Problem with this plug-in in stereo

classic Classic list List threaded Threaded
13 messages Options
Reply | Threaded
Open this post in threaded view
|

Problem with this plug-in in stereo

David R. Sky
Hi,

Attached is rectimod.ny , Rectifier with Modulation under the
effects menu. (Explanation below.)

This plug-in works great with mono signals, but when applied
sequentially more than once to a stereo signal (panned to center),
the signal gradually separates into two separate signals in the
left and right channels. Is it something with my code or Audacity
Nyquist?

Thanks

David

* rectimod.ny : Rectifier with Modulation

A rectifier takes a signal (such as a sinewave or guitar track)
which goes above and below the zero line, and either returns only
the positive portion of the signal ("half rectification"), or it
also returns the negative portion of the signal which has been made
positive ("full rectification"). The result of such rectification
is added higher frequency harmonics not in the original signal.

Regular rectification makes the entire signal positive, so re-
applying a rectifier adds no further higher frequency harmonics.
However, this plug-in takes an audio signal, applies half or full
rectification, then converts the all-positive result into another
signal which crosses the zero line. this means the plug-in can be
re-applied to the result, adding additional high-frequency
harmonics with each application.

this plug-in also has four options for modulating the resulting
rectified signal and mixing it back with the original signal. For
instance, the default settings have the original selected signal in
Audacity fade out while the rectified signal fades in.

Variables:

1. Rectification type - half or full rectification.

2. Modulation type -
up (fade in),
down (fade out),
constant (same level over time),
LFO (low frequency oscillator, a sine wave which rises and falls
below 20 cycles per second).

3. LFO frequency - frequency of the LFO signal, when LFO modulation
is chosen.

4. Rectified volume - maximum volume of the rectified signal, in
percent.


Many guitarists use rectifiers as part of their 'effects suite'.
the noisier the original guitar signal (from fuzz boxes, distortion
pedals, etc.), the noisier will be the result from this plug-in.

Works on mono and stereo audio, best results occur with mono input
signals.


Written by David R. Sky, November 27, 2005.

Thanks to Steven Jones for a Nyquist patch to make the rectifier
possible.
Released under terms of the GNU Public License
http://www.opensource.org/licenses/gpl-license.php

Other Nyquist plug-ins are freely available for download from:

http://audacity.sourceforge.net/download/nyquistplugins/

rectimod.ny (3K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

edgar-rft

Hi David,

The bug is caused anyhow by the offset calculation in the rectify function.
With stereo tracks one channel is shifted upwards and the other one downwards.

I assume you use the offset of 0.475 to compensate the direct current
component caused by the rectification. Unfortunately a value of half
maximum peak only works with square wave signals. To calculate the offset
needed for a complex signal is pretty complicated. If you find a good
formula for this you can make a lot of money.

In electronics this is usually solved by using highpass filters because
a constant signal or a slow changing signal cannot pass a highpass filter
and a highpass filter is easy to calculate. I suggest a cutoff frequency
of 20Hz or something around this.

But I first have to have a closer look at the code to find out wether
or where exactly a highpass filter could help to solve this.

I will write you more in a couple of hours.

Cheers,

- edgar




the transcript of your original request follows


[hidden email] schrieb am 27.11.05 20:13:15:

