Compressor trick...

[BrianK]

I originally posted this under a De-Essing thread, but figured it might not get seen by those who might be able to use it...

De-Essing is only one use of a frequency-selective compression. For DeEssing, they usually take a section around 8kHz and boost it into the "sensing" side of the compressor, making the "Esses" duck down faster. Unfortunately, sometimes the compressor inside the unit itself is not too nice - Or you want to take out some "boom" or midrange harshness. Here's my solution - it works wonders on nearly anything:

- Get a GOOD compressor with a sidechain input (Most have it nowadays). Buy a cheapo graphic EQ with 31 bands or so. (The sound quality of the EQ is not important, as you NEVER hear the sound through the EQ - it just affects the compressor sensing.)

- Split the signal and feed it into the EQ sidechain, and the other half into the compressor. Now - when you BOOST anything on the EQ - it cuts that range (first) on the compressor. If you [/b]CUT anything at the graphic EQ, it allows it to come through relatively uncompressed.

This is nice for many reasons:

(1) You get a good quality compression.

(2) As it IS a compressor (not just a De-Esser) you can set the overall compression needed for your track and THEN add the little De-Ess (or DeBoom, or whatever). One box, so less signal degradation than using a compressor plus a DeEsser.

(3) It can be used as an overall compressor (as above, but with the opposite effect of a DeEsser - when you CUT signal on the graphic, the vocal can be compressed normally, EXCEPT in certain ranges - allowing low quiet vocals through, or maybe high wispy frequencies to sneak through unaffected.

(4) It's FAR more controllable than any DeEsser - various bands and various amounts can be affected at one time. You can tame practically ANY track. It is amazing.

[Beech]

Thanks Brian, must remember to try that soon!

[Bmnd]

Hi!

You can achieve better results using two good separate boxes. De-esser acts differently, than compressor with side-chain.

[vaesion]

ive never done that before.........
so if i have it side chained with an eq and the vocal needs to be dessed.........what do I do? boost or cut the freq range that needs compression?

[AdAudioInc]

This works well on most digital consoles as well without leaving the board. On my O2R, I can copy an input to another channel so that I have 2 channels of the same vocal. On the second channel cut everything below 500 / boost the S - anywhere from 4k - 8k / and then remove it from the stereo bus and send it to an aux. Then use that aux to key the limiter on the main vocal channel. De - S

[Jack Ruston]

Brian I'm confused...

You seem to be implying that a sidechain eq setup is capable of compressing some frequencies in a signal more than others....

It doesn't really work like that because the compressor's threshold is level dependant not frequency dependant, and as you say, the signal passed to the compressor is unaffected by the eq...So all it really lets you do is to compress more when certain frequencies are present as the actual compression that occurs, affects every frequency in the source above a certain level.

It's a useful technique but it's not always effective. It goes hand in hand with its cohort, multiband compression, which does allow you to subject different frequencies to different processes.

Sorry if I've misunderstood your post!

J

[AdAudioInc]

Sly, if you boost the S like 15dB with a narrow Q, then you can set the threshold so that the only thing exceeding it IS the S. Therefore when the limiter (very fast attack and release) is triggered, only the S gets compressed.

[Babe Records]

Yes, but Sly is right: even if the limiter (or the comp) is triggered by "S" only, the compression will affect the whole spectrum of the signal, so that if youhave anything else than the "S" in your signal (for example a bass), it will be compressed along with the "S", which can be a problem in some situations.

[Babe Records]

Still, it's a good trick that works in most cases

[AdAudioInc]

In most cases that I find this tech usable - it wold be on the vocal track only, so no other instument would be affected. But I don't disagree totally - yes this tech can have some unfortunate side effects if not used carefully.

Experiment - have fun

[Jack Ruston]

De-essing a vocal track is not a great example because when you get an "ESS" it's the only thing happening at that moment, so ducking the track and ducking the "ESS" are one and the same thing: This technique is a short cut for going through the track and manually automating level. (Where things become a little more interesting would be when de-essing a vocal within a mix rather than in isolation...)


The essence of the thread is that its possible to use the sidechain to control other areas of the frequency spectrum too...but I'm contesting this, in that I dont think you can very easily tame boominess for example by chucking a lot of low end into a sidechain. If eq failed I'd be looking at a multiband.

[Data Entry]

Originally posted by Babe Records
Yes, but Sly is right: even if the limiter (or the comp) is triggered by "S" only, the compression will affect the whole spectrum of the signal, so that if youhave anything else than the "S" in your signal (for example a bass), it will be compressed along with the "S", which can be a problem in some situations.

Just wanted to add that if you remove ALL Frequencies but the "Sessss" then it will work.

Meaning if you filter out ALL frequencies but the "S" through the sidechain, it will work as a de-esser.

If you just boost the "S" signal only, then the compressor will reacte to the full program as well as the "S".

So just use the EQ to remove all the frequencies but the "Sesss" and thats all the compressor will see, just a narrow band of harsh frequencies.

