So i recently finished recording and mixing a song that my friend and I wrote. Near the very end of mixing I started to notice a lot of the Ess sounds really popping out(could be due to compression and nebula console I added to the vocals). I attempted to use 2 commercial DeEssers, but that was hugely disappointing. I eventually resorted to volume automation which worked fairly well.
How do you guys deal with sibilance ?
So i recently finished recording and mixing a song that my friend and I wrote. Near the very end of mixing I started to notice a lot of the Ess sounds really popping out(could be due to compression and nebula console I added to the vocals). I attempted to use 2 commercial DeEssers, but that was hugely disappointing. I eventually resorted to volume automation which worked fairly well.
How do you guys deal with sibilance ?
One of the easiest and most-overlooked techniques is to simply delete or gain down the esses. If you listen to top 40 radio carefully for an hour you will plenty of examples.
One clumsily-executed example is Lady Gaga's "Bad Romance"... if you pay attention, you will hear a ton of esses that are clumsily de-essed, including on the chorus, where the vocal basically disappears at the "ess" sound, but frequently "releasing" too early with a sharp, ugly, pronounced "sss" on the up-beat. Other times, the "ess" is just gained down entirely, so that the vocal effectively sings "caught in a bad roman...", especially in the last third of the song.
If you listen to the song closely, it's pretty obvious. Mostly nobody would ever notice it unless they were paying attention. But if you listen for the esses, you will hear pretty clearly how they are handled differently, and especially how there are occasionally sharp esses on the release, and how in other places, the vocal disappears into "caught in a bad roman..." without completing the word: that's not the singer, it's the engineer fading down the sss.
The technique is basically to simply fade down the vocal rapidly anywhere with an offending "ess"... the listener's ear is remarkably good at filling in the missing consonant, and you usually don't even notice it unless you're listening for it.
Esses are usually not as bad you think they are. The recording engineer, focused on each track and capture, tends to hear everything that is wrong, like the makeup technician who notices nothing but the wrinkles and uneven skin-tones on the supermodel.
If you can get the esses down to half the source level, most people will never notice it.
Depending how loud the sibilance is you could use reacomp's lowpass and highpass filter section to target the esses. If they're quieter than other material in the vocals that you don't want to attenuate then you could side-chain an eq that exaggerates the esses to trigger reacomp.
Another route to go is to use automation on an eq to target the sibilant sounds, but this would involve pretty much the same amount of work as manually drawing volume automation. The only advantage would be that it only affects the volume of the sibilant sound's frequency.
Esses are usually not as bad you think they are. The recording engineer, focused on each track and capture, tends to hear everything that is wrong, like the makeup technician who notices nothing but the wrinkles and uneven skin-tones on the supermodel.
If you can get the esses down to half the source level, most people will never notice it.
I'm glad i read ahead before I post. This is almost exactly what i was about to post. Engineers need to listen critically, but we also need to try to listen with a little bit more forgiving ears to get a better perspective on how music sounds as a whole, and not over-edit because we are listening a bit too critically. I have intentionally left esses that I was advised to remove, and the end product was much better as a whole.
__________________
The Sounds of the Hear and Now.
the best for me is... manual editing... non destructively.. using the Volume (Pre-FX) envelope. Why? Because not all esssss sounds are the same.. in loudness.. and I find many de esser plugins to also affect to some T's and other sounds... the only way for me is manual edit all the Ssss and it is the best control .. more work.. but better results.
Activate the Volume Pre-FX envelope. Select the time where the Sssss is (it's often clearly visible in the waveform) with the time selection drag down the volume fader in the envelope track. This will lower the volume in that time selection.. you can always tweak the points later in case you need it. It's non destructive and easy to do it this way. No item splitting necessary.
Activate the Volume Pre-FX envelope. Select the time where the Sssss is (it's often clearly visible in the waveform) with the time selection drag down the volume fader in the envelope track.
I have never used this approach to work with envelopes before. How could I have missed this? Thanks heda.
Thanks for all the replies guys, it is truly appreciated. In response to Yep's comment, I think you're absolutely right about the engineer stressing over all these little details that the average listener would never notice.
I may also try the ReaFir plugin in the future too.
the best for me is... manual editing... non destructively.. using the Volume (Pre-FX) envelope. Why? Because not all esssss sounds are the same.. in loudness.. and I find many de esser plugins to also affect to some T's and other sounds... the only way for me is manual edit all the Ssss and it is the best control .. more work.. but better results.
Activate the Volume Pre-FX envelope. Select the time where the Sssss is (it's often clearly visible in the waveform) with the time selection drag down the volume fader in the envelope track. This will lower the volume in that time selection.. you can always tweak the points later in case you need it. It's non destructive and easy to do it this way. No item splitting necessary.
I'd rather do this with per-take volume envelopes because they can be managed easily. Track envelopes can be a bit clumsy sometimes. It's dead simple to work with them and they're right there on takes/items which makes me work faster because it's like editing the waveform with a pencil tool (something I've been missing on REAPER for many years).
@CA$H FLOW "Esses" are very simple to track down, they always look like this (just don't overdo it like yep says, hear every edit like a fan not an engineer):
Even the most incredible de-esser I've heard in software, Fabfilter's Pro-DS, doesn't work better than manually editing them.
