Sibilance "s" issue and proper db level when recording vocals

810

New member
I recently set up a home studio with the following hardware:
-AT2020 Cardioid Condenser Mic
-behringer U-Phoria UM2 2x2 USB Audio Interface with XENYX Mic Preamp
-RF-X Reflexion rear Filter
-standard front p-filter
-also have a FirstAct MVM-88 Uni-directional Dynamic mic

I just spent an hour typing this, then clicked "preview post", only to be asked to sign back in, and I lost all of my work, yet I kept seeing a yellow "draft saved" at the bottom right as I typed. Does anyone know how to find saved drafts on this forum? What a nightmare.

I spent a month learning Audacity, Premiere Pro, Tracktion (now called Waveform), and the above mentioned hardware.

All of my friends vocals sound great while directly facing the AT2020 from a distance of 14 inches, they even move their heads slightly with emotion and gestures. I on the other hand am suffering from a sibilance issue with the letter "s". Every line I say has one or more "s". Individually lowering the db of each of my "s" would be time consuming, especially the regarding the 13 minute track. I have tried every combination of mic height, body rotation/angle of attack, distance, gain level, inside voice/screaming, with none of them reducing the "s" factor to a tolerable level. I happen to be the only one of us that can rap insane fast. I have a unique nasal sound to my voice which I named the "long nose effect". My lyrics reach the limit of my lungs air capacity. I prefer to speak more than yell, and increasing my voice only serves to use more precious oxygen that I can't afford. I beg for assistance regarding this "s" issue. The perk of using the Dynamic mic is that I can kiss the p-filter and use a more inside voice. Yelling does not suit my presentation, but yelling works for everyone else. I enjoy the cheap dynamic mic because I am able to kiss the p-filter and talk more than yell. For some reason, direct monitor only helps when the mic gain and output knobs are way up to the point of distortion, so the direct monitor feature appears to be useless.

Should the direct monitor button be disabled if it is of no use, or should it be kept on?

My third issue is regarding the recommended -15db level for vocal tracks. I understand that headroom is needed so alterations do not result in clipping. I was told that 0db is bad to record at, yet all of my friends sound great with the audio interface mic gain at 10-o-clock, and the beat track at 0db, with their recorded vocal track at -2db, while they are basically yelling. I do not sound as good as them with the same settings. They move and wave their hands while recording vocals and their sound remains consistent, but when I stay perfectly still, some small parts of my vocals end up noticeably quieter, not to mention the "s" factor. How do I properly achieve a -15db vocal recording? The Master Output level window on Waveform 8 only shows the green to yellow to red bars, and lack ruler like db increments, which I feel it should have. The only areas within Waveform 8 that have db values listed, are the volume plugin windows at the end of each track and the db value control window at the bottom middle of the screen for altering the db level of clips. What hardware knobs/software settings need to be set in order to record at -15db as recommended by the pros?

-Should the volume plugin windows at the end of the beat track be at 0db or lower when recording a vocal?
-Should the volume plugin windows at the end of the recorded vocal track be at 0db or lower when recording a vocal, or should the db level be changed after the vocal recording is finished, or does it even matter when the vocal db is changed?
-Is it the mic gain on the audio interface that I would use to achieve the -15db vocal track recording level?
-Is it a combination of the volume plugin window at the end of the recorded vocal track and the audio interface's mic gain that achieve the -15db?

Thank you for taking the time to help us with our efforts.
 
Wow, there's a lot to unpack here. Let's just cover the basics first.

Set the gain knob on the UMC22 to where the peak/clip LED doesn't light up when you're doing your loudest thing. Speak/sing/holler into the mic and look for the peak indicator. If it goes red, turn down the gain. That'll set you up with a solid but not clipping input signal.

For the sibilance issue, try pointing the mic just above or below your mouth instead of right at it. Try a deesser plugin with the frequency set to where your sibilance occurs (somewhere between 4-6 kHz, most likely), and the threshold set to where it attenuates those frequencies by 3-6 dB at first. Crank the threshold lower if it's still severe, but do note that the vocal will start to sound odd with more deessing.

Also, the AT2020 is kind of a bright mic. Maybe a simple high shelf EQ to attenuate the high frequencies would do the trick.

The direct monitor button only matters if you want to hear yourself straight off of the preamp (then turn it on) or if you'd rather monitor yourself through software with any effects applied (then turn it off).
 
I would suggest using the pop filter with any microphone when you are doing vocals, even dynamic mics.

