I'm looking for a digital filter (DirectX if possible) to remove
narrowband high frequency noise. The noise peaks are only a couple of
Hertz "wide" so Q must be really very very very very high.
I've tried
* Waves Q8 - frequency is not exact, Q not high enough
* SoundForge Paragraphic EQ - Q not high enough
Thanks for any recommendations.
Johann
--
pjiogfdöhuiofsdühuiosdfohuinjklvsdfasdüjiofsdüjiofasdüjiofasdoüjifsdjis
dffasdjisdfaiojdfasjisdfaüjiosdfaioüjsdfaoüjifasdüjiodfasüjiodfasioüjsd
fadfasüisdfasdfasdsdfdfaasdfasdüjsdfpjirü0jioüjiodgfüfsdgpjidgfsdgfadgf
asüpsdgf (Walter Anger in <20041020185752.5d5cb077.WalterAnger@aon.at> )
Johann Burkard wrote:
> Hello,
>
> I'm looking for a digital filter (DirectX if possible) to remove
> narrowband high frequency noise. The noise peaks are only a couple of
> Hertz "wide" so Q must be really very very very very high.
>
> I've tried
>
> * Waves Q8 - frequency is not exact, Q not high enough
> * SoundForge Paragraphic EQ - Q not high enough
>
> Thanks for any recommendations.
>
> Johann
>
You may be confusing Q for notch depth. You do not NEED the filter to
have a high Q. If the filter is deep, it will remove the NB tone. If
the Q is high, it will have less effect on the surrounding frequencies.
If the Q is too high, the tuning will be very critical. Are the noise
peaks also stable with frequency? I don't think a moderate Q will hurt
the sound much. What you really need is notch depth. N track studio
has an EQ with a nice notch filter that may work for you.
"Johann Burkard"
> Hello,
>
> I'm looking for a digital filter (DirectX if possible) to remove
> narrowband high frequency noise. The noise peaks are only a couple of
> Hertz "wide" so Q must be really very very very very high.
** Hmmmm, a high frequency, very narrow band noise is a sine wave tone or
fixed pitch whistling sound. This is so because any other noise sort of
noise is not so narrow band.
Is this what you are dealing with - or not ?
Your phrase "noise peaks " is a worry - makes it sound like brief spikes
of noise which are always wide band.
Johann Burkard <johannburkard@nexgo.de> wrote:
>
>I'm looking for a digital filter (DirectX if possible) to remove
>narrowband high frequency noise. The noise peaks are only a couple of
>Hertz "wide" so Q must be really very very very very high.
Nahh, you can get away with a surprisingly wide filter without anybody
noticing it, at high frequencies. Even an old Little Dipper is pretty
transparent when you're notching above 6 KC. Any parametric should be
fine... you might see a big chunk out on an FFT plot, but I bet you
can't hear it.
If you can, well, it might be time to skip filtration and go the CEDAR
route.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Scott Dorsey wrote:
> Johann Burkard <johannburkard@nexgo.de> wrote:
>>I'm looking for a digital filter (DirectX if possible) to remove
>>narrowband high frequency noise.
> Nahh, you can get away with a surprisingly wide filter without anybody
> noticing it, at high frequencies.
Well I have seven noise peaks to remove, from ~11100 Hz to ~16900 Hz.
With all the not-so-narrow-filters, that's a lot of high frequency
attenuation - and it's audible.
> If you can, well, it might be time to skip filtration and go the CEDAR
> route.
I assume CEDAR has FFT filters? Maybe I could write something like that
myself.
Johann
--
Stammtischprolet Burgratte mit Geweine in der Signatur.
(*Tönnes wird heiß in <cia9gt$qhh$03$1@news.t-online.com> )
>Hello,
>
>I'm looking for a digital filter (DirectX if possible) to remove
>narrowband high frequency noise. The noise peaks are only a couple of
>Hertz "wide" so Q must be really very very very very high.
You may try FFT filters. I use CoolEdit, nowadays Adobe Audition
for that.
Mark wrote:
> I know you won't like this answer, but it may be interesting to pass
> this through an MP3 encoder or other perceptual encoder and see what it
> does.
>
> It may remove some of your tones because it thinks they are masked
> anyway.
I just tried OGG @ 320 and 192 KBit/s and WMA @ 192 KBit/s - the noise
is encoded and not removed.
> What level are these tones compared to the program?
> Can you hear them?
See <http://johannburkard.de/misc/spectrum_noise_1.png> and
<http://johannburkard.de/misc/spectrum_noise_2.png>. Yes, it's audible
Johann
--
HASS! HASS! HASS! Du schäumst förmlich vor Hass.
