hello everybody!

I’m searching a way to avoid producing alias frequencies.

I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands

I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.

do someone has some recommendations for this?

Greetings,
Philipp 
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

First of all, I am not sure why you want a HP filter, since aliasing would happen above a certain frequency not
below. I assume you are talking about a LP filter instead.

Having said that, it only makes sense to use an LP filter if you are downsampling the signal at some point,
and for that you can try designing a linear phase FIR with the desired curve. GEN 53 can do this for you

https://csound.com/manual/GEN53.html

and implement it with ftconv.

If however you are not downsampling, LP filtering is not going to help you remove any aliasing.
========================
Prof. Victor Lazzarini
Maynooth University
Ireland

> On 10 Nov 2024, at 17:11, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
>
> *Warning*
>
> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>
> hello everybody!
>
> I’m searching a way to avoid producing alias frequencies.
>
> I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands
>
> I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.
>
> do someone has some recommendations for this?
>
> Greetings,
> Philipp
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here


Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Thank you, Victor, for your response!
I definitely meant LP; I’m not sure how that mistake slipped in.
I’m not certain if downsampling is actually necessary in my case, so perhaps it would help to explain my idea in more detail:
I’ve recorded some audio files containing ultrasonic frequencies. Now, I’d like to remove all audible frequencies with a very steep high-pass filter. After this, I plan to apply amplitude modulation to make the ultrasonic content audible, without changing the playback speed or pitch of the sound — only through the lower sideband generated by the AM.
Since AM also produces a higher sideband, I’d like to filter this out to prevent any aliasing.
So, I think I need both a very steep low-pass and high-pass filter.
Any thoughts?

> Am 10.11.2024 um 18:28 schrieb Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE>:
> 
> First of all, I am not sure why you want a HP filter, since aliasing would happen above a certain frequency not
> below. I assume you are talking about a LP filter instead.
> 
> Having said that, it only makes sense to use an LP filter if you are downsampling the signal at some point,
> and for that you can try designing a linear phase FIR with the desired curve. GEN 53 can do this for you
> 
> https://csound.com/manual/GEN53.html
> 
> and implement it with ftconv.
> 
> If however you are not downsampling, LP filtering is not going to help you remove any aliasing.
> ========================
> Prof. Victor Lazzarini
> Maynooth University
> Ireland
> 
>> On 10 Nov 2024, at 17:11, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
>> 
>> *Warning*
>> 
>> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>> 
>> hello everybody!
>> 
>> I’m searching a way to avoid producing alias frequencies.
>> 
>> I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands
>> 
>> I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.
>> 
>> do someone has some recommendations for this?
>> 
>> Greetings,
>> Philipp
>> Csound mailing list
>> Csound@listserv.heanet.ie
>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>> Send bugs reports to
>>       https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
> 
> 
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

I guess GEN53 is the best way to go for designing the filters that meet your specs.
========================
Prof. Victor Lazzarini
Maynooth University
Ireland

> On 11 Nov 2024, at 07:21, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
>
> Thank you, Victor, for your response!
> I definitely meant LP; I’m not sure how that mistake slipped in.
> I’m not certain if downsampling is actually necessary in my case, so perhaps it would help to explain my idea in more detail:
> I’ve recorded some audio files containing ultrasonic frequencies. Now, I’d like to remove all audible frequencies with a very steep high-pass filter. After this, I plan to apply amplitude modulation to make the ultrasonic content audible, without changing the playback speed or pitch of the sound — only through the lower sideband generated by the AM.
> Since AM also produces a higher sideband, I’d like to filter this out to prevent any aliasing.
> So, I think I need both a very steep low-pass and high-pass filter.
> Any thoughts?
>
>> Am 10.11.2024 um 18:28 schrieb Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE>:
>>
>> First of all, I am not sure why you want a HP filter, since aliasing would happen above a certain frequency not
>> below. I assume you are talking about a LP filter instead.
>>
>> Having said that, it only makes sense to use an LP filter if you are downsampling the signal at some point,
>> and for that you can try designing a linear phase FIR with the desired curve. GEN 53 can do this for you
>>
>> https://csound.com/manual/GEN53.html
>>
>> and implement it with ftconv.
>>
>> If however you are not downsampling, LP filtering is not going to help you remove any aliasing.
>> ========================
>> Prof. Victor Lazzarini
>> Maynooth University
>> Ireland
>>
>>> On 10 Nov 2024, at 17:11, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
>>>
>>> *Warning*
>>>
>>> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
>>>
>>> hello everybody!
>>>
>>> I’m searching a way to avoid producing alias frequencies.
>>>
>>> I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands
>>>
>>> I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.
>>>
>>> do someone has some recommendations for this?
>>>
>>> Greetings,
>>> Philipp
>>> Csound mailing list
>>> Csound@listserv.heanet.ie
>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>> Send bugs reports to
>>>      https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>>
>>
>> Csound mailing list
>> Csound@listserv.heanet.ie
>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>> Send bugs reports to
>>       https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
>
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here


Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Thanks Oeyvind! 
I tried it out and it sounds very interesting!
This is the Bode Frequncy Shifter, right?

