Question about parallel compression from Bob Katz's book

[wildplum]

I am not new to parallel compression, having used it many times in mixing (the old Motown trick, as we used to call it). However, I am just starting to use it in mastering. After reading Bob Katz’s book, I am confused about one point.

In the book, he states that “Output level or makeup gain adjusted to taste.” (so far, so good) “With parallel compression off (-% gain), there is no compression.” (ok, “off” through me a little, but I understand that with its gain set to zero, the parallel compressor makes no contribution to the resultant mix of processed and unprocessed signal.) Bob continues: “Above about -5dB, compression will be very noticeable … A nice subtle compression can be achieved with settings of-15 through -5 dB.” This is the part that I am unsure of. By “-15 through -5 dB” is he referring to the output level of the parallel compressor relative to the uncompressed signal or, perhaps, the gain reduction of the parallel compressor, or something else (I guess not dBFS as he does not state it that way)?

Also, how do I interpret or translate this (the “-15 through -5 dB” part) with a compressor such as the PSP Master Comp, which achieves parallel compression by applying the processed signal amount as a percentage of the of the resultant mix. What percentage range equates to “-15 through -5 dB”?

[bob katz]

Originally Posted by wildplum

I am not new to parallel compression, having used it many times in mixing (the old Motown trick, as we used to call it). However, I am just starting to use it in mastering. After reading Bob Katz’s book, I am confused about one point.

In the book, he states that “Output level or makeup gain adjusted to taste.” (so far, so good) “With parallel compression off (-% gain), there is no compression.” (ok, “off” through me a little, but I understand that with its gain set to zero, the parallel compressor makes no contribution to the resultant mix of processed and unprocessed signal.) Bob continues: “Above about -5dB, compression will be very noticeable … A nice subtle compression can be achieved with settings of-15 through -5 dB.” This is the part that I am unsure of. By “-15 through -5 dB” is he referring to the output level of the parallel compressor relative to the uncompressed signal or, perhaps, the gain reduction of the parallel compressor, or something else (I guess not dBFS as he does not state it that way)?

Hi, guys. "Off" means "negative infinity". -15 dB means that the output gain of the parallel compressor OR the mix fader for the parallel compressor is being mixed in at -15 dB relative to the main signal. The gain reduction is a separate item.

I think you've got the first edition, which outlines what I now call "transparent parallel compression". The first edition probably doesn't mention lookahead but if you want to get the most transparent-sounding parallel compression, you also need some lookahead in the compressor to catch the earliest transients. It seems counter-intuitive to many people that the fastest attack possible (even in microseconds) is a good thing, but for invisible-parallel, that's exactly how it should be used.

In the second edition I discuss a second approach to parallel compression, which I call "attitude" parallel compression.

Also, how do I interpret or translate this (the “-15 through -5 dB” part) with a compressor such as the PSP Master Comp, which achieves parallel compression by applying the processed signal amount as a percentage of the of the resultant mix. What percentage range equates to “-15 through -5 dB”?

The PSP compressor's parallel approach is very confusing to me. I do not like a system that raises the parallel while lowering the dry, in my opinion it's not very ergonomic to use. I would suggest you put in some test tones, and try to psyche it out from there.

Personally I prefer a system where the dry stays constant (0 dB) and the parallel is mixed in to taste and measured in dB. If the total level goes over when you add in more parallel, then an additional output level after the mix can be applied.

BK

[wildplum]

Thanks Bob, that does clear it up.

"I would suggest you put in some test tones, and try to psyche it out from there. "

An excellent suggestion. I will give it a try when i get some free time and post the results.

[F. Davis]

I tend to keep my parallel compression at around -28 to -30 below the un-compressed signal. It is a very subtle effect the way I use it though.

What level do other people here tend to average their parallel compression?

Also ITB what compressor plugins do you all use for the parallel compression (disregarding compressors with the parallel built in such as PSP stuff)?
I'm using the McDSP G-Console comp to good effect.

Also any opinions regarding linear phase processing with this perhaps?

[Herbeck]

I have tried waves LP Multiband and Cakewalk LP-64 Multiband for
Linear phase multiband parallel compression (LPMBPC). =)
But I found the sound of them a bit toyish, not good enough for Mastering.
However I am curious about UAD Precision Multiband, it has a dry/wet knob like PSP MasterComp.
Maybe not the best solution but better than nothing.
I have seen Bob Katz mention some UAD plugs should be good enough for Mastering,
is the Precision Multiband one of them?

Herbeck

[csl]

Originally Posted by bob katz

The PSP compressor's parallel approach is very confusing to me. I do not like a system that raises the parallel while lowering the dry, in my opinion it's not very ergonomic to use. I would suggest you put in some test tones, and try to psyche it out from there.

Personally I prefer a system where the dry stays constant (0 dB) and the parallel is mixed in to taste and measured in dB. If the total level goes over when you add in more parallel, then an additional output level after the mix can be applied.
BK

Yes! Agreed -- this is my pet peeve about compressors that simply put a mix knob on and be done with it. Especially if it's placed pre-make up gain as on the UAD-1 Precision Bus Compressor, IIRC. I mentioned that on the UAD forums, but was met with stony silence. Nice to see someone mention this.

[hakim]

Hi Bob,

I'm still working on the suggestions you gave me a few months back. The advice was priceless, thanks again. But in this thread I noticed you wrote the following:

Originally Posted by bob katz

It seems counter-intuitive to many people that the fastest attack possible (even in microseconds) is a good thing, but for invisible-parallel, that's exactly how it should be used.

So I was wondering, doesn't extremely fast attack times on compressors introduce distortion into audio signal? With that said are there certain types of compressor to use with Transparent Parallel Compression, I'm thinking soft-compressors are preferred for this. Am I right?

