - the original post is off, but the rest of the thread pretty much says it all.

[bob katz]

Originally Posted by jinksdingo

Excellent question!
The harmonics would be effected.
Harmonics are under and overtones combined with the fundamentals which make up the tones you hear.

A compressor is a "voltage volume turn it downer".
You can determine when it begins to turn it down and when it resumes from turning it down, even how quickly it does it turns it down so you can push more voltage into it to be turned down.

As such harmonics below the compressor threshold and beyond the attack setting would be increased proportionally along with the fundamentals.

Standard "downward compression" begins at the TOP of the dynamic range (highest levels) and brings them down. Parallel compression begins at the BOTTOM (lowest levels) of the dynamic range. But it's possible to "morph" between the two by raising the gain of the parallel compressor that you mix in. The higher the gain of the parallel compressor, the more it begins to affect the midlevels and eventually the higher levels of the signal. When you reach the point where the parallel compressor's mix gain is the same as the source, you are probably affecting levels JUST BELOW the top but still (hopefully) not affecting the highest levels of the source so it can still "sing" a little bit.

The ratios and thresholds of the parallel compressor that you mix in also affects the result. Hope this helps,


BK

[lucey]

I've had a few people prefer it and request it, where they like the depth of the blended signals (direct/compressed). It had nothing to do with RMS, or drums IMO.

These clients were doing what I'd call modern rock, as in radio/pop rock. And they liked the thickening of the mix, akin to the time smear of a transformer, but ... different. Generally that smear is a bad thing.

[leoaudio13]

Thank you jinksdingo and BK for your answers. however, I just read thru what BK just replied, "When you reach the point where the parallel compressor's mix gain is the same as the source, you are probably affecting levels JUST BELOW the top but still (hopefully) not affecting the highest levels of the source so it can still "sing" a little bit ", should it be considered at a certain point, parallel compression would probably make no different change in contents, let say a drum stem, with downward compression technique ? I didnt mean there's no change in the way how we listen to the sound of drum stems but in way that we will actually realize that they actually both give the same quality of a drum stem? I've created a test today.
- Mix 1 : drum with normal compression techniques
- Mix 2: drum with no compressors inserted on separate track but a compressor was inserted on the Drum Buss with a ratio of 2:1 with a gain reduction of no more than 3 dB.
- Mix3: Drum stem with parallel compression applied with around 15dB of gain reduction with a very fast attack and fairly slow release, just to get the a audible pumping effect.
I then record those mixes back into ProTools and normalized them with 0 dbFS in Peak just to equal their peaks. I then combined those drum stems back into my session with guitar, bass, etc ...and compared the sound. I was so surprised that the stem with Buss compression actually gave me a better sound quality. Are those steps that I did on the right track ? I was a bit confused cos' it seems to sound better with the parallel compressed stem on itself rather than when combined it back to the mix. :(

[Silver Sonya]

Among other things, the technique is psychologically relieving to those people who don't trust their own ears/judgement/skill with compressors.

All else fails, "sneak a little" of the un-destroyed signal in. I mean, the fidelity can't be all gone, right?

It became very trendy in the last few years and has resulted in a lot of weird-sounding, slightly smeary (as Brian says above) music, where you can't quite put your finger on what is wrong exactly. Just doesn't sound very good.

This kind of insecure mixing is in contrast to great masters (and celebrated eccentrics) like Tchad Blake, who use the technique daringly and "painterly" with a sense of purpose and design. This is the way to use it. Like an artist.

Unfortunately, the spirit behind it is often halting/hedging/timid/over-careful.

Generally I encourage people who are new and learning not to resort to this degree of complexity first. Try to get a real command of attacks and releases and ratios and different topologies of compression first. Get the basics down. This can take many, many, many years! Arguably a lifetime of learning, the way people study the sitar (or architecture) all their lives and don't consider themselves masters until they're in their 70s. Y'know? A lot to figure out before you start gettin' fancy.