> Am 11.11.2024 um 08:56 schrieb Oeyvind Brandtsegg :
> 
>  aCar oscili 1, icps ; replace this with your recording
> 
> ; single sideband modulation
>  aSin, aCos hilbert aCar
>  aModSin oscili 1, imodFreq, giSine, 0.0
>  aModCos oscili 1, imodFreq, giSine, 0.25
>  aMod1 = aSin * aModCos
>  aMod2 = aCos * aModSin
>  aSum = aMod1 + aMod2
>  aDiff = aMod1 - aMod2
> 


@Victor:
are there any ressources to read which explains these convolution filters? i never worked with them and never heard of them.
GEN53 seem a little opaq to me.
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

I think it does mye or less the same as the Bode frequency shifter, AFAIK that is an analog module.  One bonus feature by doing it like we do here is that you can use any signal as the modulator.

Thanks Partev for the idea!
Sounds like a good solution!

- Philipp

> Am 17.11.2024 um 22:45 schrieb Partev Sarkissian <0000060b2ef1338e-dmarc-request@LISTSERV.HEANET.IE>:
> 
> 
> Also, I saw mention of wanting a steeper roll-off rate. As I recall, many Csound 
> filters are second-order filters, would typically have 6dB/oct roll-off. Cascading 
> a filter into another adds an emphasis prior to cut-off, and a steeper roll-off rate. 
> 
> I've cascaded filters that way and attained 18-24dB/oct roll-off rates. I was using
>  that when doing FM and wanted to filter select groups of side-bands. Also used 
> that method on a single-side FM from an AES paper Victor did,... which I loved. 
> Did straight as in the paper, then experimented on it with my 4th and 8th order filters. 
> Two audio paths, one was upper, the other lower side band, and used 4th and 8th 
> order cascaded filters on the different parts the the upper and lower side-bands. 
> Interesting two channel effects in the headphones. 
> 
> If you find that useful, it could probably do something similar in AM. 
> 
> 
> 
> -Partev 
> 
> 
> On Monday, November 11, 2024 at 07:57:30 AM GMT, Oeyvind Brandtsegg  wrote: 
> 
> 
> ... you can then listen to the aSum or aDiff outputs, each containing the separate sideband
> 
> man. 11. nov. 2024 kl. 08:56 skrev Oeyvind Brandtsegg :
> Hi Philipp
> 
> Maybe you'd like to try single sideband modulation? There, you can get the upper and lower sidebands as separate outputs.
> Like this:
> 
>  aCar oscili 1, icps ; replace this with your recording
> 
> ; single sideband modulation
>  aSin, aCos hilbert aCar
>  aModSin oscili 1, imodFreq, giSine, 0.0
>  aModCos oscili 1, imodFreq, giSine, 0.25
>  aMod1 = aSin * aModCos
>  aMod2 = aCos * aModSin
>  aSum = aMod1 + aMod2
>  aDiff = aMod1 - aMod2
> 
> all best
> Øyvind
> 
> man. 11. nov. 2024 kl. 08:21 skrev Philipp Neumann <0000119f78f3a4f9-dmarc-request@listserv.heanet.ie>:
> Thank you, Victor, for your response!
> I definitely meant LP; I’m not sure how that mistake slipped in.
> I’m not certain if downsampling is actually necessary in my case, so perhaps it would help to explain my idea in more detail:
> I’ve recorded some audio files containing ultrasonic frequencies. Now, I’d like to remove all audible frequencies with a very steep high-pass filter. After this, I plan to apply amplitude modulation to make the ultrasonic content audible, without changing the playback speed or pitch of the sound — only through the lower sideband generated by the AM.
> Since AM also produces a higher sideband, I’d like to filter this out to prevent any aliasing.
> So, I think I need both a very steep low-pass and high-pass filter.
> Any thoughts?
> 
> > Am 10.11.2024 um 18:28 schrieb Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE>:
> > 
> > First of all, I am not sure why you want a HP filter, since aliasing would happen above a certain frequency not
> > below. I assume you are talking about a LP filter instead.
> > 
> > Having said that, it only makes sense to use an LP filter if you are downsampling the signal at some point,
> > and for that you can try designing a linear phase FIR with the desired curve. GEN 53 can do this for you
> > 
> > https://csound.com/manual/GEN53.html
> > 
> > and implement it with ftconv.
> > 
> > If however you are not downsampling, LP filtering is not going to help you remove any aliasing.
> > ========================
> > Prof. Victor Lazzarini
> > Maynooth University
> > Ireland
> > 
> >> On 10 Nov 2024, at 17:11, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
> >> 
> >> *Warning*
> >> 
> >> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
> >> 
> >> hello everybody!
> >> 
> >> I’m searching a way to avoid producing alias frequencies.
> >> 
> >> I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands
> >> 
> >> I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.
> >> 
> >> do someone has some recommendations for this?
> >> 
> >> Greetings,
> >> Philipp
> >> Csound mailing list
> >> Csound@listserv.heanet.ie
> >> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> >> Send bugs reports to
> >>       https://github.com/csound/csound/issues
> >> Discussions of bugs and features can be posted here
> > 
> > 
> > Csound mailing list
> > Csound@listserv.heanet.ie
> > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> > Send bugs reports to
> >        https://github.com/csound/csound/issues
> > Discussions of bugs and features can be posted here
> 
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here
> Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here 
> Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

