Aug 24, 2014 - On Windows I found it is necessary to apply that mda de-esser repeatedly for it to have an effect comparable with SpitFish, so you may have to.
Why sound editing software can be your friend. One common problem with recording the human voice is sssssomething called ssssssibilance. If you didn’t already know what “sibilance” meant, my odd spelling above may have clued you in. It’s a sharp, biting hissing sound that happens whenever the recorded voice utters the “S” sound.
You’ll be listening to the recording, and everything will sound great, until the speaker/singer utters the phrase “ she’ s a sen sation,” and pierces your eardrums. The good news is that you can use free sound editing software to fix this! There are lots of reasons why this might happen, over-compression (see more info on compression here: and ), wrong kind of mic for that voice, too much reverb effect, etc. Just know that if you do any recording or audio editing of human voices, you will periodically be faced with sibilance that cannot make it onto the final version of the audio file. So what do you do? You’ve may have heard of something called a “de-esser,” yeah? De-essers are effects that are designed to filter and tame sibilance, while leaving the non-sibilant part of the audio alone.
If you are familiar with the range of human hearing and the use of EQ (equalization) tools in audio (see my article about EQ here: ), you’ll know that people can only hear sounds that exist between the frequencies of 20 hertz (HZ) and 20,000 Hz. In practical terms, most adults are going to have trouble hearing anything over about 13,000 Hz. What’s most interesting about this fact is that certain familiar sounds “live” at predictable frequencies.
No it IS interesting! Don’t argue with me;).
If we know, for example, that the “bass-y, boom- y parts of a bass guitar will most often be found around 80-100 Hz. So we can then we use an EQ to turn down (or up.but usually down.never mind) the bass by ONLY turning the volume down around 80-100 Hz. Likewise, high sounds like the sizzle of a high-hat, can be found hanging out around 6,000 Hz (or 6KHz for short). So if you knew that vocal sibilance mostly hung out between 4 KHz and 10 KHz, you would know where to start looking for your particular brand of hissiness so you could fix it. A de-esser is simply a tool – found in most sound editing software programs – that zeros in on a certain group of frequencies (like 4-10KHz or 5-8KHz, etc.) and will turn the volume down in that region, but ONLY when it gets TOO loud. Smart tool huh?
As long as the “s” sounds are normal for human speech, they are left untouched. But the instant they get unnaturally hissy, they’ll get turned down just enough to make them sound normal. It takes some playing with the settings to get this just right. Overdo, though, it and your singer/speaker will develop a lovely lisp. If you don’t have a de-esser, a compressor, or an EQ (de-essing is really just a combination of compression and EQ), there is one sure-fire way to fix sibilance which is my first choice, especially if the audio is short, like in a podcast intro or a short song.
Instead of treating the entire file as you would do with a de-esser, just LOOK at the audio in an editor. It’s so cool that we can use our eyes to edit audio now. Sigh, OK I’m a geek.
Anyway, you can usually see sibilance on a computer screen pretty easily, using any sound editing software available. The squiggly blobs get all bunched up and dark for a bit. In the picture, you can see the word “serendipity’ recorded. The blue blob at the beginning. Just follow the cursor and when you hear the sibilance, you’ll likely see it as well.
Now zoom into JUST the sibilant sound and highlight it. Use the editor’s volume control to turn down ONLY the “s.” Listen to the result and if that didn’t fix it all the way, or you gave the speaker a lisp, just “undo” and try again until you get it just right.
Rinse and repeat for each offending sibilant sound, and voila! Your hissing snake problem is fixed! I recommend the latter method of fixing sibilance in your audio if the file is only a minute or two or less. For longer files (audio books, etc.) you’re probably better off using a de-esser.
James, Sibilance is hard to remove without introducing other artifacts into the recording or making it sound 'dull'. With that said, Spitfish is probably your best free option.
Sibilance is mostly in the 5-10khz range, so to help reduce it in the first place try recording with the EQ down a bit in that range. Best to use a parametric EQ to pinpoint the most offending frequencies so that you can notch them out. There are good free VST parametric EQ's you can download and plug into Audacity if you use the VST enabler. Another option is to put something over the mic that will dampen the upper frequencies a bit - like a thick foam windscreen or cotton sock. Hi Ray, Yeah, that's the obvious place to put them, however, it's not the RIGHT place to put them.! The VST Bridge is based partly on code developed from Steinberg (which is why it's a seperate download) so it expects either to have a parameter set in the OS which tells it where to find VST plugins, or else it expects you to use the defult folder: C: Program Files Steinberg VSTPlugins Obvious, isn't it???? (NOT!) Once you put the 3 dll's in here, they should be available when you next start Audacity.
Note though, that you don't get any fancy GUI's in Audacity - just direct access to the parameters! Hope this helps! Hi Ray, Yeah, that's the obvious place to put them, however, it's not the RIGHT place to put them.! The VST Bridge is based partly on code developed from Steinberg (which is why it's a seperate download) so it expects either to have a parameter set in the OS which tells it where to find VST plugins, or else it expects you to use the defult folder: C rogram FilesSteinbergVSTPlugins Obvious, isn't it???? (NOT!) Once you put the 3 dll's in here, they should be available when you next start Audacity.
Note though, that you don't get any fancy GUI's in Audacity - just direct access to the parameters! Hope this helps!
Who would figure? Brandon is 100% correct (another audio geek among us, cool!). De-essers by their very function introduce artifacts and alter the balance of the audio. They are essentially EQ at their core.
Audacity is not really the ideal tool for this but I would try zooming in, marking each SSS, then knock each instance down a few db with a pencil tool (which redraws the audio waveform). That will de-emphasize the sibilance without altering the overall timbre or tone of the voice track as you would with a cheap or free de-esser. Like anything, with audio there's a degree of finesse involved and practical experience that comes into play in terms of your outcome. The less experience you have with advanced audio editing the less likely that you'll get a satisfactory result without a fair amount of playing around with it. Doesn't mean it's not worth doing (it is.) just means that you shouldn't give up just because you can't push a button and have it come out perfect in 10 seconds. On the other hand of this equation is whether or not it's faster to just RETRACK it with attention to mic technique and minding your S's. Copy a snip of the worst sibilance and let me hear it, James and I'll be able to give you better advice.