Sense you're a student, it might be interesting to try to research/discern what the earliest uses of deliberate parallel compression were in the history of recording. Me, I really don't know. No clue. Were they using it on Beatles records? I don't think so. But I'm not an expert. My sense is it's a fairly recent phenomenon in the (still young) history of recorded music.

cheers,
c

[Adam Dempsey]

It feels to me like impressions of the use of this technique are becoming overly complicated and maybe this thread at least needs a link to this fairly well known article from 1997 by Richard Hulse, originally in 'Studio Sound', 1977.

It has been around since at least the early sixties and dozens of companies have used it as the basis of a wide range of products.

In short:
Series compression = regular downwards compression.
Parallel = upwards compression, usually minimum 2:1 ratio, fast attack to prevent overshoot of the two paths, and no gain reduction during the softest passages of the source signal. Result: as your level exceeds the threshold, less compressed signal is in your final output.

As mentioned, getting an ear for general compression first is paramount. Hear it first in your head: do you need to tame the dynamics? Would EQ be better, if it's a spectrum balance issue? If compression, do you need to tame things from the top down or raise detail from the bottom up and minimally affect the upper dynamics?

Suggestion: blend it into your dry (uncompressed) signal, rather than blending the dry into the compressed. Yes, it can be subtle, if you're talking mastering and not the slammed sound of side chain drum buss compression. Above all, keep it musical.

[dave-G]

I get a sense we're over-caveating this whole concept. Yeah, there are some pitfalls to the technique, but IME, the people who would end up in the pits are the same people that don't quite 'get' making straight/'series' compression work well either.

Ultimately, if you know what you're after and can hear what you're doing, parallel compression can be a powerful technique for a broad range of effects.

-dave

[danika]

I've seen several implementations where parallel compression is combined with high and low emphasis eq. In fact, in Ableton Live 8 there is a factory rack chain preset called Brooklyn Compression that uses high and low cut filters to roll off everything between 125 Hz and 4 kHz (which is a bit extreme IMO). What's the intent in doing that?

[jinksdingo]

The gain reduction is effecting your program material as determined by the low end voltages.
By rolling off the low end you can effect the area you desire.
When combined with the unaffected signal your low end is as it is.
Likewise, you may not want or need the higher frequencies of your material to be limited with compression so filter them out.

leoaudio

Debatable.
On one side you have those whom feel that careful use of the attack can allow the transients thru and timing the release of the compressor can occur in a similar way to parallel.

On the other, you have those attune to the differences which can be applied to each method.

[Perry Grinn]

Originally Posted by bob katz

The higher the gain of the parallel compressor, the more it begins to affect the midlevels and eventually the higher levels of the signal. When you reach the point where the parallel compressor's mix gain is the same as the source, you are probably affecting levels JUST BELOW the top but still (hopefully) not affecting the highest levels of the source so it can still "sing" a little bit.

The ratios and thresholds of the parallel compressor that you mix in also affects the result.

How about the attack time? If I'm using a slower attack time, letting peaks through, they mix in with the source at perhaps much higher levels than the source's peak value. This in theory is no long upward compression is that correct?

Would you need a zero attack time theoretically to actually produce upward compression?

[flatfinger]

Isn't it about time to bring in the terms " macro dynamics " and " micro Dynamics " ??????


This article ( though guilty of using different nomenclature !!) is a pretty scholarly dissertation http://homepages.paradise.net.nz/~rh.../sidechain.htm

One thing to keep in mind is that there is almost never a one size fits all solution . bringing up the " troughs " might not be something you want to do with certian songs/mixes.
Sometimes a good thrashingof the peaks is in order!!!!!!!!

[flatfinger]

Originally Posted by leoaudio13

Thank you for your reply. About the psychoacoustic effect, would it be possible that harmonic contents are related ? I think that dynamic is the most important thing in compression but would it be something else that might change when we apply paraleel comp./normal comp. technique to the original signal ?