Some microphones are better for some voices than others. It may be that the AT2020 is not the right mic for your voice.

If others are moving around a lot and keeping a constant sound they either have very good mic technique or have employed some compression in their post processing. Keep in mind that compression is going to make your sibilance worse if you don't get that tamed first.

If others are recording with backing track levels at 0dBFS they are clipping the master bus. If that sounds good to you and them, I can't help. Now, we could be talking different things, but you can't put one 0dBFS track into the master and add *anything* without pushing it over 0dBFS. If you bounce it down without attenuating the level, it's going to clip.

So, keep doing what you're doing, but balance your vocal track with the beat/backing track in the mix, then use the standards: level, EQ, compression, reverb, to get a mix that peaks well below 0dBFS on the master. Once you have that, you can apply a limiter to push it up for distribution.

P.S +1 on what [MENTION=22368]Tadpui[/MENTION] said.
 
Wow, there's a lot to unpack here. Let's just cover the basics first.

Set the gain knob on the UMC22 to where the peak/clip LED doesn't light up when you're doing your loudest thing. Speak/sing/holler into the mic and look for the peak indicator. If it goes red, turn down the gain. That'll set you up with a solid but not clipping input signal.

For the sibilance issue, try pointing the mic just above or below your mouth instead of right at it. Try a deesser plugin with the frequency set to where your sibilance occurs (somewhere between 4-6 kHz, most likely), and the threshold set to where it attenuates those frequencies by 3-6 dB at first. Crank the threshold lower if it's still severe, but do note that the vocal will start to sound odd with more deessing.

Also, the AT2020 is kind of a bright mic. Maybe a simple high shelf EQ to attenuate the high frequencies would do the trick.

The direct monitor button only matters if you want to hear yourself straight off of the preamp (then turn it on) or if you'd rather monitor yourself through software with any effects applied (then turn it off).

I will learn how to properly apply your two main tips and will update you on the results. Thank you so much.
 
I would suggest using the pop filter with any microphone when you are doing vocals, even dynamic mics.

Some microphones are better for some voices than others. It may be that the AT2020 is not the right mic for your voice.

If others are moving around a lot and keeping a constant sound they either have very good mic technique or have employed some compression in their post processing. Keep in mind that compression is going to make your sibilance worse if you don't get that tamed first.

If others are recording with backing track levels at 0dBFS they are clipping the master bus. If that sounds good to you and them, I can't help. Now, we could be talking different things, but you can't put one 0dBFS track into the master and add *anything* without pushing it over 0dBFS. If you bounce it down without attenuating the level, it's going to clip.

So, keep doing what you're doing, but balance your vocal track with the beat/backing track in the mix, then use the standards: level, EQ, compression, reverb, to get a mix that peaks well below 0dBFS on the master. Once you have that, you can apply a limiter to push it up for distribution.

P.S +1 on what [MENTION=22368]Tadpui[/MENTION] said.

Geegolly, I hope when you said "If that sounds good to you and them, I can't help." it meant we were on the right track instead of we are damned beyond redemption. thank you for the advice, I will let you know how it turns out.
 
Geegolly, I hope when you said "If that sounds good to you and them, I can't help." it meant we were on the right track instead of we are damned beyond redemption. thank you for the advice, I will let you know how it turns out.
Ok, maybe it was a little sarcastic, but what you posted about 0 and -2 was nonsense, or someone is blowing smoke.
 
Ok, maybe it was a little sarcastic, but what you posted about 0 and -2 was nonsense, or someone is blowing smoke.

It could be that I do not understand the nature of the various ways to alter db levels. The fact is that my friend yelling at the mike from a 14 inch distance with the mic gain at 10-o-clock, and the beat track unaltered and the same as the moment I added it to the timeline which shows it at 0db, makes him sound properly paired with the beat track. His recorded vocal is lowered from 0db to -2db only because it seems a tad in front of the beat. Yes indeed there is serious clipping going on with the beat track. The waveform is almost a solid block. When I lower the db of the beat track as one clip, the waveform shrinks, but retains it's solid nature with teeth that are only visible when zoomed in. You mentioned that I should "attenuate the level" of the beat track before bouncing it down. Does bouncing down mean lowering the db of the beat track? I beg for you to assist me with properly attenuating the levels within Waveform 8. All Tracktion 5+/Waveform 8+ versions are similar in their layout.
 
