This forum is sponsored by

**T-RackS Group Buy - EXTENDED!**

Up to 10 for the price of 1 possible.

Mixing and mastering processors, including Lurssen Mastering Console.

Read more at**www.ikmultimedia.com**

- KVRist
- 362 posts since 2 Jul, 2010

I had a good experience recently using frequency-dependant expansion to put some life back into the heavily saturated vocal in a full mix. ("Re-essing" but with a bell around 2.5k). Waves C1 did the job in split-band mode and was quick to dial in. I kept the version done with TDR Nova GE instead, which is super-clean and designed for full-mix-processing. Either would have been fine in this case.

Obviously the usual caveats about multiband processing and fixing stuff in the mix apply!

Obviously the usual caveats about multiband processing and fixing stuff in the mix apply!

- KVRAF
- 11097 posts since 13 Oct, 2009

Tearing Riots wrote:ghettosynth wrote:This is generally referred to as "upward compression,"

I know the differences, reaxcomp does upward expansion, not sure what you are getting at.

Good post though.

I was referring to this:

Tearing Riots wrote:Most 'dynamics' devices can be used as upward expanders, your daw might have one. Ratio < 1 usually indicates this.

Given the limited number of examples here, I don't think that there is any real consensus, certainly not enough to say "usually." For Neutron, it's not true, negative ratios mean upward compression. Granted, I was working under an unsupported assumption as well, but, Mogawi boy has it right, few support this. I'd actually have to go look to see which of my 40+ compressor plugins support this, but I'm happy enough with the FLux tools that "I" don't care that much.

I think that the only true thing that can be said is that you need to check your manual on what negative ratios mean.

- KVRAF
- 11097 posts since 13 Oct, 2009

Tearing Riots wrote:Ratio between 0 and 1 is upward expansion, no? I certainly don't know everything that's out there, have seen negative ratios only on elysia mpressor where it 'overcompresses', that's the way it makes sense to me.

Yes, assuming all else equal, but that's not what you said. All negative numbers are less than 1. Since I was not familiar with how ReaX implements it I only had Neutron as a reference which doesn't support ratios in the range (-1,1).

- KVRAF
- 11097 posts since 13 Oct, 2009

Etienne1973 wrote:Tearing Riots wrote:... have seen negative ratios only on elysia mpressor where it 'overcompresses', that's the way it makes sense to me.

Makes sense to me too. A downward compressor that provides negative ratios is STILL downward compressing.

Yes, again, all else being equal. That's not what happens in neutron however. The sign is used to change the behavior to upward compression.

From the manual:

THRESHOLD

Threshold sets the point at which the dynamics processing begins to take place. With positive ratios, this means signals overshooting the threshold, and with negative ratios, this means signals falling below the threshold.

RATIO

Ratio allows you to adjust the amount of level adjustment the compressor will apply from any given input signal. A ratio of 3:1 means the for every 3 dB a signal overshoots the Threshold, only 1 dB of gain increase will occur. Neutron’s Compressor is capable of both upwards compression (with a negative ratio) and downwards compression (with a positive ratio).

So now we have three separate examples with three different behaviors for "ratios less than 1."

- KVRAF
- 11097 posts since 13 Oct, 2009

Tearing Riots wrote:Yes i was not explicit enough.

And the neutron nomenclature is wrong

Sure, it's not representing a "negative ratio." However, IMHO, it's far more useful than the "correct" case that the mpressor uses. Waves C1 also uses negative ratios for the same feature as mpressor, for those that care. Waves points out though that it's of limited interest. With iZotope's choice, it's almost certainly just a U/I thing. It allows you to state that negative ratios do something different without having a separate control and without having to explain that the "upward compression" ratios are below the "downward compression" ratios in the list.

A better solution might be to not have the parens at all and just use an upward or downward facing arrow, or include a "range" control like ProR. Again though, my guess is that the use of the negative ratio was the simplest solution that iZotope felt would confuse the fewest customers without cluttering the interface further.

