|
Post by scumbum on Mar 20, 2016 21:33:41 GMT -6
Chris Lord-Alge taught me something I've had problems with ever since I started recording , how to fix vocal sibilance .
In this video he says to EQ vocals before the compressor , not after the compressor , if you EQ after the compressor you'll get vocal sibilance . Never knew that . I gave it a try and hes right !!
For me this saves a huge amount of time automating faders and crap , battling sibilance .
|
|
|
Post by svart on Mar 21, 2016 7:21:44 GMT -6
Yeah, I found that on my own a while ago, but never knew it was worthy of a tip. I usually run two different EQs, one pre compressor, one post compressor.
I do all the cutting pre-compressor, and all the boosting(if any) post compressor. Seems to work best that way. Boosting into the compressor can be useful, but can make those boost frequencies sound strange if pushed too much.
|
|
|
Post by NoFilterChuck on Mar 21, 2016 7:38:24 GMT -6
right.
1) remove the crap you don't want compressed. 2) compress the remaining good stuff. 3) remove any crap resulting from compressing
|
|
|
Post by Ward on Mar 21, 2016 7:52:18 GMT -6
Anyone ever heard of a de-esser?
|
|
|
Post by svart on Mar 21, 2016 8:14:04 GMT -6
Anyone ever heard of a de-esser? Yeah, it's essentially a frequency based compressor. Sometimes I use the de-esser, sometimes the cut-before-compression, depending on what works better. I find that the cut-before-compressor just sounds better, since you're actually taking away the offending area so that the compressor does not act upon it, unlike the de-esser which singles out that frequency for the compressor to act upon it. Just another tool in the toolbox I suppose though.
|
|
|
Post by Johnkenn on Mar 21, 2016 9:11:51 GMT -6
Honestly, I had never really thought about this...and I've been using EQ's after the compressor...I'm certainly gonna give it a try.
|
|
|
Post by tonycamphd on Mar 21, 2016 9:15:05 GMT -6
Improper mic placement, and over compressing is the prime cause of sib/slice and tizz, prox effects on mics that are too close bring up less energetic sibs/slice/tizz, and abused comps bring louder freqs down toward the level of naturally less energetic sibs/slice/ tizz freqs, then gaining them up, both mistakes give unnatural and unpleasant presence energy to the ugly stuff, and both mistakes are often made in tandem.
|
|
|
Post by scumbum on Mar 21, 2016 9:49:35 GMT -6
Honestly, I had never really thought about this...and I've been using EQ's after the compressor...I'm certainly gonna give it a try. Me too , always used to boost after the compressor , and I've always had to deal with nasty sibilance . Thats a thing of the past now . Setup an EQ before and after a compressor with the same settings on both EQ's . Then listen to what each one does to the sibilance . Post compressor will make some nasty sibilance that will need to be dealt with big time . Pre-compressor EQ , you might only need a light touch of a de-esser or none at all .
|
|
|
Post by jazznoise on Mar 21, 2016 9:49:35 GMT -6
It's usually bad mic technique + extremely toppy mic + heavy handed compression + bright reverb that causes this problem. I rarely find myself having issues with sibilance, but obviously doing a high freq boost before your limiter makes it more likely to tame hard S and T sounds.
|
|
|
Post by tonycamphd on Mar 21, 2016 10:23:29 GMT -6
Honestly, I had never really thought about this...and I've been using EQ's after the compressor...I'm certainly gonna give it a try. Me too , always used to boost after the compressor , and I've always had to deal with nasty sibilance . Thats a thing of the past now . Setup an EQ before and after a compressor with the same settings on both EQ's . Then listen to what each one does to the sibilance . Post compressor will make some nasty sibilance that will need to be dealt with big time . Pre-compressor EQ , you might only need a light touch of a de-esser or none at all . Scum, what eq r u using after comp that causes that? Curious minds want to know ?
|
|
|
Post by Johnkenn on Mar 21, 2016 10:38:15 GMT -6
I hardly ever boost anything...Maybe 2khz on a bass...or the occasional Acoustic top end.
|
|
|
Post by warrenfirehouse on Mar 21, 2016 10:57:23 GMT -6
Ive seen him talk about this somewhere else too. I think he is referring to boosting top end into the compressor instead of after. Obviously a technique if the goal is a bright vocal, or you wouldnt be boosting at all.
