Elecraft

Noise Reduction

Classic

List

Threaded

6 messages Options

cheng076

Noise Reduction

I have been involved with a project developed by Peter Rhodes of the UK that started as a way to improve the NR in his Pick a Star trx. It appears that the project is nearing the limits of what can be done and will hopefully go public soon. The project is called nitPic and operates on the audio output (phones) under Peter's software. It is truly amazing what it can do. Briefly it has two modes, SSB and CW. It processes the audio in 'bins' or slices. As each bin is processed the SW decides if it is signal or noise; if signal it is enhanced and if noise it is reduced. Sounds simple. It almost completely eliminates a CW signal if in SSB mode and vice versa eliminates a SSB signal if in CW mode. I have been able to bring out CW signals from the noise floor that do not show the slightest trace on my PowerSDR waterfall/spectrum display.

So good things are happening and improvements are in the works.

PJH, N7PXY
______________________________________________________________
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
Message delivered to [hidden email]

k4vd

Re: Noise Reduction

I'd love to hear some A/B audio samples of nitPic if they become available.
It operates at AF? Operates well in congested bands?

I've always been told that by the time the signal is demodulated to AF the
damage is done. Why I believe this I have no clue. So when I hear something
about noise reduction or filtering on AF it always makes me think of
howling CW and scratchy voice. The best filter I used at audio was the old
Autek QF-1A. Boy could I get that howling! I think I need to catch up and
learn new old ham tales.

73,
Kev K4VD

On Wed, Aug 23, 2017 at 1:24 PM, P.J.Hicks <[hidden email]> wrote:

> I have been involved with a project developed by Peter Rhodes of the UK
> that started as a way to improve the NR in his Pick a Star trx. It appears
> that the project is nearing the limits of what can be done and will
> hopefully go public soon. The project is called nitPic and operates on the
> audio output (phones) under Peter's software. It is truly amazing what it
> can do. Briefly it has two modes, SSB and CW. It processes the audio in
> 'bins' or slices. As each bin is processed the SW decides if it is signal
> or noise; if signal it is enhanced and if noise it is reduced. Sounds
> simple. It almost completely eliminates a CW signal if in SSB mode and vice
> versa eliminates a SSB signal if in CW mode. I have been able to bring out
> CW signals from the noise floor that do not show the slightest trace on my
> PowerSDR waterfall/spectrum display.
>
> So good things are happening and improvements are in the works.
>
> PJH, N7PXY
> ______________________________________________________________
> 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
> Message delivered to [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
Message delivered to [hidden email]

Scott-2

Re: Noise Reduction

In reply to this post by cheng076

Is there a website or software repository yet?

Scott
AD5HS

On 8/23/2017 12:24 PM, P.J.Hicks wrote:

> I have been involved with a project developed by Peter Rhodes of the UK that started as a way to improve the NR in his Pick a Star trx. It appears that the project is nearing the limits of what can be done and will hopefully go public soon. The project is called nitPic and operates on the audio output (phones) under Peter's software. It is truly amazing what it can do. Briefly it has two modes, SSB and CW. It processes the audio in 'bins' or slices. As each bin is processed the SW decides if it is signal or noise; if signal it is enhanced and if noise it is reduced. Sounds simple. It almost completely eliminates a CW signal if in SSB mode and vice versa eliminates a SSB signal if in CW mode. I have been able to bring out CW signals from the noise floor that do not show the slightest trace on my PowerSDR waterfall/spectrum display.
>
> So good things are happening and improvements are in the works.
>
> PJH, N7PXY
> ______________________________________________________________
> 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
> Message delivered to [hidden email]
>

k6dgw

Re: Noise Reduction

Doesn't appear to say very much. Other than the BW difference [RF/IF BW
might be more than ultimate AF BW], and provided that the demodulated
audio is free of IMD and other obnoxious distortion, I don't see why NR
at AF wouldn't be essentially the same as at IF. Obvious differences in
the mechanisms to do it, and I'm sure there are dozens of nits that will
be picked over this, but if you can find the signal in the noise [or the
noise obscuring the signal] at the IF frequency, I can't see any reason
why you can't do it at AF just as well.

73,

Fred ["Skip"] K6DGW
Sparks NV DM09dn
Washoe County

On 8/23/2017 4:06 PM, Walter Underwood wrote:
> Discussion appears to be here.
>
> https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t=directory&ch=web&pub=groups&sec=dir&slk=363 <https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t=directory&ch=web&pub=groups&sec=dir&slk=363>
>
> wunder
> K6WRU
> Walter Underwood
> CM87wj
> http://observer.wunderwood.org/ (my blog)

______________________________________________________________
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
Message delivered to [hidden email]

k4vd

Re: Noise Reduction

Maybe the additional noise in the stages between where the filter would be
in IF and where it is in AF? I know I'm thinking of multi-stage RF/IF/AF
superhets and that might be what is tripping me up when it comes to a
single conversion SDR capability. Maybe there is no difference? Or maybe
components that make up a filter in an IF stage are more effective (higher
Q) than in an audio stage?