__________________ Pressure is what turns coal into diamonds - Michael a.k.a. Runaway
a big +1 to envelope automation. Like Mercado I also use take envelopes for this. Just like with pre-fx envelope, you can visually see the change in waveform which is cool bonus.
Of course the best thing is to get it right with source, you know mic placement, right mic and singing technique etc. But all in all take automation works very well, even though it takes little time.
BTW I sometimes use SPITFISH de-esser too. The trick is to get signal loud enough before de-essing plugin so I use some gain plugin there and another one after de-esser to set the gain back to where it was.
Location: Near Cambridge UK and Near Questembert, France
Posts: 22,754
And of course the simplistic way of dealing with it at source.
If you have a singer up who you can hear as likely to be problematic in this respect, oldest trick in the book is to set up a wooden pencil right in front of the diaphragm (mic, not singer!) which will reduce both sibilance and plosive effects without killing the actual signal you want.
Many good suggestions above. Of course if you have the time, Mercado Negro’s technique is one of the most efficient/transparent ways to do it and I often use that technique for the final mixing/mastering stage. However, if you want the de-essing to be done automatically and with the least amount of side effect, there are a couple of de-esser plugins that are relatively quite efficient/transparent and they are not necessarily the most expensive ones, they were simply designed better.
Of all the de-essing plugins I've tested (~ 20 different ones); here are my favorite ones in order of transparency:
Freeware:
- 2 to 4 instances in series of the Limiter No6, using the "HF limiter" section only at HF threshold of -10 (the MinLatency option has to be on for real-time processing)
- Spitfish De-esser
Commercial ($) :
- Tone Booster De-esser
- Oxford Supresser
Note that in general, precise adjustment of de-esser parameters is CRITICAL to both its efficiency in reducing sibilance, as well as for its transparency. Compared to other “standard” processors (gates, compressors, eqs, etc) de-essors are really hard to setup if transparency is the main factor. To end on a positive note; of all the de-esser I’ve tested to date, Limiter No6(x4) is the easiest to setup while producing the least side-effect, and it’s also free.
And of course the simplistic way of dealing with it at source.
If you have a singer up who you can hear as likely to be problematic in this respect, oldest trick in the book is to set up a wooden pencil right in front of the diaphragm (mic, not singer!) which will reduce both sibilance and plosive effects without killing the actual signal you want.
Thanks for the suggestion, I'll try that next recording session. Most likely it's because the pencil is just big enough to produce an interference pattern that disperse the very high frequencies, thus reducing sibilance. Not very nice looking though!
The best method is to back away from the mic, which reduces sibilance. Being right up on the vocal mic almost always results in more sibilance. On the "fix" side, manually editing sounds better than any de-esser but takes time. The pencil trick GPunk mentioned is an old tried and true method.
But being right up the mic increases sibilance, as does heavy vocal compression.
Even the most incredible de-esser I've heard in software, Fabfilter's Pro-DS, doesn't work better than manually editing them.
Agree fully. It's often a good bit more work though.
Melodyne does that really well, just go in and "draw down" the gain on sibilant blobs, without splitting or tuning or crossfading or otherwise affecting the vocals.
If you have a singer up who you can hear as likely to be problematic in this respect, oldest trick in the book is to set up a wooden pencil right in front of the diaphragm (mic, not singer!) which will reduce both sibilance and plosive effects without killing the actual signal you want.
Quote:
Originally Posted by Lawrence
The pencil trick GPunk mentioned is an old tried and true method.
Lol. Credit where credit due.
__________________
The Sounds of the Hear and Now.
One of the easiest and most-overlooked techniques is to simply delete or gain down the esses. If you listen to top 40 radio carefully for an hour you will plenty of examples.
The technique is basically to simply fade down the vocal rapidly anywhere with an offending "ess"... the listener's ear is remarkably good at filling in the missing consonant, and you usually don't even notice it unless you're listening for it.
Esses are usually not as bad you think they are. The recording engineer, focused on each track and capture, tends to hear everything that is wrong, like the makeup technician who notices nothing but the wrinkles and uneven skin-tones on the supermodel.
If you can get the esses down to half the source level, most people will never notice it.
I learned this manual de-essing process from Kenny Gioia's mixing rock tutorial vid. Seems like the pros prefer this method. I had struggled with automated de-essers for a long time. Bottom line is that they are hit or miss and more miss than hit IMO. But it also seemed so tedious to go in and duck every ESS in a vocal track which already has an amazingly tedious amount of volume automation.
One way to speed things up is to use a spectral analyzer like Stillwells spectro. With spectro You can SEE all the ESS's. Then you just lasso them and turn them down. Done.
Melodyne does that really well, just go in and "draw down" the gain on sibilant blobs, without splitting or tuning or crossfading or otherwise affecting the vocals.
Oh thats cool
Also as mercado pointed out (and Kenny showed me this as well) the ESS component has a specific waveform that is easily identifiable so if you do use the take envelopes it is pretty quick.
So for me there are at least 3 good and fast ways to deal with sibilance
Spectral analyser
Take envelopes
Melodyne (if you own it)