De-esser before or after compressor on vox....?

[thephatboi]

I have a very nice compressor set up to do de-essing with a sidechain eq to use in conjuntion with a normal full band compressor on vocal tracking. Wondered if anyone had a preference as to which comes first in the chain and why. Just curious about other's takes on this. thanks.

[lowfreq33]

It can go either way. My preference is usually comp>de-esser, because if you put the comp last it sort of "un-does" what you used the de-esser for. Occasionally I'll go the other way if the de-esser is being too aggressive.

[paterno]

Originally Posted by thephatboi

I have a very nice compressor set up to do de-essing with a sidechain eq to use in conjuntion with a normal full band compressor on vocal tracking. Wondered if anyone had a preference as to which comes first in the chain and why. Just curious about other's takes on this. thanks.

Are you de-essing while tracking?

[thephatboi]

Originally Posted by paterno

Are you de-essing while tracking?

yes, that is my intention because I don't like alot of plugin de-essers and usually end up fixing it manually with quick volume ride edits.

I only need this for certain singers that come in here that may not be the most polished. etc. Thanks for the input.

BTW the compressor I am using for de-essing is a Cranesong STC8 with a sidechain eq, which does work great.

[paterno]

Originally Posted by thephatboi

yes, that is my intention because I don't like alot of plugin de-essers and usually end up fixing it manually with quick volume ride edits.

I only need this for certain singers that come in here that may not be the most polished. etc. Thanks for the input.

BTW the compressor I am using for de-essing is a Cranesong STC8 with a sidechain eq, which does work great.

If you only ever take one piece of advice from anywhere on the internet, this is it:

DO NOT DE-ESS A LEAD VOCAL WHILE TRACKING.

Unless, of course, you are using the de-esser sound as an effect. Because sooner or later, you will de-ess too much while tracking and you will be royally screwed. In my close to 18 years of doing this, it is almost the only thing that is impossible to do anything with later on.

If you must use your sidechain technique, do it while mixing. It should be relatively easy in most DAW's that allow hardware inserts.

You have been duly warned. Just don't do it...

Cheers,
John

[underworld]

i thought de-essing was a form of band-specific compression? so ... depending on the extent of the sibilance .... de-essing before nominal compression would likely be more effective, while de-essing after nominal compression would be more subtle.

[thephatboi]

Thanks John, you are right, I just try sometimes to get the finished sound to tape to save time (on low budget clients) in mixing. It is dangerous I agree. But nevertheless sometimes with time constraints I have to take a chance here and there. Of course those type of clients are the ones who need the most help.... cause they probably can't sing that well c'est la vie.

[paterno]

Originally Posted by underworld

i thought de-essing was a form of band-specific compression? so ... depending on the extent of the sibilance .... de-essing before nominal compression would likely be more effective, while de-essing after nominal compression would be more subtle.

No, not really. The best de-essers are not band-specific. They compress the whole signal -- just super fast. The key is in when it does the compression, the shape of the attack and release, and the amount of gain reduction. If you can find a dBX 902 manual out there anywhere, there is a great section in it on de-esser theory. I've haven't gone out and tried to find more info, but I'm sure there s something out there...

John

[paterno]

Originally Posted by thephatboi

Thanks John, you are right, I just try sometimes to get the finished sound to tape to save time (on low budget clients) in mixing. It is dangerous I agree. But nevertheless sometimes with time constraints I have to take a chance here and there. Of course those type of clients are the ones who need the most help.... cause they probably can't sing that well c'est la vie.

Hey Sean --

I totally understand. I believe any recording is about getting as close to a 'finished' sound every time you hit record. Maybe try a different mic if you are running into big sibilance issues...

Cheers,
John

[TonyBelmont]

Try this signal chain on for size...

compressor > de-esser > compressor

Sometimes it's better to have it in the middle of two compressors compressing in different ways.

[matucha]

If the ESS problem is only related to a volume/loudness of the esses and not distortion, then I use de-esser as a last thing in the chain and as a "fader" for the esses. I wouldn't do it in the tracking, because other processing will mess with that and de-essing is quite a dangerous process, so you can ruin your takes pretty easily.

My reasons why I use it as a last process are:

1) if you EQ (HF) afterwards, your esses will likely change and you'll need to readjust the desser for that
2) if you compress after de-esser, it will(can) bring the esses up again

not that if you change compressor or eq in front of the de-esser you'll leave it set as it was, but there is still big chance it is going to eat the changes nicely.

Try new UAD precision de-esser, it is the best I've tried (in the plugin realm)... it is very fast and if set well quite invisible.