You need headroom. turn the tracks down till your getting about 10db of headroom before 0db. Or get some kind of meters VU or others. and set your tracks up to around -18db RMS, with no more than -6db peaks. Sibilance. keep the mic up above the mouth area pointing toward the lower nose/upper lip. You can also try this, get a pencil, take a rubber band and put the pencil on center of the mic screen.so it's in the center of the diaphragm, use the rubber band to hold it in place (old trick to kill overly sibilant vocals while tracking).
 
I'm going to suggest something else. How does your voice sound rapping on stage - into an SM58 or something? Also - you could also try something as a guide to the problem. Use your 2020 - not a very sibilance prone in my experience - and record your voice from 3 feet away, with enough gain to get decent level. It's going to sound a bit thin - with the bass tailing off. However, you're not interested in the low end - just the high end. See if you can hear the problem sound with the distant technique. If you cannot - then it's a closeness issue, and could be physical, not electronic. One of my regular clients has a small gap between her two front teeth, and when she sings, a jet of air seems to fire right at the mic. To record her, the mic has to be at her forehead height, firing downwards - a very strange position, but none of the pop shields or foam filters can stop the usual positions being quite horrid. Years ago, I had another client who had a strange nasal sound - and again, we discovered so much of his voice actually came out of his nose that the narrow nasal passages did something similar. A mix of position and EQ cured that one - BUT - he never liked his voice, but weirdly, what he disliked in the recordings was what we hear all the time. I guess the nose sound never got back to his ears.

Have you considered that your voice is just one of those very difficult ones to record? Your sibilant recordings - what do your friends think? Do they sound a problem to them? I'm just wondering if you are suffering from trumpeter syndrome? They never like the sound people record because they have never, ever stood facing their own trumpet and have no idea what it sounds like.

On levels. I'm afraid that I'm ancient, and have never shifted my analogue reel to reel way of working which was never EVER to go into the red, unless I made a mistake. In analogue, mistakes were rarely fatal. They are in digits! I simply adjust the gain knob until the loudest signal doesn't go into the red. If I'm in doubt, I just turn the knob anticlockwise a bit and don't worry. I still make mistakes.
 
It could be that I do not understand the nature of the various ways to alter db levels. The fact is that my friend yelling at the mike from a 14 inch distance with the mic gain at 10-o-clock, and the beat track unaltered and the same as the moment I added it to the timeline which shows it at 0db, makes him sound properly paired with the beat track. His recorded vocal is lowered from 0db to -2db only because it seems a tad in front of the beat. Yes indeed there is serious clipping going on with the beat track. The waveform is almost a solid block. When I lower the db of the beat track as one clip, the waveform shrinks, but retains it's solid nature with teeth that are only visible when zoomed in. You mentioned that I should "attenuate the level" of the beat track before bouncing it down. Does bouncing down mean lowering the db of the beat track? I beg for you to assist me with properly attenuating the levels within Waveform 8. All Tracktion 5+/Waveform 8+ versions are similar in their layout.
Here's how I'd tackle this:
  1. Just use the AT2020 with a pop filter a couple inches in front of the mic and the reflexion screen behind it (optional).
  2. Plug the mic into the Behringer interface, and connect that to the computer.
  3. Create a new project in whichever DAW you are most familiar with.
  4. Add the beat/backing track, but mute it for now.
  5. Stand so you are 6-8" in front of the microphone and begin to sing at the level you want to record at.
  6. Set the GAIN on the interface so it is not clipping and the peak level you see in the DAW for your voice is about -10dBFS.
  7. Now, unmute the backing track, and using "clip gain" or whatever you are used to (perhaps a gain plugin), reduce the beat track level so when you sing along you hear it and your voice.
  8. NOW, record your track along with the backing track playing.
  9. Mix them together, using EQ, compression (watch compression because of your noted "S" problems), etc. on your track (backing track is already mixed and super-compressed).
  10. Watch the level on the Master bus/stereo out. It should be well below 0dBFS, peaks in the -6 to -8dBFS range, perhaps.
  11. Add a limiter to the master bus, and set it to add gain and limit so the mix will peak around -1dBFS. You can play the mix and watch it usually, so it's only adding enough gain to push the peaks a bit over the -1dBFS setting.
  12. Bounce it down and listen/share here for comments.

P.S. You should be monitoring with headphones while recording, and the main point is to lower the beat track level to your vocal, and record at that level. Don't make your vocal match the beat track's mastered/0dB level!
 
