Help needed with SSSibilance

I’m attempting to narrate and produce my first audiobook after having pros narrate my first two books. I’m using Win 8.1 with a Blue Yeti (60% gain, Pop filter, 8" distance) and 2.2.2 Audacity. After doing a sample audio and mastering with your recommended guidelines of Equalize-Low Roll ff,5000, then RMS Normalize -21, then Limiter-Soft,0,0,-3.5,10,No, then 6,6,6 Noise Reduction, I have a sample that passes ACX Check.

Unfortunately, I have a very noticeable whistle sibilance when I pronounce s or sh. It’s the same problem in the raw file or mastered. I also tried fishspit de-esser but could not improve it. Maybe I don’t have the settings right or need to use a different plug-in. I figured out that it is in my normal speaking voice and I never noticed it until now. I’ve tried correcting my voice but a lifetime of habit won’t let me sustain different than normal tongue positions without screwing up my delivery. I sent my sample out to two friends for feedback without telling them what I was concerned about. Both of them said that the s’s are a problem for the listener. One said the t’s were also too shrill.

Later I tried recording with a thin sock over the mic as well as the pop filter a couple inches from my mouth. It helped somewhat but knocks down the RMS 2db. Maybe I’ll have to live with that and let the RMS Normalize boost it back up.

Here’s a 20 second unmastered sample that I hope you can have a listen to and give me some ideas. Thanks.

Your 28 foot what?

That’s classic “Essing,” and you can fight with the de-essing tool, or you can try a simple equalizer I drew. I switch from raw to corrected about 9 seconds.

It should have about the same effect as the sock, but it’s more tuned to the crisp tones, so less effect on everything else. You can’t take out too much because that will cause serious muffling.

Let us know.


very noticeable whistle sibilance

Actually, you don’t. You have plain sibilance, an evil boosting of tones between about 4000Hz and 7000Hz. Whistle sibilance is very serious. That’s a wholesale changing some voice tones into other ones. I don’t know you can fix that in post production.


I got a similar effect with the DeEsser tool with these settings.

Screen Shot 2018-09-11 at 10.10.15.png
I don’t think DeEsser comes with stock Audacity. I need to find it.


The idea is to leave your normal “S” tones alone and just get rid of the “ice pick in the ear” sound.

These adjustments get harder as your ears fall apart.




I’m pretty sure this is current. I know they made changes to it and updated this one.

I’m also including DistantSeas.xml which is the equalizer custom setting. That’s basically fancy-pants tone controls.

DistantSeas.XML (461 Bytes)
Download that and import it into Effect > Equalizer.

Open any audio > Effect > Equalization > Save/Manage… > Import > Select the DistantSeas.xml file > Open.


I know how to change the XML file to increase the effect. Let me know.

I do have some homework for you. Crumple up a newspaper sheet about a foot in front of the microphone and record it as you would normally. Try to find a sheet of a real paper like New York Times or Washington Post. Some free newspapers use different paper and the effect is different.

If you start overloading Audacity (meter goes all the way up), just back up until it stops.

This is a poor man’s microphone analysis tool.



Thanks so much for that info and instructions. I listened to your corrected file and downloaded your equalizer settings
and applied them to the entire raw file without any other mastering. I heard an improvement to the final 9 seconds
you edited but the strongest sibilance is in the first 7 seconds on the words “swarming” and “sailing” and I am not
hearing much difference there when I applied your custom equalizer. It did seem to clear up all the other ones to an
acceptable level. I’ll do some more experiments. Could you go back to my original file and see if you can correct the
first part as well and let me know if we need to change some settings?

Tomorrow I’ll try to download the de-Esser plug-in you linked to and get back to you on the results. Assuming I can get
that to work, do you have a preference for which one is simpler to use or provides better correcting for sibilance? I’m
new to mastering so it’s been quite a struggle to get this far.

I don’t have a good newspaper here but uploaded the crinkled paper file here using advertising paper that is similar
I hope. I guess you will use that to check the health of my mic? I kept the gain on 60% same as I used for my other
tests but was redlining the meter too much so had to back away to about 18 inches.


Spitfish is too blunt an instrument to fix whistles.
Paul-L’s de-esser plugin on high-resolution settings can remove it, (takes real-time though) …

remove whistle on 's'.gif
Then follow with Spitfish.

I went over the sample and mastered it in this order:

  1. EQ
  2. RMS Normalize
  3. Limiter
  4. De-Esser set to -25, 2.0, 3000 to 5555, 10, 5.0
  5. Noise Reduction 6,6,6

If you could have a listen to this sample compared to the unmastered one at the top of the thread I’d like to know if you think it is good enough for the listener not to get annoyed at the remaining sibilance. I can only detect a small improvement but maybe it’s enough?

Does anyone think I can improve it further with different de-esser settings?

I heard an improvement to the final 9 seconds you edited but the strongest sibilance is in the first 7 seconds on the words “swarming” and “sailing”

I got lost. My sample post was raw, uncorrected for the first 9 seconds and then “DistantSeas” filtered for the rest. You were intended to apply that custom equalizer plugin to a different raw reading and see if it’s enough.

We only correct a portion of the sample in order to be clear the difference between before and after. If we didn’t do that, you would have to swap back and forth between two different sound files to hear the difference. That’s a bit more of a juggling act and it’s easy to be listening to the wrong thing by accident. Of course, that never happens to me.

I need to leave for a while.


We’re splitting fine hair. Try one with the first DeEsser value Threshold set to -30dB rather than -25dB. That should make the corrections a little quieter and hopefully less obvious. That’s what it did last time I tried it.

Remember start with a fresh, raw read file each time. This is a good time to get used to Exporting live performances as WAV (Microsoft) 16-bit. You can do many different chores with that file and not have to read it again.


OK will do. I just tried using -30 on the original file and that helps. I think I can go forward at this point. When I save files in Audacity I have the setting at Windows 16 bit but for some reason when I open the files later they have reverted back to 32 bit.

Everything opened in Audacity is converted into 32-bit.