By the way ,

when you start asking about the "psychoacoustic effect" and drums , you are into sound design territory , and , a compressor is the toy for designing individual components of drum sounds !!!!!!!! you can use a sampler and Layering ( combined w/gated reverb !) to make rock ( and hip hop !) musics famous " Drums from hell " !! Once you start to get into it , you realize that some of the snare sounds( for example) from the realm of " Loudspeaker music " are pretty ridunckulas!! Thundering drums that could never exist in the real physical world !!!


( Sorry for the Tangent but you did mention rock !!!!!!!)

[flatfinger]

Originally Posted by Silver Sonya

Among other things, the technique is psychologically relieving to those people who don't trust their own ears/judgement/skill with compressors.

All else fails, "sneak a little" of the un-destroyed signal in. I mean, the fidelity can't be all gone, right?

It became very trendy in the last few years and has resulted in a lot of weird-sounding, slightly smeary (as Brian says above) music, where you can't quite put your finger on what is wrong exactly. Just doesn't sound very good.

This kind of insecure mixing is in contrast to great masters (and celebrated eccentrics) like Tchad Blake, who use the technique daringly and "painterly" with a sense of purpose and design. This is the way to use it. Like an artist.

Unfortunately, the spirit behind it is often halting/hedging/timid/over-careful.

Generally I encourage people who are new and learning not to resort to this degree of complexity first. Try to get a real command of attacks and releases and ratios and different topologies of compression first. Get the basics down. This can take many, many, many years! Arguably a lifetime of learning, the way people study the sitar (or architecture) all their lives and don't consider themselves masters until they're in their 70s. Y'know? A lot to figure out before you start gettin' fancy.

Sense you're a student, it might be interesting to try to research/discern what the earliest uses of deliberate parallel compression were in the history of recording. Me, I really don't know. No clue. Were they using it on Beatles records? I don't think so. But I'm not an expert. My sense is it's a fairly recent phenomenon in the (still young) history of recorded music.

cheers,
c

+1

Spot on SS . Another thing that seems like it should be so easy and obvious is the Moving of controls during mixdown ( not automation; though thats still bettter than nothing !!).

There is no rule that says you can't change that compressors threshold or ratio DURING the final printing !!!!!!!!!

[Eloheim]

Here's how I think of using parallel compression: The compressed signal is smashed to fatten the signal and bring up the tracks' ambiance that you couldn't even hear in the first place. The dry version is there to maintain the transient 'pop' of the original. Also, compression can destroy some of the groove of the track, because the LOUD-soft-LOUD-soft, would become less distinct, ie. LOUD-SOFT-LOUD-SOFT, for example. So using the dry version might help keep some of the original dynamic groove in place. (Although proper/in tempo release times can do wonders as well...)

To apply this to something like drums, I'd suggest using fast attack and release on the compressed version. Make sure the attack is fast enough to flatten + fatten the transient to give it a broader punch, enough that you can bring it's level up without the transients poppin out all over the place. As for the release, try to set it fast enough so that the compressor is releasing just AS the sound is decaying away. This way is almost like upward expansion, because you're (in effect) RAISING the level of the drum hit's tail, allowing you to hear room/ambiance/subtlety that wasn't apparent in the first place.

Now you can mix the wet/dry levels so you get the fatness and punch of the compressed version, ALONG with the pop and dynamics of the original.

I'm sure there are countless other ways to approach the topic at hand, but the above is how I tend to think about it most often.

Peace.

[Adam Dempsey]

Originally Posted by Perry Grinn

How about the attack time? If I'm using a slower attack time, letting peaks through, they mix in with the source at perhaps much higher levels than the source's peak value. This in theory is no long upward compression is that correct?

See my post #11 above.

And for the record (literally), I've not felt the need to use the technique for some years.

[Jerry Tubb]

I never use PC.

JT

[dcollins]

Originally Posted by Jerry Tubb

I never use PC.
JT

You really need to get with the 20th century, Jerry. Next you're going to tell us you don't use multi-band parallel expansion either.

Luddites.


DC