Last edited:
You need headroom. turn the tracks down till your getting about 10db of headroom before 0db. Or get some kind of meters VU or others. and set your tracks up to around -18db RMS, with no more than -6db peaks. Sibilance. keep the mic up above the mouth area pointing toward the lower nose/upper lip. You can also try this, get a pencil, take a rubber band and put the pencil on center of the mic screen.so it's in the center of the diaphragm, use the rubber band to hold it in place (old trick to kill overly sibilant vocals while tracking).

That is the first time I heard of the pencil/band application. Does the pencil somehow distort the "s" waves into something else less annoying?
 
I'm going to suggest something else. How does your voice sound rapping on stage - into an SM58 or something? Also - you could also try something as a guide to the problem. Use your 2020 - not a very sibilance prone in my experience - and record your voice from 3 feet away, with enough gain to get decent level. It's going to sound a bit thin - with the bass tailing off. However, you're not interested in the low end - just the high end. See if you can hear the problem sound with the distant technique. If you cannot - then it's a closeness issue, and could be physical, not electronic. One of my regular clients has a small gap between her two front teeth, and when she sings, a jet of air seems to fire right at the mic. To record her, the mic has to be at her forehead height, firing downwards - a very strange position, but none of the pop shields or foam filters can stop the usual positions being quite horrid. Years ago, I had another client who had a strange nasal sound - and again, we discovered so much of his voice actually came out of his nose that the narrow nasal passages did something similar. A mix of position and EQ cured that one - BUT - he never liked his voice, but weirdly, what he disliked in the recordings was what we hear all the time. I guess the nose sound never got back to his ears.

Have you considered that your voice is just one of those very difficult ones to record? Your sibilant recordings - what do your friends think? Do they sound a problem to them? I'm just wondering if you are suffering from trumpeter syndrome? They never like the sound people record because they have never, ever stood facing their own trumpet and have no idea what it sounds like.

On levels. I'm afraid that I'm ancient, and have never shifted my analogue reel to reel way of working which was never EVER to go into the red, unless I made a mistake. In analogue, mistakes were rarely fatal. They are in digits! I simply adjust the gain knob until the loudest signal doesn't go into the red. If I'm in doubt, I just turn the knob anticlockwise a bit and don't worry. I still make mistakes.

I have yet to rap on stage. I have three main voices, my normal rap voice with a bit of nasal,.. a strong news anchor voice with less of a nasal effect, and a sleazy voice that makes me sound more youthful and cartoonish like I am up to no good. The sleazy voice is my favorite because it is very unique, if I were to compare it to someone well known, I would pick Lil Wayne during his close-up calmer vocals. I only have experience with the two mic's within my initial hardware list. I am very glad to hear that the 2020 is not sibilance prone. at a distance of three feet, the "s" sibilance is gone or no longer an issue. It is a closeness issue, but sadly, the 14 inch distance that my friends prefer, does not do much to remove my "s" sibilance. I believe it is physical in nature. However, my teeth do not have a gap, so it must be directly related to my tongue and the roof of my mouth. """the mic has to be at her forehead height, firing downwards""", by downwards, do you mean the mic is hanging down from above, instead of being mounted on a stand? My nose is not abnormally large, but my nasal passages appear to create a long nose effect. My nasal effect is more noticeable to myself as I rap, and less noticeable during playback. I believe my voice is difficult to record compared to my friends. My friends think that a deesser will deal with the "s" sibilance enough, but I would prefer to kill it close to the source to remove the need to edit each "s" individually or hopefully to remove the need for a deesser. I am effected by trumpeter syndrome in a way because the "S" sibilance is exaggerating the "s" more than anyone would notice if I were talking to them in person. It must be the air passing through my mouth that allows a mic to amplify the "s" sibilance to the point of annoyance. The strange aspect is that, even with my head turned 45 degrees from the mic, the "s" sibilance is still noticeable, so the "s" sibilance must be of a more omnidirectional nature instead of a linear stream of air.
 
I think that there will be a position that works best - by reducing the unwanted sounds and emphasising the wanted ones. What I meant by the pointing down positioning was that for the singer I was working with - the mic was on it's normal stand, but the grill was just above her eyebrow line, aiming virtually downwards, aiming at her chest. See singing went forward through the mic's capture zone. This does give a thinner sound than in front, but not much, and was easily EQ's to sound fine. Her 'sss' stream not landing on the mic in any way. You could do the same from the side but then you need to not turn your he'd which is tricky. Why not put the headphones on, hold the mic at arms length and wave it around till you find the magic position.
 