[Jack Ruston]

No....

That's all the THRESHOLD will see...and it really doesn't matter too much if there's other stuff there as well, provided the offending frequency is significantly prominent enough to get a trigger for the compressor....BUT:

That's irrelevant, because the compressor still ACTS on the signal that is fed to it's input regardless of frequency content.

J

[Kenny Gioia]

Originally posted by BrianK

(1) You get a good quality compression.

A quality compressor is not really what you want from a de-esser.

You need a really quick attack and a really quick release.

The warmth from a vintage compressor is really not applicable.

And a real de-esser should employ some kind of crossover to avoid compressing everything.

That's not to say that your idea has no value.

It's a good idea for a vocal with slightly too much sss coming thru.

In fact you can use it for any offending freq.

Thanks

[AdAudioInc]

In my example, I use an eq'ed version of the vocal for the key - cut everything below 500 - and sharply boost the offending frequency - with a true zero attack (digital limiter, and maybe 5 or 10 ms release on a single vocal only track - I don't think that can affect anything else other than the S ??

[Jack Ruston]

OK but that's because there's nothing else other than the "ESS" to affect...In a vocal track there's nothing else happening at that second that the "ESS" occurs...That's why I'm saying that essing is a bad example...

Lets take boominess....It occurs with a lot of other stuff at the same time, and it's not just a split second kind of thing that can be ducked without anyone noticing. So therefore it's not going to work very well if you try to remove it with this sort of technique. What will happen is that if a section is a bit boomy, it will remain boomy but just be turned down a lot. The vocal will jump up and down while retaining its broader tonal charachteristics.

J

[AdAudioInc]

agreed - the only other thing I use this type of thing for is to key a compressor on a music track with the VO track - a ducker I think it's popularly called.

Boominess - I don't know??

[BrianK]

I did not say it compresses JUST one frequency range - I said it does it "faster" which means "is more sensitive to those ranges" if boosted.

A "good quality" compressor - Distressor is one fine example. There are many others.

I do think DeEssers have serious limitations over this technique, and I usually find their range of choices very limited. (And yes, a DeEsser IS a frequency selective limiter). As I said - this is not just a DeEsser, but the mechanics of DeEssing make this a very useful setup.

Until you've used it as much as I have (I hook it up about one in every three mix sessions to tame something) I wouldn't say "it doesn't work" - yes it compresses the whole signal, the threshold is ALWAYS affecting the whole sound. I usually use only a few dBs of cut or boost on the EQ control chain. But it DOES make your vocal/guitar/drum loop respond in a totally controllable way that is unavailable with a "normal" compressor or DeEsser.

To clarify and simplify: setting up this way, when you "boost" a frequency on the EQ, it cuts it more readily on the compressor. When you "cut" on the EQ, it allows that range to "come out" more, less compression. Sometimes, this allows you to get by without corrective EQ to the signal, which saves on phase shift, etc., you know....

[BrianK]

Boominess - I know what you're saying. But when someone is in CLOSE on a mic and makes it "boom" that IS the main signal content at that point, and it dwarfs the rest of the sound anyway. With this setup, it DOES turn it all down, but only for the fragment of a second when it booms, and the rest of the "tone" is still there, maybe a few dBs lower.

This is NOT to correct overall boominess that continues throughout a track - that's an EQ thing - anything CONSTANT is not fixed by a dynamic action.

[AdAudioInc]

For pops - I sometimes define the pop in the DAW and apply a low cut filter to it, usually its just a few frame event. This can filter out the plosive without affecting what's around it.

Not exactly on topic, but a sort of trick anyhoo

[Jack Ruston]

Hey Brian...

Sorry if I misunderstood your original post...It just seemed very much as if you were suggesting that the rest of the signal could be left unaffected...My apologies for any misinterpretation.

I do use the technique quite often and I take your point (and mentioned this above somewhere) that the technique does work when the offending frequency appears momentarily....

But as the problem crops up more and more, I like a band of compression set to limit it. Have you tried the tubetech hardware multiband? I haven't had the chance yet and I dont know if it lets you 'get in there' enough, but I've had some good results with the waves linear multiband (makes PT delay compensation work overtime) and the McDSP one (although I really dont like the gui etc)....Is there a multiband that has a 'normal' compressor gui on each band?

J

[BrianK]

As a multiband compressor (by its very nature) is FOUR or more compressors (hardware version), so the quality is rarely as good as the single. Compromise either way. And this way STILL has more options for parameter control.

All of it works, all of it works "differently"... This one is - by far - the easiest and cheapest to try. (As mentioned above - before I had the graphic - I used a channel of the board EQ'd into the compressor.)

Try it on an acoustic guitar WITHOUT problems; duck down a bit of the boominess and bring out a bit of the stringiness - without EQing the audible track. It works.

[lukasbrooklyn]

sly is right. i find dynamic eq to work best in this area - freq. selective dynamic treatment & no band splitting and subsequent joining.