- KVRist
- 276 posts since 10 Sep, 2015, from You haven't unlocked this character yet

a:b compression

a=amplitude above threshold going in and b=amplitude going out

a:b can be expressed as (b/a)=amplitude attenuation...

So, if a:b=

2:1 original signal above threshold is attenuated to 1/2

3:1 original signal above threshold is attenuated to 1/3...and so on

(0.2):1 original signal above threshold is changed (increased) to 5/1 or just 5 - this is upward compression

(thanks to Tearing Riots for helping make this section a more precise explanation)

Why? Because 0.2 = 2/10 or 1/5, then 1/5:1 as we now know is 1/(1/5) or just 5.

Concept exploit:

Which is more, 1/2 or 1/100? RIght, 1/2 or 0.5

i.e., would you want half a dollar or one hundredth of a dollar?

Which is more, 1/1 or 1/(1/5)? 1/(1/5) is

Because 1/(1/5) = 1/1 * 5/1 = 5 if you remember how to divide fractions by flipping the second fraction and then multiplying.

The use of negative signs in context of compression is best understood by the use of negative signs in kinematic physics. Negatives used on vectors for motion mean that something is going the other direction of what forwardis defined to be, based on the reference frame. In other words, putting your car in reverse and punching it, can be mathematically modeled by -v, negative velocity.

So, a value smaller than 1 for (a), or (-) sign, in this case, mean essentially the same thing. Why do some developers use one over the other if they mean the same thing? It's whatever they think you'd find to be more intuitive to understand. Personally, I like the use of negative signs because it intuitively implies direction but the decimal is actually mathematically correct.

Physics people have a more intuitive approach to explaining things while the Mathematicians have a more formal way. So, in this case, I can almost bet the negative sign was the idea of the physics guy and the decimal idea was more some "mathy" person.

Now, if you understand this, applying BOTH concepts, you should be able to take this little test:

What does -0.2:1 mean?

Anyone? Anyone? Bueller?

a=amplitude above threshold going in and b=amplitude going out

a:b can be expressed as (b/a)=amplitude attenuation...

So, if a:b=

2:1 original signal above threshold is attenuated to 1/2

3:1 original signal above threshold is attenuated to 1/3...and so on

(0.2):1 original signal above threshold is changed (increased) to 5/1 or just 5 - this is upward compression

(thanks to Tearing Riots for helping make this section a more precise explanation)

Why? Because 0.2 = 2/10 or 1/5, then 1/5:1 as we now know is 1/(1/5) or just 5.

Concept exploit:

Which is more, 1/2 or 1/100? RIght, 1/2 or 0.5

i.e., would you want half a dollar or one hundredth of a dollar?

Which is more, 1/1 or 1/(1/5)? 1/(1/5) is

Because 1/(1/5) = 1/1 * 5/1 = 5 if you remember how to divide fractions by flipping the second fraction and then multiplying.

The use of negative signs in context of compression is best understood by the use of negative signs in kinematic physics. Negatives used on vectors for motion mean that something is going the other direction of what forwardis defined to be, based on the reference frame. In other words, putting your car in reverse and punching it, can be mathematically modeled by -v, negative velocity.

So, a value smaller than 1 for (a), or (-) sign, in this case, mean essentially the same thing. Why do some developers use one over the other if they mean the same thing? It's whatever they think you'd find to be more intuitive to understand. Personally, I like the use of negative signs because it intuitively implies direction but the decimal is actually mathematically correct.

Physics people have a more intuitive approach to explaining things while the Mathematicians have a more formal way. So, in this case, I can almost bet the negative sign was the idea of the physics guy and the decimal idea was more some "mathy" person.

Now, if you understand this, applying BOTH concepts, you should be able to take this little test:

What does -0.2:1 mean?

Anyone? Anyone? Bueller?

Last edited by Mathematics on Fri Jul 13, 2018 7:14 pm, edited 5 times in total.

...and the electron responded, "what wall?"