>
> Hi,
>
> Attached is rectimod.ny , Rectifier with Modulation under the
> effects menu. (Explanation below.)
>
> This plug-in works great with mono signals, but when applied
> sequentially more than once to a stereo signal (panned to center),
> the signal gradually separates into two separate signals in the
> left and right channels. Is it something with my code or Audacity
> Nyquist?
>
> Thanks
>
> David
>
> * rectimod.ny : Rectifier with Modulation
>
> A rectifier takes a signal (such as a sinewave or guitar track)
> which goes above and below the zero line, and either returns only
> the positive portion of the signal ("half rectification"), or it
> also returns the negative portion of the signal which has been made
> positive ("full rectification"). The result of such rectification
> is added higher frequency harmonics not in the original signal.
>
> Regular rectification makes the entire signal positive, so re-
> applying a rectifier adds no further higher frequency harmonics.
> However, this plug-in takes an audio signal, applies half or full
> rectification, then converts the all-positive result into another
> signal which crosses the zero line. this means the plug-in can be
> re-applied to the result, adding additional high-frequency
> harmonics with each application.
>
> this plug-in also has four options for modulating the resulting
> rectified signal and mixing it back with the original signal. For
> instance, the default settings have the original selected signal in
> Audacity fade out while the rectified signal fades in.
>
> Variables:
>
> 1. Rectification type - half or full rectification.
>
> 2. Modulation type -
> up (fade in),
> down (fade out),
> constant (same level over time),
> LFO (low frequency oscillator, a sine wave which rises and falls
> below 20 cycles per second).
>
> 3. LFO frequency - frequency of the LFO signal, when LFO modulation
> is chosen.
>
> 4. Rectified volume - maximum volume of the rectified signal, in
> percent.
>
>
> Many guitarists use rectifiers as part of their 'effects suite'.
> the noisier the original guitar signal (from fuzz boxes, distortion
> pedals, etc.), the noisier will be the result from this plug-in.
>
> Works on mono and stereo audio, best results occur with mono input
> signals.
>
>
> Written by David R. Sky, November 27, 2005.
>
> Thanks to Steven Jones for a Nyquist patch to make the rectifier
> possible.
> Released under terms of the GNU Public License
> http://www.opensource.org/licenses/gpl-license.php
>
> Other Nyquist plug-ins are freely available for download from:
>
> http://audacity.sourceforge.net/download/nyquistplugins/
>

--
The author of this email does not necessarily endorse the
following advertisements, which are the sole responsibility
of the advertiser:

______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

David R. Sky
Hi Edgar,

Thanks very much, that was terrific advice! I applied a highpass8 filter
at cutoff 20Hz to both stereo channels after I had used the rectifier,
then re-applied the rectifier then the highpass8 filtrer, and did this 3
or 4 times - the signal sounds like it remains in the center pan position.
so I'll incorporate the filter into the plug-in.

Does this same thing happen with the mono signal? - that is, does a low
frequency component get added to the result and appear like a lop-sided
rectangle on the screen?

0.475 is used as the offset because the input signal is normalized to a
level of 0.95 I subtract that from the rectified signal and multiply by 2
to give it a +/-0.95 peak-to-peak amplitude, at least I assume that's what
actually happens, that was my intent. *chuckle*

Thanks Edgar!

David

On Sun, 27 Nov 2005, [hidden email] wrote:

>
> Hi David,
>
> The bug is caused anyhow by the offset calculation in the rectify function.
> With stereo tracks one channel is shifted upwards and the other one downwards.
>
> I assume you use the offset of 0.475 to compensate the direct current
> component caused by the rectification. Unfortunately a value of half
> maximum peak only works with square wave signals. To calculate the offset
> needed for a complex signal is pretty complicated. If you find a good
> formula for this you can make a lot of money.
>
> In electronics this is usually solved by using highpass filters because
> a constant signal or a slow changing signal cannot pass a highpass filter
> and a highpass filter is easy to calculate. I suggest a cutoff frequency
> of 20Hz or something around this.
>
> But I first have to have a closer look at the code to find out wether
> or where exactly a highpass filter could help to solve this.
>
> I will write you more in a couple of hours.
>
> Cheers,
>
> - edgar
>
>
>
>
> the transcript of your original request follows
>
>
> [hidden email] schrieb am 27.11.05 20:13:15:
>>
>> Hi,
>>
>> Attached is rectimod.ny , Rectifier with Modulation under the
>> effects menu. (Explanation below.)
>>
>> This plug-in works great with mono signals, but when applied
>> sequentially more than once to a stereo signal (panned to center),
>> the signal gradually separates into two separate signals in the
>> left and right channels. Is it something with my code or Audacity
>> Nyquist?
>>
>> Thanks
>>
>> David
>>
>> * rectimod.ny : Rectifier with Modulation
>>
>> A rectifier takes a signal (such as a sinewave or guitar track)
>> which goes above and below the zero line, and either returns only
>> the positive portion of the signal ("half rectification"), or it
>> also returns the negative portion of the signal which has been made
>> positive ("full rectification"). The result of such rectification
>> is added higher frequency harmonics not in the original signal.
>>
>> Regular rectification makes the entire signal positive, so re-
>> applying a rectifier adds no further higher frequency harmonics.
>> However, this plug-in takes an audio signal, applies half or full
>> rectification, then converts the all-positive result into another
>> signal which crosses the zero line. this means the plug-in can be
>> re-applied to the result, adding additional high-frequency
>> harmonics with each application.
>>
>> this plug-in also has four options for modulating the resulting
>> rectified signal and mixing it back with the original signal. For
>> instance, the default settings have the original selected signal in
>> Audacity fade out while the rectified signal fades in.
>>
>> Variables:
>>
>> 1. Rectification type - half or full rectification.
>>
>> 2. Modulation type -
>> up (fade in),
>> down (fade out),
>> constant (same level over time),
>> LFO (low frequency oscillator, a sine wave which rises and falls
>> below 20 cycles per second).
>>
>> 3. LFO frequency - frequency of the LFO signal, when LFO modulation
>> is chosen.
>>
>> 4. Rectified volume - maximum volume of the rectified signal, in
>> percent.
>>
>>
>> Many guitarists use rectifiers as part of their 'effects suite'.
>> the noisier the original guitar signal (from fuzz boxes, distortion
>> pedals, etc.), the noisier will be the result from this plug-in.
>>
>> Works on mono and stereo audio, best results occur with mono input
>> signals.
>>
>>
>> Written by David R. Sky, November 27, 2005.
>>
>> Thanks to Steven Jones for a Nyquist patch to make the rectifier
>> possible.
>> Released under terms of the GNU Public License
>> http://www.opensource.org/licenses/gpl-license.php
>>
>> Other Nyquist plug-ins are freely available for download from:
>>
>> http://audacity.sourceforge.net/download/nyquistplugins/
>>
>
> --
> The author of this email does not necessarily endorse the
> following advertisements, which are the sole responsibility
> of the advertiser:
>
> ______________________________________________________________
> Verschicken Sie romantische, coole und witzige Bilder per SMS!
> Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Audacity-nyquist mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
>


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

edgar-rft
In reply to this post by David R. Sky

Hi David,

answer to your question about the rectifier effect on mono signals:

yes, the effect is the same. The whole track starts to get shifted more and
more negative the more you apply the rectifier. The amount of shifting is
dependent to the waveform. It seems as if it gets shifted towards the
negative normalisation value of -0.95.

I still have no other solution than a highpass filter but as soon as the
Audacity website translation is finished I will explore the whole thing
because with guitar effects my all time favourite was first a rectifier,
then a fuzz limiter and finally a wahwah like bandpass filter.

- edgar



______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

David R. Sky
Hi Edgar,

I did a brief patch which isn't really complete yet, I gotta put the hp8
filter somewhere else. And an lp8 filter too, I think with enough
rectification the harmonics will go above 20khz.

Cheers

David

On Mon, 28 Nov 2005, [hidden email] wrote:

>
> Hi David,
>
> answer to your question about the rectifier effect on mono signals:
>
> yes, the effect is the same. The whole track starts to get shifted more and
> more negative the more you apply the rectifier. The amount of shifting is
> dependent to the waveform. It seems as if it gets shifted towards the
> negative normalisation value of -0.95.
>
> I still have no other solution than a highpass filter but as soon as the
> Audacity website translation is finished I will explore the whole thing
> because with guitar effects my all time favourite was first a rectifier,
> then a fuzz limiter and finally a wahwah like bandpass filter.
>
> - edgar
>
>
>
> ______________________________________________________________
> Verschicken Sie romantische, coole und witzige Bilder per SMS!
> Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Audacity-nyquist mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
>


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

David R. Sky
In reply to this post by edgar-rft
Hi Edgar,

I've attached rectimod.ny , with highpass8 at 20Hz and lowpass8 at 20khz,
filtering the (sim ...) statement. Is this adequate, or do I need to apply
the filters to the rectified signal before it gets re-mixed with the dry
signal?