Originally Posted by bob katz

The PSP compressor's parallel approach is very confusing to me. I do not like a system that raises the parallel while lowering the dry, in my opinion it's not very ergonomic to use. I would suggest you put in some test tones, and try to psyche it out from there.

I recently got the PSP Master Comp, and now I am hearing this... are you saying this compressor is a poor design?

[bob katz]

Originally Posted by F. Davis

I tend to keep my parallel compression at around -28 to -30 below the un-compressed signal. It is a very subtle effect the way I use it though.

What level do other people here tend to average their parallel compression?

Also ITB what compressor plugins do you all use for the parallel compression (disregarding compressors with the parallel built in such as PSP stuff)?
I'm using the McDSP G-Console comp to good effect.

Also any opinions regarding linear phase processing with this perhaps?

The level that you need is VERY dependent on the model of the compressor, how much it is compressing, attack, release, the musical material and what you are setting out to do, so it's too complex a question to answer by just discussing the level you are mixing at.

In other words, the more gain reduction you are using in the compressor, the more level of the compressor in the mix you can "tolerate" or "desire". The more the compressor is bouncing (coming in and out of compression) the less level of the compressor you are probably going to want or need in the mix for the same sonic effect.

So if I said to you that sometimes I use -18 dB mix level and sometimes I might use -7 it's a meaningless answer unless you know the model of compressor, attack, release, lookahead, type of music... and {snore, sorry}

That said, -28 to -30 level for the compressor level that's mixed in is definitely going to be very subtle and "sweet"!

I hope this helps,

BK

[bob katz]

Originally Posted by Herbeck

I have tried waves LP Multiband and Cakewalk LP-64 Multiband for
Linear phase multiband parallel compression (LPMBPC). =)
But I found the sound of them a bit toyish, not good enough for Mastering.
However I am curious about UAD Precision Multiband, it has a dry/wet knob like PSP MasterComp.
Maybe not the best solution but better than nothing.
I have seen Bob Katz mention some UAD plugs should be good enough for Mastering,
is the Precision Multiband one of them?

Herbeck

I wish I could say anymore, but I never had a mastering computer with enough PCI slots to do more than an "I like this but I can't use it yet" to my UAD, so I gave it to my assistant Robin to use in the mix room. :-(

BK

[bob katz]

Originally Posted by hakim

Hi Bob,


I recently got the PSP Master Comp, and now I am hearing this... are you saying this compressor is a poor design?

I think it is one of the nicest sounding plugin compressors, but I'm not impressed with the ergonomics, especially in the parallel section. I'd love to work with PSP to improve the ergonomics of this plugin.

BK

[Cellotron]

fwiw - amount of times I've been happier with the sound of parallel compression in mastering vs. using other methods of processing to achieve desired results = 0.

Obviously ommv! But I'd say it's very likely that the amount of "classic" releases known for the excellence of their sonics that had parallel compression used during mastering is extremely low.

Best regards,
Steve Berson

[hoop]

FWIW, Vlado Meller mastered an album for me 2 years ago, and he heavily used parallel compression using the Weiss DS-1. The project was a classical/pop album, ala Josh Groban. It was not a subtle difference between the mastered and the unmastered tracks - the loud material with drums didn't feel squashed at all, and the really quiet piano/vocal sections were a bit more forward.

And I absolutely loved what he did....

- Kent

[balcony]

Hi Bob I have just been tracking using "attitude" parallel compression (I would not know the words if not for your book) With a few trial and error takes, I can create a warm "tape?" type of sound ratio around 2 (RNC or sintefex ) or really driven at 10 (fairchild on the sintefex) I tracked 12 part harmonies with the softer ratio's and it all just fits.No plugins!
Less, if any comp needed on the buss.

I believe mixing will be soooo much easier and thanks again for the real deal advice.

[dcollins]

Originally Posted by Cellotron

fwiw - amount of times I've been happier with the sound of parallel compression in mastering vs. using other methods of processing to achieve desired results = 0.

Obviously ommv! But I'd say it's very likely that the amount of "classic" releases known for the excellence of their sonics that had parallel compression used during mastering is extremely low.

According to what I've read on the Internets, it's used on every release.


DC

[XAXAU]

What level of thershold do you use? I think mine´s too low cause when I mix in at -20dBFS it´s usually a little bit too much with a singleband comp. Low ratio, fastest attack, release that fits the track.

-15 to -5dBFS is totally crazy in your face I think.

Just curious if I´m missing something here?

Cheers...

[dcollins]

Originally Posted by XAXAU

What level of thershold do you use? I think mine´s too low cause when I mix in at -20dBFS it´s usually a little bit too much with a singleband comp. Low ratio, fastest attack, release that fits the track.

-15 to -5dBFS is totally crazy in your face I think.

Just curious if I´m missing something here?

Cheers...

Mix it in to where you can just barely hear it, then turn it down another 40dB.


DC

[Virtalahde]

Originally Posted by dcollins

According to what I've read on the Internets, it's used on every release.

I believe you're wrong. It's multiband parallel M/S compression that gets used in every release!

[dcollins]

Originally Posted by Virtalahde

I believe you're wrong. It's multiband parallel M/S compression that gets used in every release!

Thanks for the clarification. Sometimes it's hard to up to date with all the latest technics.


DC

[Virtalahde]

Originally Posted by dcollins

Thanks for the clarification. Sometimes it's hard to up to date with all the latest technics.

I know. I'm sometimes a little lost with the amount of reverb to apply on a typical project.

[mastertone]

Originally Posted by Virtalahde

I know. I'm sometimes a little lost with the amount of reverb to apply on a typical project.

Oh, thats easy!
wet/dry -> 70/30 works great here, but you have to do it before multiband parallel M/S limiting, not compression, please..