[matucha]

Or me ;-)...

so whatever works for you WORKS ;-)

[thephatboi]

Now now boys
I can respectfully see both points, in a perfect world I would de-ess after recording, but on these particular sessions (which don't allow me enough time to do it like I would like) this singer is not very polished, and time to mix is very limited. BTW my mic is not hyped, I am not even EQing, great pres and converters and still too much sibillance (it's the singer) so I resorted to de-essing while tracking (carefully). I did find after playing around that it does sound more natural with the de-esser before the compressor but that said, the effect is slightly minimized because as said above the full spect. compressor brings the sib. back up a bit, but perhaps that keeps me a bit safer when printing that way. On my own projects and ones that have a decent budget I of course would not de-ess while tracking.

I would like to try that UAD de-esser, heard good things but I am on Protools HD and am out of PCI slots so just to use any UAD stuff I will have to invest a pretty hefty sum to get into an expansion chassis and then the card/plugs etc. I love their plugins, I have the TDM 1176s, LA2As, Pultecs etc but not many of them ever made it to TDM land unfortunately.
Thanks for the input guys.

[druhms]

It's a bit embarrassing, but I have only ever tried putting the de-esser before the comp. Next mix, I'm going to try the opposite. I do admit to using two though, one before and one after the comp.

Not to get too mushy here, but things like this is what I like most about this forum. There are all of those little things you do just out of habbit, then you read about them here and you actually think about them for a moment.

Druhms

[vernier]

Put a de-esser on "Suffering Succotash" ..and what do you get?

[bob katz]

I've had numerous experiences of hearing a highly compressed, very sssssss-heavy recording that could not be de-essed in any way because there were no transients left for the de-esser to grab. It just doesn't work.

So if you patch a compressor in front of a de-esser, and use too much compression, there will be no transient signal for the de-esser to grab, no threshold to set on the de-esser... the hyped up esses just end up remaining, permanently printed in there, distorted and all.

You can get away with compression in front of de-essing as long as the compression is fairly light and you leave enough dynamic range in the s frequencies for the threshold of the de-esser to grab. So, as a general rule, I recommend de-essing in front of compression. And also, it's easier to set the threshold of the de-esser if it comes first as it would be a variable if it follows a compressor which is changing the sound. The de-esser would definitely not work as efficiently if it follows a compressed signal.

De-essers work on thresholds and dynamic differences in the high frequency region. If you reduce the dynamic range of the material coming into the de-esser, it won't work as well!

[matucha]

Sometimes you need 2 de-essers, each tuned to different freq, but that's where things went really bad in the tracking.

Mic does make difference for sure. So does the mic. position. How do you approach the positioning with sibilant voices?
It is obvious to go off axis (with mic that isn't peaky on the sides). I had bad luck putting the mic to the side, because people always turn their heads (instinctively?) towards the mic. I like what happends when you put the mic lower on some male voices (takes more lows from neck'n'chest and tames the sibilance). Putting it higher worked also, but... for some reason I don't do it this way often. Putting mic further away also helps often... if room (sound) allows.

De-essers work on thresholds and dynamic differences in the high frequency region. If you reduce the dynamic range of the material coming into the de-esser, it won't work as well!

Good points. But strangely when I put de-esser before the comp, it brings esses up again (most of the time). On other hand whenever I can I just cut the problematic esses and lower the volume on them manually. Works best and it is definitely before compression ;-).

When I think about the process... most of the time the recorded voice doesn't sound sibilant, but after the compression it becomes a bit so that's why I tend to use de-esser at the end of chain. If the compressor had a sidechain... ;-)... it could be done in one process. Nice, shame that UAD comps don't have one ;(.

[Benmrx]

Call me crazy, but I usually put a de-esser on a vocal at the point in which it's starting to bother me. So.....If I'm doing a mix, and I bring up the vocal, and there's some nasty shit going on with sibilants, I'll bust out a de-esser, therefore being the first thing in the chain.

If I'm working my way through a mix, and I've got the vocals sounding great with no de-esser, but then I find that by the time I've done all my vocal processing that the sibilants in the chorus' are too strong, then I'll strap on a de-esser, automate the bypass so it's only on in the chorus', and will be the last thing in the chain.

Then sometimes (like the last song I mixed), there was just this one "ssss" that was too strong in the intro of the song, so I just did a quick volume dip.

[macleodgrant]

i'm thinking melodyne may be the best for de-essing control (select the esses and decrease the volume) and i think you should always process the best possible quality source so in that way a vocal with minimised esses and if you need it then at the end of the chain should suffice

i've started using 2 de-essers to control them automatically, one at a lower freq and one at a higher freq and just selectively. a multiband compressor would probably be the ticket here

[matucha]

multiband = 2 many crossovers

wideband de-esser = no crossovers, it just lowers the volume on esses


esses are always separate, so there is no need for crossing it unless there is a specific problem