Does anyone know how to eliminate the whistling effect when a commentator speakes words with an S at the beginning and at the end. I have tried some FX available in CS5 but they do not seem very effective.

Depending on your workstation, there are all sorts of de-esser/sibilance plug-ins that do a pretty good job, although a multi band compressors works quite well, tuned for the 'ssss' frequency and set so it drops just that bit of the band. What do you use to edit audio?

PAUL. Thanks for your very quick reply. I have up to now edited the audio on the timeline of Adobe Premiere pro 5 by using all the audio Fx available including De- esser. They don't seem to do the job. I have attempted with a bit of success zooming right in on the audio track and by using the razor cut out some of the whistle before the word starts. You can image how long that takes on a 14 min script with hundreds of words with an S present. With regard to your mention of Multi band compressors, this is above my head, as I have no knowledge of what they are.

Not helpful when it's already recorded, but the root of the issue is (generally) proximity and angle. Talking over instead of into the mic helps the Ps; moving the mic further away helps both. An EQ will help with 'S'ing as it's recorded, so I assume would also help in post.

If you have a parametric equalizer plugin, try this to identify the frequency of the sibilance and then eliminate it. First locate the offending frequency(s):

Make a notch filter by setting the highest Q and highest boost your plugin has. Scan up the frequencies until the offending sound really jumps out at you. (You have found the max of the offending sound.) Then pull the boost down to the lowest cut you have, or at least until the sound is minimized to suit you. If there are harmonics on this sound, go up 2x the frequency and cut there also. You will often find the sibilance between 4 and 5 kHz.

The problem with De-essers is that when it's bad, and you have to use higher settings, what is left sounds very odd, or you get strange 'pumping-swooshing' sounds as each sss happens. A trick in the recording studios is to do what Battle suggests, but with a difference. Create a copy of the audio track. This gives you an uncompressed audio track for everything including the ssss. On the copy of the track identify the range of frequencies where the noise sits, and using the eq - a 32 band graphic often works best, notch out everything apart from the ssss sound. What you then have is a trigger track. Using a dynamics plug in - compressor, and perhaps a gate, increase the difference between silence and sss, paying attention to attack and decay. Next thing is to try to remove/control the sss. If you have a plug-in that is coupled to a graphic, (the multi-band compressor) then you use the side chain input to trigger the compression that drops the level of the SS. The trigger track contains only sound where the ssss is - so this track is routed to the compressor and every time the track has sss, it drops just the band the sss is in.

Other people tackle it by using a 32 band graphic to highlight the sss, but they make two copies of the audio - one with just the sss, and the other with an upside down graphic curve which leaves everything apart from the ssss - then they use their de-esser on just the sss sound, or use it as a trigger, similar to above.

So much depends on how bad it is and what software you have. You are right about the number of times the problem happens - you have lots of work to do, and some may well have to be manually tweaked.

It's not just bad mic technique - some people with gaps in their teeth whistle - and nothing you can do with the mic stops you hearing that!

Does anyone know how to eliminate the whistling effect when a commentator speakes words with an S at the beginning AND at the end.

At the start AND end of words, he or she has a very bad sibilance problem.

My experience is, you either get someone else or try to fix it at the source, especially if it's a series.

If it's that bad, the person knows it and is probably fairly self conscious. If not play him/her a sample of it and ask questions like,
have you commentated before and how did they fix the sibilance? Have you tried to solve the problem yourself?

It could well be a dental problem and maybe easy and inexpensive to fix. Call around to local orthodontists and get some advice,
a checkup maybe in order.

If all else fails, try some fixes above, you need patience and small adjustments so you don't end up with a lisping narrative.
I've seen it happen when the client yelled GET RID OF THAT DAMN SSSS. :)

Cheers.

Thanks everyone for the info and advice offered, unfortunetly I have only the FX that are available in CS5. If I had a 32 multiband graphic I could probally suss out the advice given. But as I haven't, it's probally easier to find another narrator who hasn't got the whistling problem. By the way can you tell where I can get the type of multiband graphic you refer to.

BlueCat make a variety of very well done plugins. This simi-parametric is freeware: Blue Cat's Triple EQ - Semi-Parametric 3 Bands Equalizer Audio Plug-in (VST, AU, RTAS, DX) (Freeware) (http://www.bluecataudio.com/Products/Product_TripleEQ/)
but have a look at their non-free mixers and eqs. usual disclaimer, I just like their stuff, I have no relationship with them.

Their VST version shows up in PPro6 audio effects when loaded into plugins folder.

Also, check out these freeware reviews: http://bedroomproducersblog.com/2011/05/11/bpb-freeware-studio-best-free-graphic-equalizer-vst-plugins/
If you are on Windows, this 31band freeware plugin may be of interest: http://www.softpedia.com/get/Multimedia/Audio/Audio-Plugins/Eq31.shtml