Why not put the headphones on, hold the mic at arms length and wave it around till you find the magic position.

My lord, why didn't I think of that, I often do not see the simple ways of going about things. I will do it until I find a sweet spot and update you.
 
Here's how I'd tackle this:
  1. Just use the AT2020 with a pop filter a couple inches in front of the mic and the reflexion screen behind it (optional).
  2. Plug the mic into the Behringer interface, and connect that to the computer.
  3. Create a new project in whichever DAW you are most familiar with.
  4. Add the beat/backing track, but mute it for now.
  5. Stand so you are 6-8" in front of the microphone and begin to sing at the level you want to record at.
  6. Set the GAIN on the interface so it is not clipping and the peak level you see in the DAW for your voice is about -10dBFS.
  7. Now, unmute the backing track, and using "clip gain" or whatever you are used to (perhaps a gain plugin), reduce the beat track level so when you sing along you hear it and your voice.
  8. NOW, record your track along with the backing track playing.
  9. Mix them together, using EQ, compression (watch compression because of your noted "S" problems), etc. on your track (backing track is already mixed and super-compressed).
  10. Watch the level on the Master bus/stereo out. It should be well below 0dBFS, peaks in the -6 to -8dBFS range, perhaps.
  11. Add a limiter to the master bus, and set it to add gain and limit so the mix will peak around -1dBFS. You can play the mix and watch it usually, so it's only adding enough gain to push the peaks a bit over the -1dBFS setting.
  12. Bounce it down and listen/share here for comments.

P.S. You should be monitoring with headphones while recording, and the main point is to lower the beat track level to your vocal, and record at that level. Don't make your vocal match the beat track's mastered/0dB level!

I reached step 8 using Audacity. I attached an image of my work that includes a red arrow at the bottom left. The arrow is pointing at the -1.0 to 1.0 range window. I do not know what this window is called, or why it only ranges from -1 to 1. Seems to me that it should have a broader range like the top right playback level has so the user can easily see if wave peaks are not above -6db, without the need to play the track and watch the playback/master level. The forum rules stated that this is not the area to ask if someones voice is good or not, so I fear having this thread deleted if I upload the short .mp3 clips of my voice. Would that be ok to do? Thank you so much for the step by step layout. Yourself and everyone else assisting me will get acknowledgement when we become well known.
 

Attachments

  • Audacity 810.png
    Audacity 810.png
    146.5 KB · Views: 20
Last edited:
Mic

I think that there will be a position that works best - by reducing the unwanted sounds and emphasising the wanted ones. What I meant by the pointing down positioning was that for the singer I was working with - the mic was on it's normal stand, but the grill was just above her eyebrow line, aiming virtually downwards, aiming at her chest. See singing went forward through the mic's capture zone. This does give a thinner sound than in front, but not much, and was easily EQ's to sound fine. Her 'sss' stream not landing on the mic in any way. You could do the same from the side but then you need to not turn your he'd which is tricky. Why not put the headphones on, hold the mic at arms length and wave it around till you find the magic position.

I attached an image of my mike in the configuration that I believe you explained, did I get it correctmic.jpg?
 
I reached step 8 using Audacity. I attached an image of my work that includes a red arrow at the bottom left. The arrow is pointing at the -1.0 to 1.0 range window. I do not know what this window is called, or why it only ranges from -1 to 1. Seems to me that it should have a broader range like the top right playback level has so the user can easily see if wave peaks are not above -6db, without the need to play the track and watch the playback/master level. .....
Audacity Waveform Help Page
 
That is the first time I heard of the pencil/band application. Does the pencil somehow distort the "s" waves into something else less annoying?
I'm not sure what the science is. I just know I've seen it used (I'm old lol) and it worked. I also tried it myself when the only vocal mic I had was a very bright mic (my vocals don't do well with bright mics). Take your mic and try it see if it helps.
Some big name recording engineers have use the pencil thing. I think it originally started out as a way to save ribbon mics from getting destroyed by singers, with extra hard xplosives. They found it also can help with the sibilance too. I saw something some years back about,Michael Wagener using it with the vocalist from King's X on a LDC mic.
 
Last edited:
RE: position. No that's 90 degrees wrong! Your mic is a side fire mic, it picks up sideways on the face with the audio technica label. Move the mic so it is horizontal, try with the opposite end to the xlr nearly on your forehead with the brand name on the bottom.
 
Back
Top