Thanks

David


rectimod.ny (3K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

edgar-rft
In reply to this post by David R. Sky

Hi David,

I just did something similar the last two hours:

I have added a high pass filter to the rectifier function. Now you can
apply the effect as often as you like wihout causing a up or down
shifting or a bias, as I assume, is the correct term. Shifting usually
means a shifting in time. But it was me who had started to use confusing
terms.

The new function is:

(defun rectify (signal rectifyt offset)
(normalize (mult 2.0 (hp
(if (= rectifyt 1)
(sum (pos s) (neg s))
(pos s))
20)) ; cutoff frequency of the highpass filter
))

The offset now can be completely ignored because it is eliminated by the
high pass filter. A little drawback is that in the first fifty milliseconds
now there is a short glitch because the highpass filtes needs some time to
come to work. But this is the same with electronic filters and not a
Nyquist bug.

I have tried several other high pass filters up to highpass8 but the higher
the q factor and therefore the cutoff steepness the worse becomes the ripple
in the first few milliseconds. A simple high pass worked best.

I have sent the modified plugin to your private adress. Please have an ear
on it wether it still sounds like you had wanted it to be and if you like
it we can post it back to the list.

Have a nice day,

- edgar


the rest is the transcript of the original message

[hidden email] schrieb am 28.11.05 09:39:07:

>
> Hi Edgar,
>
> I've attached rectimod.ny , with highpass8 at 20Hz and lowpass8 at 20khz,
> filtering the (sim ...) statement. Is this adequate, or do I need to apply
> the filters to the rectified signal before it gets re-mixed with the dry
> signal?
>
> Thanks
>
> David
>
>

--
The author of this email does not necessarily endorse the
following advertisements, which are the sole responsibility
of the advertiser:

______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

edgar-rft
In reply to this post by David R. Sky

Hi David,

I have just tested both versions yours versus mine and I have discovered
that both versions have the same problem: the glitch in the beginning.

If you apply the effect several times in a row the normalize function
tries to keep the glitch which is much higher then the rest of the sound
so the overall volume starts to get lower and lower.

I still have no real good solution for this but will keep trying.

- edgar



the rest is the transcription of your last mail

[hidden email] schrieb am 28.11.05 08:26:09:

>
> Hi Edgar,
>
> I did a brief patch which isn't really complete yet, I gotta put the hp8
> filter somewhere else. And an lp8 filter too, I think with enough
> rectification the harmonics will go above 20khz.
>
> Cheers
>
> David


--
The author of this email does not necessarily endorse the
following advertisements, which are the sole responsibility
of the advertiser:

______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: Problem with this plug-in in stereo

David R. Sky
Hi Edgar,

Do you mean the glitch which starts when the highpass filter is first
activated?

Anyways, I've attached undcbias.ny , "DC bias removal", which came with
the Audacity 1.3.0 beta I got from leland, it looks like one line of code

(hp s 0.1)

and something commented out which I don't understand. Anyways, thought
this might help.

David

On Mon, 28 Nov 2005, [hidden email] wrote:

>
> Hi David,
>
> I have just tested both versions yours versus mine and I have discovered
> that both versions have the same problem: the glitch in the beginning.
>
> If you apply the effect several times in a row the normalize function
> tries to keep the glitch which is much higher then the rest of the sound
> so the overall volume starts to get lower and lower.
>
> I still have no real good solution for this but will keep trying.
>
> - edgar

undcbias.ny (244 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: hp filter was Problem with this plug-in...

David R. Sky
In reply to this post by edgar-rft
Hi Edgar,

I just had an idea - what about an 'inverse' normalization process to
remove the dc bias?, using s-min to find the lowest point of the audio,
comparing this with the result of the s-max (or peak) function, and
adjusting accordingly? That way we wouldn't need the hp filter?

Thanks

David


On Mon, 28 Nov 2005, [hidden email] wrote:

> The offset now can be completely ignored because it is eliminated by the
> high pass filter. A little drawback is that in the first fifty milliseconds
> now there is a short glitch because the highpass filtes needs some time to
> come to work. But this is the same with electronic filters and not a
> Nyquist bug.



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: hp filter was Problem with this plug-in...