And I agree Scum....it sounds alot better to me too.
I also do this on parallel drum bus...boosting top and bottom into the compressor
|
|
|
Post by scumbum on Mar 21, 2016 11:59:32 GMT -6
Me too , always used to boost after the compressor , and I've always had to deal with nasty sibilance . Thats a thing of the past now . Setup an EQ before and after a compressor with the same settings on both EQ's . Then listen to what each one does to the sibilance . Post compressor will make some nasty sibilance that will need to be dealt with big time . Pre-compressor EQ , you might only need a light touch of a de-esser or none at all . Scum, what eq r u using after comp that causes that? Curious minds want to know ? Sony Oxford EQ I've always had sibilance issues at mix time . I usually boost at 3.5K . I started using an sm57 to try and avoid sibilance . I'm gonna go back to my condenser and see if I can use it again now that I'm boosting before the compressor .
|
|
|
Post by cowboycoalminer on Mar 21, 2016 17:54:42 GMT -6
Chris Lord-Alge taught me something I've had problems with ever since I started recording , how to fix vocal sibilance . In this video he says to EQ vocals before the compressor , not after the compressor , if you EQ after the compressor you'll get vocal sibilance . Never knew that . I gave it a try and hes right !! For me this saves a huge amount of time automating faders and crap , battling sibilance . Look at all these plug ins behind him in that video picture.
|
|
|
Post by geoff738 on Mar 21, 2016 18:02:10 GMT -6
Going in and editing the offending frequencies out seems to work best for me. You can actually get pretty quick at it, and usually see where they are just from the waveform. It's pretty unique.
And remember too that de-essing is not just for troublesome vocals. Cymbals, fizzy highs on a guitar amp? Try de-essing.
Cheers, Geoff
|
|
|
Post by rowmat on Mar 22, 2016 5:39:27 GMT -6
I generally EQ before compressing. EQ'ing the 'woof' out of an acoustic guitar BEFORE the compressor is a no brainer unless you want the compressor working overtime trying to clampdown the low end while hardly touching anything else. As far as de-essing goes, if it is needed, we have a SPL Hardware de-esser, several plugin de-essers and the good old 'do it manually in an editor'. Which de-essing method is used depends on $$. However the best method I've found is to duplicate the vocal track, cut out everything leaving only the 'esses', any plosives and hard consonants intact and then phase reverse the 'esses' track and adjust the level of the track to control the level of sibilance cancellation against the original vocal track. It sounds better than any hardware unit or plugin but can add anywhere up to an hour of extra work per vocal take. (The SPL hardware de-esser also uses phase cancellation to attenuate the sibilance but still requires arbitarily chosen filters which are subject to some degree of compromise IMO.)For the manual de-essing method somewhere between 4 to 7db of sibilance cancellation (attenuation) does the job nicely and will allow a fair bit of high end boost without getting spitty. Others boost the highs before de-essing but I usually de-ess before EQ'ing. I'll give the client a demo of this method and most do prefer it once they hear how natural the esses sound even after adding top end EQ. The SPL hardware de-esser spends most of its time patched into both the reverb sends to the 480L.
|
|
|
Post by M57 on Mar 22, 2016 6:28:04 GMT -6
However the best method I've found is to duplicate the vocal track, cut out everything leaving only the 'esses', any plosives and hard consonants intact and then phase reverse the 'esses' track and adjust the level of the track to control the level of sibilance cancellation against the original vocal track. So when you say cut out, do mean not just slicing, but high and low pass as well, right? I tried this recently, and maybe it's because it was my first time, but it was pretty labor intensive. It did work, but I found that unless the levels are dead nuts on, things can go wonky pretty quickly. I found myself automating the fader to get it right. Bottom line for me: I ended up using it only on the most egregious of problem sibilances because of how tricky it was to get right. Cool trick though.
