I have an audio track for a movie which has hiss at 'S' letters. When you hear words like 'messy' or 'assessor' you hear the 'S' letters as strong, distracting hiss. Does anyone know the name of this kind of noise, and how to remove? Please comment... any help will be greatly appreciated.
P.S.: There's no problem with the hiss on idle.
+ Reply to Thread
Results 1 to 8 of 8
Originally Posted by redwudz
One question: will it affect the whole audio track (and may lead to quality loss in general), or it has only effect on the 's' sounds?
I haven't needed to use the filter, but it seems popular. You can run it with Audacity by installing the VST Enabler: http://audacity.sourceforge.net/help/faq?s=install&i=vst-enabler
Just about any filter will cause some collateral 'damage'. Most times the key is finding a compromise between removing the noise or 'over-filtering' the audio to the point it's damaged worse. I always seek improvement, never a perfect output. Hiss or 'S' sounds are particularly hard to filter as they often contain many frequencies, similar to 'white noise'. But Spitfish is designed for your type of problem.
If your source is vocal, you can sometimes cut off the higher frequencies and lower the 'hiss' level with a low pass filter without too much damage to the rest of the audio range. Expirerimenting with a software audio equalizer may help to show what's possible.
It sounds like someone may have already made a mess of the audio file and it may be very difficult to even get it back to it's original condition.
Never delete the original audio till you are sure you have made an improvement.
Spitfish will only affect a very small frequency window, where the "S" sounds live - around 3500hz (or 5000hz for femle voices)
Originally Posted by Soopafresh
What's your opinions about this (a hint for de-essing in Adobe Audition 1.5):
To get rid of the sibilance, try going to Spectral View. Once you get the resolution set high enough (I like to use a setting of 4096), the sibilant frequencies will pop right out as bright yellow strips. You can then use the marquee selection tool, select the sibilant frequencies (the brightest yellow ones somewhere in the 4k-8k range, usually), and use the Amplitude effect to drop the selection down by 12 dB or so. Best de-esser in the business, in my experience!
Any of them work
I've stuck with this problem. The Adobe Audition trick was interesting, but somehow Spitfish seemed better. I've tried many settings but I still wasn't satisfied with the end result. I had to apply Spitfish at least 5 times to hear noticable difference, but the difference was still little, while all the rest of the audio (I mean the parts where weren't any recognizable 's' sound) became a little muffled, flat. So overally Spitfish didn't make it better.
I guess this "deesser" thing would be more effective if not a frequency range in the whole file but only the 's' parts would be touched. Isn't there a plug-in or audio editor which can look for the parts where the 's' sounds are really recognizable and applies filtering only for that certain parts?
I suspect this may not be the only problem with the audio track I'm processing, only a symptom of bad recording in general. I've uploaded a 24 seconds long sample, so if an expert could give it a listen and would tell his opinion, I'd appreciate that.
http://www.badongo.com/audio/11435763 (Thank you in advance. )