(*Tönnes in <cd15mj$lkk$06$4@news.t-online.com> )
In article <32vs76F3s6k0bU1@individual.net> philallison@tpg.com.au writes:
> "Johann Burkard"<
> >
> > Well I have seven noise peaks to remove, from ~11100 Hz to ~16900 Hz.
>
>
> ** No you do not.
>
> There are no such things as "noise peaks".
>
> That must be your own invented term .
A commonly accepted definition of "noise" is anything coming out that
you didn't put in. Distortion also meets this criteria, and by many,
it's lumped in with "noise."
I would suggest, since the frequencies were well defiined, that the
original poster is seeing amplitude peaks at those frequencies when he
looks at a "spectrum analyzer" display in his DAW. Since he didn't put
them there in the first place (or so he thinks - we don't know that),
to him this is "noise" and they are definitely "peaks" so while it may
not be correct to call them "noise peaks" his description is clear.
> Answer my previous points - or admit you are yet another PITA troll.
No, a troll is someone who posts a message that knowingly will cause
controvery, often off topic of the forum or newsgroup. Like the one to
which I am responding.
--
I'm really Mike Rivers (mrivers@d-and-d.com)
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me here: double-m-eleven-double-zero at yahoo
Johann Burkard wrote:
> Mark wrote:
> > I know you won't like this answer, but it may be interesting to
pass
> > this through an MP3 encoder or other perceptual encoder and see
what it
> > does.
> >
> > It may remove some of your tones because it thinks they are masked
> > anyway.
>
> I just tried OGG @ 320 and 192 KBit/s and WMA @ 192 KBit/s - the
noise
> is encoded and not removed.
>
> > What level are these tones compared to the program?
> > Can you hear them?
>
> See <http://johannburkard.de/misc/spectrum_noise_1.png> and
> <http://johannburkard.de/misc/spectrum_noise_2.png>. Yes, it's
audible >
> Johann
> --
> HASS! HASS! HASS! Du schäumst förmlich vor Hass.
> (*Tönnes in <cd15mj$lkk$06$4@news.t-online.com> )
Try N Track Studio, it has a free demo and even if you buy it, its not
expensive.
The EQ can be set for band stop i.e. notch filter, and with the
highest Q setting it can create a notch 10 kHz 20 dB deep with only 1
to 2 dB loss at 9 kHz. (At least thats what the display shows, I
didn't actually sweep it to verify that).
> Scott Dorsey wrote:
> > Johann Burkard <johannburkard@nexgo.de> wrote:
> >>I'm looking for a digital filter (DirectX if possible) to remove
> >>narrowband high frequency noise.
>
> > Nahh, you can get away with a surprisingly wide filter without anybody
> > noticing it, at high frequencies.
>
> Well I have seven noise peaks to remove, from ~11100 Hz to ~16900 Hz.
> With all the not-so-narrow-filters, that's a lot of high frequency
> attenuation - and it's audible.
>
> > If you can, well, it might be time to skip filtration and go the CEDAR
> > route.
>
> I assume CEDAR has FFT filters? Maybe I could write something like that
> myself.
The code for such filters isn't terribly complicated - you'll find some good
sources online.
Getting the code to talk to your application would be the tricky bit imho.
Norbert Hahn wrote:
> Johann Burkard <johannburkard@nexgo.de> wrote:
>>I'm looking for a digital filter (DirectX if possible) to remove
>>narrowband high frequency noise.
>
> You may try FFT filters. I use CoolEdit, nowadays Adobe Audition
> for that.
Thanks, Audition seems like a good idea. Will check it out.
Johann
--
du bist nur ein doofer beleidiger warum verziehst du dich nich ich eins
von deine forums
("Bernd P." in <b1mm67$13p8g0$1@ID-156972.news.dfncis.de> )
Mark wrote:
> The EQ can be set for band stop i.e. notch filter, and with the
> highest Q setting it can create a notch 10 kHz 20 dB deep with only 1
> to 2 dB loss at 9 kHz. (At least thats what the display shows, I
> didn't actually sweep it to verify that).
The problem is that I have at least seven of these frequencies that I
need to take out.
Johann
--
Millionen Menschen koennen dich nicht unterhalten, da muss erst Toennes
im From: stehen und schon erweist du dem Kollegium° die Ehre deiner
aktiven Niederkunft im Torfmoor°. Fragen?
(*Tönnes in <cj4inl$irg$05$1@news.t-online.com> )
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.