|
|
|
Post by rowmat on Mar 22, 2016 6:54:29 GMT -6
However the best method I've found is to duplicate the vocal track, cut out everything leaving only the 'esses', any plosives and hard consonants intact and then phase reverse the 'esses' track and adjust the level of the track to control the level of sibilance cancellation against the original vocal track. So when you say cut out, do mean not just slicing, but high and low pass as well, right? I tried this recently, and maybe it's because it was my first time, but it was pretty labor intensive. It did work, but I found that unless the levels are dead nuts on, things can go wonky pretty quickly. I found myself automating the fader to get it right. Bottom line for me: I ended up using it only on the most egregious of problem sibilances because of how tricky it was to get right. Cool trick though. Sorry I was somewhat lacking in detail with my explanation I slice the DUPLICATED vocal track each side of the 's' and remove all the audio between the 'esses'. So I end up with an entire track of just little individual 's' waves. Then I glue all these 'esses' together into a single wave file. So now I have one track containing a single wave file of 'esses'. This track is then phase reversed and the level adjusted against the original vocal track (on the same buss) to control the amount of sibilance reduction. If both tracks are set at the same level the 'esses' cancel completely. As you begin to lower the level of the 'esses' track you will hear the sibilance begin to return. Reducing the level of the 'esses' track between -4 to -7db (down from '0') is usually about right. If you want to boost the high EQ quite heavily for a very bright airy vocal then de-ess say at -4 as this will provide more de-essing than at -7 but the EQ will make up for it. Basically use you ears and adjust to taste. Some additional targeted level adjustment maybe required but generally I find not that much is needed. The louder 'esses' meet their louder out-of-phase counterparts while the quieter 'esses' meet their quieter out-of-phase versions. You could even say that the 'esses' meet their 'nem-esses'. I then usually apply and EQ, verb etc. I like fine tuning the amount of cancellation with EQ and any verbs etc enabled as it lets you adjust the de-essing in context as you can balance the amount of sibilance to remain natural with the applied EQ and especially bright plate reverbs. I'll point out the more you do, the faster you get at it, and you can begin to identify which 'esses' are going be more, or less, of an issue by simply their visual appearance. Some 'esses' you can simply ignore altogether if they are low enough in level.
|
|
|
Post by warrenfirehouse on Mar 22, 2016 7:04:27 GMT -6
You could even say that the 'esses' meet their 'nem-esses'. Bravo. Slow clap.
|
|
|
Post by M57 on Mar 22, 2016 7:21:43 GMT -6
>>I slice the DUPLICATED vocal track each side of the 's' and remove all the audio between the 'esses'. So I end up with an entire track of just little individual 's' waves. Then I glue all these 'esses' together into a single wave file. So now I have one track containing a single wave file of 'esses'.<< I think you were clear from the start. What I'm getting at is that you "have one track containing a single wave file of.." the sections of the track that have not just esses, but ALSO whatever other sounds are accompanying the esses in that moment. I'm suggesting that high and low pass could be applied to the 's' track so the focus is on phase reversing the esses and only the esses.
|
|
|
Post by rowmat on Mar 22, 2016 7:22:38 GMT -6
Bravo. Slow clap. I take donations!
|
|
|
Post by henge on Mar 22, 2016 7:22:42 GMT -6
Chris Lord-Alge taught me something I've had problems with ever since I started recording , how to fix vocal sibilance . In this video he says to EQ vocals before the compressor , not after the compressor , if you EQ after the compressor you'll get vocal sibilance . Never knew that . I gave it a try and hes right !! For me this saves a huge amount of time automating faders and crap , battling sibilance . Look at all these plug ins behind him in that video picture. hahahah!!