Sami Jumppanen
On 28/11/05, David R. Sky <[hidden email]> wrote:
> Hi Edgar,
>
> I just had an idea - what about an 'inverse' normalization process to
> remove the dc bias?, using s-min to find the lowest point of the audio,
> comparing this with the result of the s-max (or peak) function, and
> adjusting accordingly? That way we wouldn't need the hp filter?

Hi,

if you guys just understood something about the audio itself instead
of just witing DSP code... ;)

Firstly, we cannot "center" the audio by the use of min / max peaks.
What counts, is the "weight balance" of the signal. To find out that,
we need to calculate the average with n samples. However, we cannot do
that with a fixed number of samples, nor with the whole audio file -
because the average varies constantly.

We need to remove the DC component, and DC (or high-pass) filter is
just the tool for doing that. In real life, we let the audio go
through a capacitor with suitable resistance to ground to discharge
the capacitor.

--
Sami "Some-E" Jumppanen
[hidden email]
http://netti.nic.fi/~some-e/


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
<a href="http://ads.osdn.com/?ad_idv37&alloc_id865&op=click">http://ads.osdn.com/?ad_idv37&alloc_id865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
Reply | Threaded
Open this post in threaded view
|

Re: hp filter was Problem with this plug-in...

David R. Sky
Hi Sami,

the problem Edgar was speaking about when applying a highpass filter was
the initial 50 milliseconds of 'ripple' or something, which I can hear
after repeated application of the rectifier plug (I have to rely on
sighted people's feedback about this since I can't see the screen to
check). Do you know what that initial ripple is and how to remove it
without further distortion? Or how to at least reduce it?

David

On Tue, 29 Nov 2005, Sami Jumppanen wrote:

> Hi,
>
> if you guys just understood something about the audio itself instead
> of just witing DSP code... ;)
>
> Firstly, we cannot "center" the audio by the use of min / max peaks.
> What counts, is the "weight balance" of the signal. To find out that,
> we need to calculate the average with n samples. However, we cannot do
> that with a fixed number of samples, nor with the whole audio file -
> because the average varies constantly.
>
> We need to remove the DC component, and DC (or high-pass) filter is
> just the tool for doing that. In real life, we let the audio go
> through a capacitor with suitable resistance to ground to discharge
> the capacitor.
>
> --
> Sami "Some-E" Jumppanen
> [hidden email]
> http://netti.nic.fi/~some-e/
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> <a href="http://ads.osdn.com/?ad_idv37&alloc_id865&opÌk">http://ads.osdn.com/?ad_idv37&alloc_id865&opÌk
> _______________________________________________
> Audacity-nyquist mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/audacity-nyquist
>
>
Reply | Threaded
Open this post in threaded view
|

Re: hp filter was Problem with this plug-in...

Sami Jumppanen
Hi,

I think it relates to the delaying nature of the process. What's
"inside" the filter when the filtering begins? Is it garbage? I think
not. How do we define the initial condition? In real (analogue) life
we practically never face this kind of situation, because the filters
just have been on and work forever. Anyway, hard to say what the
"ripple" means without seeing it. But let's assume this:

The audio has offset so that it's heavily on the positive side. How
does the filter see the beginning of the audio file? It sees a jump
from zero level to highly positive level, like a step. Smoothing out
the step takes time, determined by the cutoff frequency of the filter.
During that time the resulting DC level "crawls" towards zero. Test
it. Create a moment of silence in the middle of this kind of audio,
HP-filter the audio and see if the effect is the same as you see in
the beginning.

Also, crossed my mind that high-order filter algorithms can "bounce".
If we need low delay and steep cutoff, I think that's an obvious side
effect.

Sami


On 29/11/05, David R. Sky <[hidden email]> wrote:

> Hi Sami,
>
> the problem Edgar was speaking about when applying a highpass filter was
> the initial 50 milliseconds of 'ripple' or something, which I can hear
> after repeated application of the rectifier plug (I have to rely on
> sighted people's feedback about this since I can't see the screen to
> check). Do you know what that initial ripple is and how to remove it
> without further distortion? Or how to at least reduce it?
>
> David


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
<a href="http://ads.osdn.com/?ad_idv37&alloc_id865&op=click">http://ads.osdn.com/?ad_idv37&alloc_id865&op=click
_______________________________________________
Audacity-nyquist mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/audacity-nyquist