A good rule-of-thumb to use in cases of resampling (which is more or
less what this is all about): Eliminate all frequecies above half of
the target sample rate.

To reproduce a sound, you'll need at least two samples, and a 11025
Hz tone will require a sample rate of 22050 bytes / second to
reproduce.

Therefore, if your target sample rate is 22050 Hz, you won't be able
to reproduce anything past 11025 Hz, meaning you can eliminate it. A
few applications (such as Sound Forge, if I remember correctly)
automatically apply various EQ's and anti-alias filters before
resampling.

Also, frequencies in the direct vicinity of said frequency (in this
case, everything from about 9000 to 11025 Hz) will probably suffer
from distortion and weird effects as well (known as mirroring, I
think) and you'd do well in reducing them a bit. The de-esser might
help, but there's no way to completely avoid these artifacts. You
cannot reproduce the frequencies involved, since they are destroyed
in the compression.

Good luck!

Regards, Johan @ o3

Brian Franz wrote:
>>
>> I stand corrected, Thanks Charles.
>>Charles de Montebello wrote:
>>>>Well, yes and no. First off, Sonic Foundry DOES make a De-Esser.
>>It's
>>>>in the ever-useful Multi-band dynamics plugin which ships with
>>Vegas
>>>>2.0. There you essentially could have a de-esser, a de-popper,
etc.
>>>>all in the same plugin.
>>>>
>>>>But, I think you are most probably dealing with the artifacts
>>>>produced by encoding to this format. Generally the high
frequencies
>>>>are the ones to demonstrate quality-loss the most obviously. The
>>>>sound you will hear is not unlike the sound you hear when you
over-
>>>>noise reduce. Like a chirping sound. Esses are in that frequency
>>>>range and longer in length, so you will hear problems with
encoding
>>>>more easily in these cases. You really need to encode at very
high
>>>>bit rates to remove sibilant esses from an .asf file. De-essing
and
>>>>rolling off the high end are BOTH good suggestions for the
>>prevention
>>>>of overly jarring side-effects.
>>>>
>>>>c.
>>>>
>>>>Brian Franz wrote:
>>>>>>The Lisp you're talking about is a common problem with certain
>>>>Vocals
>>>>>>and voice over people. The real answer that Dave isn't able to
>>>>tell
>>>>>>you, is that you need a De-esser. Some compressors that have a
>>>>>>sidechain would also work....but then you have to know how a de-
>>>>esser
>>>>>>works. Waves makes a De-esser plugin. A De-esser turns down
>>>>certain
>>>>>>frequencies when they exceed a thresold value (that's how they
>>>>>>technically say it). Basically the "S" sound is in the
frequency
>>>>>>range of 4Khz-8Khz. SF doesn't make a De-esser, so that's why
>>Dave
>>>>>>probably can't recommend the real solution. The Waves de-esser
>>is
>>>>>>very simpe to use and works well. If you're not able to get
>>ahold
>>>>of
>>>>>>a De-esser then try using Dave's method, more specifically, try
a
>>>>>>notch EQ and notching out around 7Khz.
>>>>>>
>>>>>>Dave Hill wrote:
>>>>>>>>Seems like some audio compression artifacts- perhaps apply a
>>>>little
>>>>>>>>eq to the voice and roll of some of the high end? It might
take
>>a
>>>>>>bit
>>>>>>>>of experimenting to properly tweak your settings, both on the
>>>>Vegas
>>>>>>>>side and the encoder side-
>>>>>>>>
>>>>>>>>"Preview in Player" is the tool to use for this- select a
small
>>>>>>>>region with the offending sounds (3-4 secs) and fine tune
your
>>>>>>>>settings, then save as a template.
>>>>>>>>
>>>>>>>>
>>>>>>>>Joseph G. Lansing wrote:
>>>>>>>>>>How can I prevent a "lisp" sound in narration in an asf?
>>>>>>>>>>I'm limited to using MPEG-3, MS MPEG4 V3 at 100kbps. I've
>>>>>>>>>>tried boosting audio to 24 and 32 bit, but no great results.
>>>>>>>>>>
>>>>>>>>>>The original .wav I'm using is very clean, and
>>>>>>>>>>professionally recorded. Any ideas?