|
|
|
Post by rowmat on Mar 22, 2016 7:33:12 GMT -6
>>I slice the DUPLICATED vocal track each side of the 's' and remove all the audio between the 'esses'. So I end up with an entire track of just little individual 's' waves. Then I glue all these 'esses' together into a single wave file. So now I have one track containing a single wave file of 'esses'.<< I think you were clear from the start. What I'm getting at is that you "have one track containing a single wave file of.." the sections of the track that have not just esses, but ALSO whatever other sounds are accompanying the esses in that moment. I'm suggesting that high and low pass could be applied to the 's' track so the focus is on phase reversing the esses and only the esses. If you mean as applies to an entire mix then absolutely. However with a discrete vocal track I think you'll find, in most cases, the 'esses' are predominantly fairly well isolated on their own in terms of frequency content. However I'm not in disagreement with you about avoiding attenuating non sibilant frequencies. If you zoom right in and observe any 's' waveforms there is generally not much 'non sibilant' energy superimposed within the main body of most 'esses'. The non sibilant frequencies tend to taper off leading into the sibilance when the 's' occurs at the end of a word. The reverse tends to apply when an 'S' begins a word. It's like a crossfade from sibilance to 'non sibilance'. However 'Z's tend to have more of a mixed content. Try it and you'll see! Obviously some kind of bandpass filtering would reduce attenuating non-sibilance related material that can occur in concurrently with the sibilance itself but I think filtering may then also introduce some non-lineraties in the phase cancellation process of the sibilance itself as I believe occurs to a degree in the SPL de-esser. If the SPL hardware de-esser and the manual DAW method I use both employ phase cancellation to reduce sibilance then why do I (and others) find the DAW method more natural sounding? I find with most de-essers it's just too fine a line between "Not enough de-essing" to all of a sudden "I can hear it actually de-essing."
|
|
|
Post by M57 on Mar 22, 2016 8:15:40 GMT -6
^^ All true - I've been aware that most of the time the s snippets contain little in the way of pitch information outside of the target frequency - I'm just wondering how critical it is to be persnickety about these things. My guess is that unless you have a growly gravely singer, taming the lower frequencies probably isn't that critical, but higher frequencies include the "air" in the room (when you're capturing the space in any way - intentionally or not) will be sucked out along with the bath water and could potentially cause unwanted breathing, right?
|
|
|
Post by tonycamphd on Mar 22, 2016 8:15:42 GMT -6
I think you were clear from the start. What I'm getting at is that you "have one track containing a single wave file of.." the sections of the track that have not just esses, but ALSO whatever other sounds are accompanying the esses in that moment. I'm suggesting that high and low pass could be applied to the 's' track so the focus is on phase reversing the esses and only the esses. If you mean as applies to an entire mix then absolutely. However with a discrete vocal track I think you'll find, in most cases, the 'esses' are predominantly fairly well isolated on their own in terms of frequency content. However I'm not in disagreement with you about avoiding attenuating non sibilant frequencies. If you zoom right in and observe any 's' waveforms there is generally not much non sibilant energy superimposed within the main body of most 'esses' except tapering off either the very beginning, or right at the tail end of the 'esses'. However 'Z's tend to have more of a mixed content. Try it and you'll see! Obviously some kind of bandpass filtering would reduce attenuating non-sibilance related material that can occur in concurrently with the sibilance itself but I think filtering may then also introduce some non-lineraties in the phase cancellation process of the sibilance itself as I believe occurs to a degree in the SPL de-esser. If the SPL hardware de-esser and the manual DAW method I use both employ phase cancellation to reduce sibilance then why do I (and others) find the DAW method more natural sounding? I find with most de-essers it's just too fine a line between "Not enough de-essing" to all of a sudden "I can hear it actually de-essing." this is a great trick, the idea is to duplicate the vocal track EXACTLY, all plugs to the letter, then delete all but the esses in the duplicated tracks waveform, then you just slide the duplicated tracks fader up to attenuate the esses to taste, if you feel the need to filter above and below the center freq of your ess problem, just use a linear phase eq to filter, otherwise the resonances and phase smearing of a traditional eq/filter will do some weird stuff.
|
|