73,
Kev K4VD

On Wed, Aug 23, 2017 at 7:58 PM, Fred Jensen <[hidden email]> wrote:

> Doesn't appear to say very much. Other than the BW difference [RF/IF BW
> might be more than ultimate AF BW], and provided that the demodulated audio
> is free of IMD and other obnoxious distortion, I don't see why NR at AF
> wouldn't be essentially the same as at IF. Obvious differences in the
> mechanisms to do it, and I'm sure there are dozens of nits that will be
> picked over this, but if you can find the signal in the noise [or the noise
> obscuring the signal] at the IF frequency, I can't see any reason why you
> can't do it at AF just as well.
>
> 73,
>
> Fred ["Skip"] K6DGW
> Sparks NV DM09dn
> Washoe County
>
> On 8/23/2017 4:06 PM, Walter Underwood wrote:
>
>> Discussion appears to be here.
>>
>> https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t=
>> directory&ch=web&pub=groups&sec=dir&slk=363 <
>> https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t
>> =directory&ch=web&pub=groups&sec=dir&slk=363>
>>
>> wunder
>> K6WRU
>> Walter Underwood
>> CM87wj
>> http://observer.wunderwood.org/ (my blog)
>>
>
> ______________________________________________________________
> 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
> Message delivered to [hidden email]
>

K9MA

Re: Noise Reduction

Impulse noise gets stretched out in time when passed through a narrow
filter, so it must be removed ahead of the narrow filter. This is true
whether in hardware or software. Since this is a very common type of
man-made noise, effectively removing or reducing it would be very
useful. I can't see how this could possibly be done in the audio
range. It could, perhaps, at baseband in a direct sampling receiver.

In any case, a future Elecraft radio which could effectively reduce
impulse noise when a band is full of strong signals would be very
attractive, especially to urban hams. One scheme used long ago was to
use a second receiver, tuned a bit out of the band to a relatively clear
segment, to control a noise blanker in the main receiver. Could
Elecraft possibly add a mode to use the sub-receiver in this manner?
Or, because the noise-sensing receiver doesn't have to have great
selectivity, perhaps a dedicated receiver could be offered.

73,

Scott K9MA

On 8/23/2017 19:20, Kevin der Kinderen wrote:

> Maybe the additional noise in the stages between where the filter would be
> in IF and where it is in AF? I know I'm thinking of multi-stage RF/IF/AF
> superhets and that might be what is tripping me up when it comes to a
> single conversion SDR capability. Maybe there is no difference? Or maybe
> components that make up a filter in an IF stage are more effective (higher
> Q) than in an audio stage?
>
> 73,
> Kev K4VD
>
> On Wed, Aug 23, 2017 at 7:58 PM, Fred Jensen <[hidden email]> wrote:
>
>> Doesn't appear to say very much. Other than the BW difference [RF/IF BW
>> might be more than ultimate AF BW], and provided that the demodulated audio
>> is free of IMD and other obnoxious distortion, I don't see why NR at AF
>> wouldn't be essentially the same as at IF. Obvious differences in the
>> mechanisms to do it, and I'm sure there are dozens of nits that will be
>> picked over this, but if you can find the signal in the noise [or the noise
>> obscuring the signal] at the IF frequency, I can't see any reason why you
>> can't do it at AF just as well.
>>
>> 73,
>>
>> Fred ["Skip"] K6DGW
>> Sparks NV DM09dn
>> Washoe County
>>
>> On 8/23/2017 4:06 PM, Walter Underwood wrote:
>>
>>> Discussion appears to be here.
>>>
>>> https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t=
>>> directory&ch=web&pub=groups&sec=dir&slk=363 <
>>> https://uk.groups.yahoo.com/neo/groups/SNR_group/info?v=1&t
>>> =directory&ch=web&pub=groups&sec=dir&slk=363>
>>>
>>> wunder
>>> K6WRU
>>> Walter Underwood
>>> CM87wj
>>> http://observer.wunderwood.org/ (my blog)
>>>
>> ______________________________________________________________
>> 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
>> Message delivered to [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
> Message delivered to [hidden email]

--
Scott K9MA

[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
Message delivered to [hidden email]