Why's it hard to record "s" sounds cleanly?

D

dan_aka_jack

Well-known member
Why do "s" and (to a lesser extent) "f" sounds produce problems for sound recording kit? You know - the way that an "s" sound in dialogue sounds a bit like it's distorting, even when the levels are nice and safe?

What's going on? Does the production of the "s" sound pump out a bunch of white noise which freaks out the audio recording hardware?

Thanks,
Jack
 
Plosives cause air to move across the mic which can cause this problem. The solution is use a pop screen / softie or change the position of the mic to minimize the movement of air. I'm making the assumption that your mic is close to the speaker.
 
Consonants such as s, ch, sh and z can produce sibilance on a recording. This typically involves a pronounced level spike (sometimes as much as 6 dB) in the 5-10 kHz frequency range. Close-mic'ed vocals are rather prone to sibilance, and many shotgun mics have a real problem with this. Try repositioning the mic to reduce sibilance. A specialized type of dynamics compresser called a "de-esser" (widely available as a software plugin) can be used to reduce or remove sibilance from a recording.

Of course, you could also be dealing with a plosive issue, as Tigeba is sugesting. If the unwanted sound is high-pitched, its probably a sibilance. If the unwanted sound is low-pitched, its probably a plosive.
 
HI guys,

Thanks loads for your quick replies! Most of the audio recording that I've been doing recently has been using a lavalier mic over a Sony wireless system. I was wondering if perhaps the wireless system is accentuating the sibilance problem? I guess the answer is "test, test, test"! Before posting, I spent a while reading some of the great threads here on DVXUser including threads about why wires are always better than wireless and the importance of good cables.

How does a de-esser work? Does it detect high levels in the 5-10kHz range and soft-clip them?

Many thanks,
Jack
 
dan_aka_jack said:
How does a de-esser work? Does it detect high levels in the 5-10kHz range and soft-clip them?
Click to expand...

Kind of. It compresses (turns down) the signal when it detects excessive levels in that range. It's not really anything you couldn't do yourself, lowering the volume everytime it happens, the de-esser just takes the work out of it. :smile: (And does it consistantly.)
 
Ah, interesting - thanks. I already use a little bit of dynamic range compresion - will a de-esser work well alongside a "normal" compressor? Should I put the de-esser before the "normal" compressor in the processing pathway?
 
dan_aka_jack said:
... using a lavalier mic over a Sony wireless system.
Click to expand...

Sibilance may be overloading your transmitter. Since sibilance can hit much hotter than the rest of dialogue, you may have to back off your compander setting on the transmitter. If this is the case, and you're distorting on your transmitter, then a de-esser will only only de-emphasize the distortion.
 
For what it is worth a de-esser is merely the combination of an equalizer and a compressor with a sidechain input. Instead of compressing the full spectrum, the equalizer on a de-esser allows you to select a more narrow spectrum to apply the compression. It isn't the same as just lowering the volume in selected spots because this would attenuate the entire signal.
 
Um, maybe I'm misunderstanding, but if you use an equalizer as a sidechain for a compressor, that will only affect what triggers the compressor, not what frequecies are compressed. Not saying a de-esser isn't frequency-specific compression, but it would be a little more complicated than that, no?
 
So then it would work as I described, compressing the entire bandwidth whenever the signal goes over the threshold in the frequencies specified by the eq. Unless you have one of the newer phase-inverting de-essers, but that's different than the compressor + eq sidechain we're both talking about.
 
You must log in or register to reply here.
Back
Top