$$$ for help removing high pitch "ssss" sound

[paudieshea]

Hi, I am looking for some help removing high pitch “ssss” whistle sounds from my audio clip? I have tried many things so far but I am new to audio editing and need some guidance. I have about 80 clips with 160 minutes of audio. I need to get this problem fixed ASAP. I am willing to pay $50 to any one that can give me clear directions as how to remove this high pitch whistle “sss” sound.
So far I have tried de-esser in the following programs without much success:
Adobe Audition
Adobe Sound Booth
ProTools.
I have also tried to copy “s” without the whistle to replace the bad “s” but this does not sound natural.
I have tried fadeouts on the “s” sounds but it keeps the high pitch whistle.
I have tried amplitude reduction on the “s” sounds but it keeps the high pitch whistle.
And I have tried some mixing and matching of all the above without much look.
I can improve the sound but not completely remove it.


m7s10.take3.wav

[paudieshea]

Right Click on the m7s10.take3.wav link above and select "save target as" to download a clip. Note the words "goals", "LifeGoals", "objectives", "Tasks" "activities", "timlines".

[lordsmurf]

I've been restoring audio for 15+ years now, digitally.
While I have a lot of posts here at Videohelp on the topic, there are quite a few more here:
- http://www.digitalfaq.com/forum/forumdisplay.php/restore-filter-improve-17.html
- http://www.digitalfaq.com/guides/video/index-restore.htm

I'll look at your file.

UPDATE: Sorry, but that can't be fixed. The person reading is the problem, she hisses and whistles when she speaks, holding on to plural letters like a snake. Software can't correct that, it needs to be done up front, both with a person that is trained to speak, as well as with a better microphone at an optimal speaking distance. The recorder sounds like its gain is too high, too -- cooking the audio a bit.

[MOVIEGEEK]

I did this in Audacity using low pass filter set to 800Hz:


after.wav

I agree with LS that the problem is her speech pattern, you can't do much with a lisp. Another problem is the source is MS ADPCM which accentuates the hiss, if you could get the files in PCM then the quality would be better:
http://support.microsoft.com/kb/89879

[paudieshea]

Hey LordSmurf,

Let me know how you get on with the .wav file.

[paudieshea]

Hey MovieGeek,

I think that the file you have worked on is no better than the versions i have being able to create myself. Any other ideas??

[MOVIEGEEK]

Hey MovieGeek,

I think that the file you have worked on is no better than the versions i have being able to create myself. Any other ideas??

Get a microphone and narrate it yourself.

[lordsmurf]

Or get it redone by a proper voice actor.

[netmask56]

Re-record with a better microphone, pay attention to peak levels and maybe try getting the narrator to speak across the microphone rather than directly into it. Most of the problem is the nature of her voice. I tried using the Waves de-essor plugin to no avail. Where there are gaps between words I was able to manually attenuate the sss's but not within sentences - you really need to re-record under better conditions or find another narrator. You could fiddle around with a notch filters. The easiest way to find the frequency band is to set it to peak and try to maximise the annoying sound then reverse to attenuate.

[freebird73717]

paudieshea
First of all welcome to the Forums!

Secondly
There is really no need to offer a reward or compensation here. Pretty much all of the regulars here answer questions for free just because they like helping people.

I didn't check out your file because I'm not great at audio so I'd be no help to you there but the others that answered are very good at that stuff so if they say that there isn't much hope to correct then I'd say you are pretty much out of luck.

[Cornucopia]

Where's my $50???

Seriously, it would take some more tweaking to get it just right, but here is a sample of just doing this in Audition:

1. Couple of Deep & Narrow Notch filters @ ~4700Hz, ~6800Hz, ~8600Hz
2. Linear Fade-down ramps on the terminal "S"es.
3. A Dynamic FFT filter that has a broad notch cut centered at ~6300Hz (kicks in when level gets above certain dB).

There are other things you could do. Contrary to what others have said, it's NOT unusable, just a challenge.

Scott

m7s10.take3_dess.wav

EDIT: (took me about 10 minutes of fiddling, could do more with more time, and using ProTools instead)

[Xpenguin17]

I'm able to isolate the hissing and shrills from the rest of the audio, but I can either remove the top-freq shelf of them completely, making her sound like a typical AZN assgoblin who never pronounces plurals, or I can attenuate them, which makes them less annoying but still sounds unnatural and out of whack.

Choose.

[joecass]

there's really no way to eliminate the exaggerated sibilance that I can figure out..... using many audio software programs I have...... but I did try one thing, highlighting the ends of the sentences, and fading out the offensive
sounds to a minimum.... doesn't get rid of the problem, but sounds a bit better

edited_m7s10.wav

[Deter]

I can actually help on something.
This was an easy fix. However the recording itself is terrible.

The problem is how the lady speaks in to the mic.

I would re-record it, since it is just 1 person talking and the quality is bad.

[Ruggie]

I think this does what you want.
Although there is no filter that I know of that does this automatically, it is a rather simple fix.

When viewing the waveform, her S's show up as a very tight wave form(even when in the middle of a word) and is fairly easy to spot right away.
I just found the first peak and the last peak of the waveform and deleted everything between the two(You can't delete at the peak because you will still get the ramp up, so find a spot where its fairly level)

It took me maybe 5 minutes to do, and shortens her S's from around .2 seconds to around .05 seconds.

I attached my edited version based off the original sample.

Please let me know if this was helpful!

[spankurmonkey]

HI.

I did a google search and found this thread and was wondering if some of you can review my file and let me know what can be done if anything to the files to remove the high pitchs/whistles, etc.

I have asked the person in the audio to narrate a book for me . I like his voice but the "s" noise was not apparent at the time he did an audition recording for me.

Looking for any help. I have asked him to move the mic, etc. However beyond that I am not sure he is technical to know how to clean the audio if anyway.

I know a little bit about Audacity... well I know how to open and run but beyond that I am green.

[Mephesto]

Downsample to lower samplerate. It'll soften the Ses to Fs. Simplest way to deal with this.
I'd help out more if I didn't lose access to my main drive.
I noticed the audio is talking about addiction treatment. As someone who survived a crippling withdrawal after going cold turkey I can only be pessimistic. Don't get your hopes up about treating anyone. Drugs succeeded to do to me what being locked up at age 14 without a conviction didn't. Your patients are already dead.

[newpball]

HI.

I did a google search and found this thread and was wondering if some of you can review my file and let me know what can be done if anything to the files to remove the high pitchs/whistles, etc.

You need a de-esser, any good audio editor has one.

[spankurmonkey]

Is that something to be used with Audacity ?

[Steve(MS)]

I have seen a free plugin for audacity for that somewhere, do a search and see if you can find it.
However, I have no clue how well it would work.

[newpball]

Is that something to be used with Audacity ?

I am sure there are plugins for it.

I usually use Sonar and like the IZotope Nectar plugin for voice filtering.

[davexnet]

Here's one example using the Spitfish de-esser:
https://www.youtube.com/watch?v=6EGqbtXfAJA

[joecass]

I used a few software programs to try and "de-ess" the sound, however it didn't have much of an effect. The only way to solve the problem is to have the narrator re-record the dialogue using a compressor-limiter on the microphone before the recording device. In typical recording studios, they use what's called a "pop filter" on the mic, which reduces 'plosives' and 'esses' on a recording.