The K4's new noise reduction

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

The K4's new noise reduction

wayne burdick
Administrator
We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.

Once the full implementation is complete, we'll post audio samples and additional details.

(In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)

73,
Wayne
N6KR

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

stengrevics
Hi Wayne,

The K3S’ NR algorithm is not operable on digital modes.  Will that be the case on the K4 as well?

73,

John
WA1EAZ

> On May 31, 2019, at 1:11 PM, Wayne Burdick <[hidden email]> wrote:
>
> We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>
> Once the full implementation is complete, we'll post audio samples and additional details.
>
> (In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>
> 73,
> Wayne
> N6KR
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Barry-2
In reply to this post by wayne burdick
Wayne,
     Will this be implimented for the K3/K3s also??

73
Barry
K3NDM

------ Original Message ------
From: "Wayne Burdick" <[hidden email]>
To: "Elecraft Reflector" <[hidden email]>
Cc: [hidden email]
Sent: 5/31/2019 1:11:23 PM
Subject: [Elecraft] The K4's new noise reduction

>We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>
>Once the full implementation is complete, we'll post audio samples and additional details.
>
>(In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>
>73,
>Wayne
>N6KR
>
>______________________________________________________________
>Elecraft mailing list
>Home: http://mailman.qth.net/mailman/listinfo/elecraft
>Help: http://mailman.qth.net/mmfaq.htm
>Post: mailto:[hidden email]
>
>This list hosted by: http://www.qsl.net
>Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Drew AF2Z
In reply to this post by wayne burdick
That sounds promising! A side-by-side comparison of the same noise
samples on a K3 would be interesting as well.

73,
Drew
AF2Z


On 05/31/19 13:11, Wayne Burdick wrote:

> We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>
> Once the full implementation is complete, we'll post audio samples and additional details.
>
> (In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>
> 73,
> Wayne
> N6KR
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Oliver Dröse
In reply to this post by Barry-2
Didn't Wayne answer your question with his original mail already? Hint:
Read his last sentence. ;-)

73, Olli

Contest, DX & radio projects: https://www.dh8bqa.de/


Am 31.05.2019 um 19:27 schrieb Barry:

> Wayne,
>     Will this be implimented for the K3/K3s also??
>
> 73
> Barry
> K3NDM
>
> ------ Original Message ------
> From: "Wayne Burdick" <[hidden email]>
> To: "Elecraft Reflector" <[hidden email]>
> Cc: [hidden email]
> Sent: 5/31/2019 1:11:23 PM
> Subject: [Elecraft] The K4's new noise reduction
>
>> We've recently started testing a new DSP NR (noise reduction)
>> algorithm for the K4. Our early testers are saying it's as good or
>> better than anything else they've used, removing stochastic noise
>> without the characteristic "hollow" sound of earlier algorithms. In
>> my own tests, I'm finding weak SSB signals I simply couldn't copy
>> without it. It goes beyond just reducing listening fatigue. This type
>> of NR operates in the frequency domain, like spectral subtraction,
>> but without the usual artifacts (musical tones) or long sampling delays.
>>
>> Once the full implementation is complete, we'll post audio samples
>> and additional details.
>>
>> (In answer to the inevitable question:  Unfortunately, we won't be
>> able to port this to the K3/K3S. The new algorithm requires the new,
>> much faster DSP chip we're using in the K4.)
>>
>> 73,
>> Wayne
>> N6KR
>>
>> ______________________________________________________________
>> Elecraft mailing list
>> Home: http://mailman.qth.net/mailman/listinfo/elecraft
>> Help: http://mailman.qth.net/mmfaq.htm
>> Post: mailto:[hidden email]
>>
>> This list hosted by: http://www.qsl.net
>> Please help support this email list: http://www.qsl.net/donate.html
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
> ---
> Diese E-Mail wurde von AVG auf Viren geprüft.
> http://www.avg.com
>
>
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Drew AF2Z
In reply to this post by wayne burdick
I'm fairly happy with my K3's NR/NB. I could always use more and better
NR, or course, but a major negative for me is the long time it takes for
the filter to rebuild when changing NR settings or WIDTH. I hope the K4
will be speedier...

73,
Drew
AF2Z



On 05/31/19 13:11, Wayne Burdick wrote:

> We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>
> Once the full implementation is complete, we'll post audio samples and additional details.
>
> (In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>
> 73,
> Wayne
> N6KR
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Barry-2
In reply to this post by Oliver Dröse
Olli,
     You are correct. He did give the answer. Mea culpa. Somehow, I
missed that.

