Mix/Production : De-essing

@budhabuilding ...thanks for that. Interesting chain in that plug. I bet you can change the order of things. I have always compressed before limiting in my mastering chain and I thought that perhaps a mastering de-esser might want to pull those highs down before they hit the compressor. My computer is not a hot rod so I stay away from third party plugs....do everything with stock Digi plugs.

@tuff bransch ....U89...have I heard your vocals through that...maybe not. I'd like to. My two microphones are Peluso 22-251 and Shure KSM32. The vocals all go through the Peluso and it is bright in a way that I find pleasing for my rough voice. Back when I tracked them with the KSM32 I did not do much, if any, de-essing.

@Leonard Scaper The Lurssen plugin chain can't be changed ...it's has just a few options and works like charm....If i send some preview mixes for clients, i use Lurssen.

Most de-essing is a kind of compressing just on the high end.. So it's not really a dynamic EQ. They split the frequentie in 2 parts. The low end is  will not be changed and the high end will be compressed in a ratio between 2:1 and 8:1. Attack and release times are auto.
So if you want full control the best way is to use a multiband compressor that uses only the high or the mid high band.

If you want  a smooth and free multiband dynamic eq then you could use the NOVA. http://www.tokyodawn.net/tdr-nova/
It's a great alternative for the high end fabfilter MB.

To set the attack and release times, i look at sound in this way:
Every sound exist of tone and articulation or body and presence.
Presence is in the high end and body in the low end.
If you want to hear a frequentie of 110hz you need at least 1000 ms / 220hz = 9.09 ms to hear just one cycle of this low A.
But you need to hear a couple of those cycles. In the high end it goes much faster.
if you compress low ends the attack times are much slower then the high end.
If you have problems to recognize the tone of your low end (body) with the compression change the attack time and the knee.

Just my personal opinion.

To set the attack and release times, i look at sound in this way:
Every sound exist of tone and articulation or body and presence.
Presence is in the high end and body in the low end.

@budhabuilding ....there is some real insight in your discussion of attack times, particularly that little gem about body and presence.

I have always known that low frequency sound waves are much longer than high frequency waves so it makes great sense to keep low frequency attack times slower for better low end body. That realization has finally, after all these years, led me to a better understanding of the benefit of multi-band compression because you can tailor the attack/release settings to the different frequencies. Thanks for that insight.

And you tossed that in about the knee parameter as well.  My mixes have taken on a different feel since I began using a softer knee. I'll look even more closely at that now for instruments with more low frequency information. I would suspect that a de-esser will have a fixed hard knee.

I wonder if just using a multi-band comp for de-essing would allow fast enough attack and release. I seem to recall back when I was experimenting with my Waves multi-band that the attack times were not nearly as fast as something like an 1176.

ha @Leonard Scaper @budhabuilding @Mar T. @Cool Hand @tuff bransch

fascinating read, i can only follow half of it......but.... that is a lot more then i would have a few years ago 

and i am picking up bits and pieces of this conversation....thanks for sharing 

That's certainly a fantastic collection of ideas fellows.. I'll add this thread to our 'knowledge base' (stickied!)
Thanks all so far for contributing.
Feel free to continue the debate though!

De-essing at my place is mostly done by eye and hand - that means listening to and watching the isolated vocal track and as soon as an obstrusive -ess or sharp t-sound is heard I zoom-in and lower the volume, or compress it or EQ it.
- for Cubase 8 users.

Although I sometimes use desser, personally it seems to me that it degrades notably the samples and destroys the brightness and body of vocals. But, if you add it, you should forget how it sounded before and stay satisfied, because if you bypass it again to check a before and after, doubts will make your brain explode.
I know that our songs sometimes squeak by the eses, something must be done, but personally I also prefer to zoom in the problematic parts and automate the volume by hand.
anyway, my favorite desser is FabFilter Pro-DS, very selective and efficient.
Thanks for the tips!

Although I sometimes use desser, personally it seems to me that it degrades notably the samples and destroys the brightness and body of vocals.

I'm glad you bumped this topic up @BLACK MIRROR because I really think de-essing can make a good recording sound even better when done right. I think folks need to just become aware of this issue and then take the time to get comfortable with a system that works for them.

I definitely agree with you about second guessing.

Hi, very interesting discussion.

My swiss army knife for deessing is a tool by Toneboosters (called Silbalance). Its working on our material quite well - I hope. And it is cheap (20 or 30 Euros only)

Best,

Fabian

I just took a look at that Toneboosters Sibilance de-esser and it looks nice and simple........just a very fast frequency based dynamics processor.

I have problems with de-essing - usually found after I have mixed and playing back later on a tinny speaker. I got a Rhode mic as it had great reviews for vocals - trouble is it picks up all the sssss and voice clicks. I heard it was helped to reduce sibilance by putting the centre gold spot to an angle when recording - not sure if it works but I try it. Often I get a better result just using my SM58 mic.

As for de-essing the track, this is my method. Copy the vocal track into a fresh track and name it as "de-esser". Using narrow band of EQ find out what frequency is the hot spot for sssss (usually around 7kHz). Create a notch (say 6.5 to 7.5) using a bell curve with Q at its highest setting. Sweep around to see exactly where sss are worst and tighten up the filter round it, only letting the sss through. Put output of de-esser track to zero. On vocal track insert a compresser and side chain to the de-esser track. Every time ssss comes through the compressor will react. Then tweak the compressor till you get a subtle de-essing effect.

Once you have got this to work for you, you can save presets in both your EQ and compressor plug-ins to use when you are mixing other songs.