- KVRist
- 172 posts since 10 May, 2006, from Ireland

- KVRian
- 758 posts since 18 Oct, 2014

Mathematics wrote:3:1 original signal above threshold is attenuated by 1/3

To be precise, the output signal above threshold (a.t.) here gets attenuated to 1/3 of the input signal a.t.

And if we substitute 'attenuate' with 'changed' your next line makes sense to me

(0.2):1 original signal above threshold is changed to 5 times the input a.t. - this is upward compression

That's upward expansion though, unless i missed a bigger point.

The use of negative signs in context of compression is best understood by the use of negative signs in kinematic physics. Negatives used on vectors for motion mean that something is going the other direction of what forwardis defined to be, based on the reference frame. In other words, putting your car in reverse and punching it, can be mathematically modeled by -v, negative velocity.

You could also see it this way: any input level above threshold results in an output level above threshold - unless you reverse the thing with a negative ratio. Just a different reference frame.

Why do some developers use one over the other if they mean the same thing? It's whatever they think you'd find to be more intuitive to understand.

Agreed, that's also why i didn't even mention negative ratios in my first post, limited usefulness. Still a bit arbitrary I'd argue.

Now, if you understand this, applying BOTH concepts, you should be able to take this little test:

What does -0.2:1 mean?

With my preferred definition -1:5 means go in 1dB above threshold, get a signal 5dB below threshold back. Izotope would increase the volume above threshold 5 times.

- KVRAF
- 2006 posts since 28 Dec, 2015, from Germany

U-HE Presswerk

Music was my first love...

Flowing atmospheric music

http://www.sonoryth.bandcamp.com/music

- KVRAF
- 11097 posts since 13 Oct, 2009

Tearing Riots wrote:Mathematics wrote:3:1 original signal above threshold is attenuated by 1/3

To be precise, the output signal above threshold (a.t.) here gets attenuated to 1/3 of the input signal a.t.

And if we substitute 'attenuate' with 'changed' your next line makes sense to me(0.2):1 original signal above threshold is changed to 5 times the input a.t. - this is upward compression

That's upward expansion though, unless i missed a bigger point.

Yes, unless I'm missing something also, this in no way changes how the threshold will be interpreted which is necessary to switch from downward compression to upward compression.

Downward compression and upward expansion work on signals ABOVE the threshold, whereas upward compression and downward expansion work on signals BELOW the threshold.

The use of negative signs in context of compression is best understood by the use of negative signs in kinematic physics. Negatives used on vectors for motion mean that something is going the other direction of what forwardis defined to be, based on the reference frame. In other words, putting your car in reverse and punching it, can be mathematically modeled by -v, negative velocity.

You could also see it this way: any input level above threshold results in an output level above threshold - unless you reverse the thing with a negative ratio. Just a different reference frame.

I think trying to understand negative compression ratios is of limited interest in practice. In iZotope's case they are NOT, in fact, negative ratios at all. The sign is simply an indicator to show a switch of mode. I don't think that there's any disagreement that it's "technically" wrong.

For me, iZotope's approach is useful as it switches between two types of compression and allows me to quickly compare which approach that I think will work better to control the dynamics for a particular part in context. I don't generally ask myself whether a compressor or an expander is the right choice.

Last edited by ghettosynth on Fri Jul 13, 2018 8:54 pm, edited 2 times in total.

- KVRian
- 623 posts since 14 Dec, 2014

Tearing Riots wrote:Most 'dynamics' devices can be used as upward expanders, your daw might have one. Ratio < 1 usually indicates this.

Else for free there's ReaXComp (multiband with upward exp.), MCompressor (freely drawable transfer curve so can be up/down expander/compressor), probably lots of others as well.

Yup, Live has Multiband Dynamics.

It can do the upper and lower threshold in 3 bands at the same time, with an easy to use UI.

Its manual has a nice explanation that may help non-Live users too:

https://www.ableton.com/en/manual/live- ... ing-theory