73,
Barry
K3NDM

------ Original Message ------
From: "Oliver Dröse" <[hidden email]>
To: [hidden email]
Sent: 5/31/2019 1:35:02 PM
Subject: Re: [Elecraft] The K4's new noise reduction

>Didn't Wayne answer your question with his original mail already? Hint: Read his last sentence. ;-)
>
>73, Olli
>
>Contest, DX & radio projects: https://www.dh8bqa.de/
>
>
>Am 31.05.2019 um 19:27 schrieb Barry:
>>Wayne,
>>     Will this be implimented for the K3/K3s also??
>>
>>73
>>Barry
>>K3NDM
>>
>>------ Original Message ------
>>From: "Wayne Burdick" <[hidden email]>
>>To: "Elecraft Reflector" <[hidden email]>
>>Cc: [hidden email]
>>Sent: 5/31/2019 1:11:23 PM
>>Subject: [Elecraft] The K4's new noise reduction
>>
>>>We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>>>
>>>Once the full implementation is complete, we'll post audio samples and additional details.
>>>
>>>(In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>>>
>>>73,
>>>Wayne
>>>N6KR
>>>
>>>______________________________________________________________
>>>Elecraft mailing list
>>>Home: http://mailman.qth.net/mailman/listinfo/elecraft
>>>Help: http://mailman.qth.net/mmfaq.htm
>>>Post: mailto:[hidden email]
>>>
>>>This list hosted by: http://www.qsl.net
>>>Please help support this email list: http://www.qsl.net/donate.html
>>
>>______________________________________________________________
>>Elecraft mailing list
>>Home: http://mailman.qth.net/mailman/listinfo/elecraft
>>Help: http://mailman.qth.net/mmfaq.htm
>>Post: mailto:[hidden email]
>>
>>This list hosted by: http://www.qsl.net
>>Please help support this email list: http://www.qsl.net/donate.html
>>
>>---
>>Diese E-Mail wurde von AVG auf Viren geprüft.
>>http://www.avg.com
>>
>>
>______________________________________________________________
>Elecraft mailing list
>Home: http://mailman.qth.net/mailman/listinfo/elecraft
>Help: http://mailman.qth.net/mmfaq.htm
>Post: mailto:[hidden email]
>
>This list hosted by: http://www.qsl.net
>Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

Edward R Cole
In reply to this post by wayne burdick
Shoot,  That my just be the tipping point for me toward  the K4.

73, Ed - KL7UW


From: Wayne Burdick <[hidden email]>
To: Elecraft Reflector <[hidden email]>
Cc: [hidden email]
Subject: [Elecraft] The K4's new noise reduction
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=us-ascii

We've recently started testing a new DSP NR (noise reduction)
algorithm for the K4. Our early testers are saying it's as good or
better than anything else they've used, removing stochastic noise
without the characteristic "hollow" sound of earlier algorithms. In
my own tests, I'm finding weak SSB signals I simply couldn't copy
without it. It goes beyond just reducing listening fatigue. This type
of NR operates in the frequency domain, like spectral subtraction,
but without the usual artifacts (musical tones) or long sampling delays.

Once the full implementation is complete, we'll post audio samples
and additional details.

(In answer to the inevitable question:  Unfortunately, we won't be
able to port this to the K3/K3S. The new algorithm requires the new,
much faster DSP chip we're using in the K4.)

73,
Wayne
N6KR


73, Ed - KL7UW
   http://www.kl7uw.com
Dubus-NA Business mail:
   [hidden email]

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

David Woolley (E.L)
In reply to this post by stengrevics
I would think that, at best, there would be no  benefit over using a
fixed filter, optimised for the mode, and at worst, they would
completely break the mode.

The ideal digital mode signal looks like noise!  COFDM as used for ADSL,
digital TV, digital radio, and 4G mobile phones gets quite close to that
ideal.

--
David Woolley
Owner K2 06123

On 31/05/2019 18:13, John Stengrevics wrote:
>
> The K3S’ NR algorithm is not operable on digital modes.  Will that be the case on the K4 as well?

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

David Woolley (E.L)
There is a good reason why cross-posting email is a bad idea.  This
failed on [hidden email]  Could anyone replying please use just
the standard elecraft list.

On 31/05/2019 21:22, David Woolley wrote:
> I would think that, at best, there would be no  benefit over using a
> fixed filter, optimised for the mode, and at worst, they would
> completely break the mode.
>
> The ideal digital mode signal looks like noise!  COFDM as used for ADSL,
> digital TV, digital radio, and 4G mobile phones gets quite close to that
> ideal.
>


______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Reply | Threaded
Open this post in threaded view
|

Re: The K4's new noise reduction

n6wg
In reply to this post by wayne burdick
Thanks for the NR news Wayne.  I love working 160m, and I expect the new NR will be a welcome new tool.  Really eager to hear some examples of it at work.
73, Bob N6WG

> On May 31, 2019 at 10:11 AM Wayne Burdick <[hidden email]> wrote:
>
>
> We've recently started testing a new DSP NR (noise reduction) algorithm for the K4. Our early testers are saying it's as good or better than anything else they've used, removing stochastic noise without the characteristic "hollow" sound of earlier algorithms. In my own tests, I'm finding weak SSB signals I simply couldn't copy without it. It goes beyond just reducing listening fatigue. This type of NR operates in the frequency domain, like spectral subtraction, but without the usual artifacts (musical tones) or long sampling delays.
>
> Once the full implementation is complete, we'll post audio samples and additional details.
>
> (In answer to the inevitable question:  Unfortunately, we won't be able to port this to the K3/K3S. The new algorithm requires the new, much faster DSP chip we're using in the K4.)
>
> 73,
> Wayne
> N6KR
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html