since i’m not a member of AES i can’t see how to get this paper 🤔

> Am 19.11.2024 um 20:41 schrieb Partev Sarkissian <0000060b2ef1338e-dmarc-request@LISTSERV.HEANET.IE>:
> 
> 
> Also check out Victor's paper in the Journal of the Audio Engineering Society 
> on single side band FM. 
> 
> 
> -Partev 
> 
> 
> 
> On Monday, November 18, 2024 at 11:10:46 PM PST, Philipp Neumann <0000119f78f3a4f9-dmarc-request@listserv.heanet.ie> wrote: 
> 
> 
> Thanks Partev for the idea!
> Sounds like a good solution!
> 
> - Philipp
> 
> > Am 17.11.2024 um 22:45 schrieb Partev Sarkissian <0000060b2ef1338e-dmarc-request@LISTSERV.HEANET.IE>:
> > 
> > 
> > Also, I saw mention of wanting a steeper roll-off rate. As I recall, many Csound 
> > filters are second-order filters, would typically have 6dB/oct roll-off. Cascading 
> > a filter into another adds an emphasis prior to cut-off, and a steeper roll-off rate. 
> > 
> > I've cascaded filters that way and attained 18-24dB/oct roll-off rates. I was using
> >  that when doing FM and wanted to filter select groups of side-bands. Also used 
> > that method on a single-side FM from an AES paper Victor did,... which I loved. 
> > Did straight as in the paper, then experimented on it with my 4th and 8th order filters. 
> > Two audio paths, one was upper, the other lower side band, and used 4th and 8th 
> > order cascaded filters on the different parts the the upper and lower side-bands. 
> > Interesting two channel effects in the headphones. 
> > 
> > If you find that useful, it could probably do something similar in AM. 
> > 
> > 
> > 
> > -Partev 
> > 
> > 
> > On Monday, November 11, 2024 at 07:57:30 AM GMT, Oeyvind Brandtsegg  wrote: 
> > 
> > 
> > ... you can then listen to the aSum or aDiff outputs, each containing the separate sideband
> > 
> > man. 11. nov. 2024 kl. 08:56 skrev Oeyvind Brandtsegg :
> > Hi Philipp
> > 
> > Maybe you'd like to try single sideband modulation? There, you can get the upper and lower sidebands as separate outputs.
> > Like this:
> > 
> >  aCar oscili 1, icps ; replace this with your recording
> > 
> > ; single sideband modulation
> >  aSin, aCos hilbert aCar
> >  aModSin oscili 1, imodFreq, giSine, 0.0
> >  aModCos oscili 1, imodFreq, giSine, 0.25
> >  aMod1 = aSin * aModCos
> >  aMod2 = aCos * aModSin
> >  aSum = aMod1 + aMod2
> >  aDiff = aMod1 - aMod2
> > 
> > all best
> > Øyvind
> > 
> > man. 11. nov. 2024 kl. 08:21 skrev Philipp Neumann <0000119f78f3a4f9-dmarc-request@listserv.heanet.ie>:
> > Thank you, Victor, for your response!
> > I definitely meant LP; I’m not sure how that mistake slipped in.
> > I’m not certain if downsampling is actually necessary in my case, so perhaps it would help to explain my idea in more detail:
> > I’ve recorded some audio files containing ultrasonic frequencies. Now, I’d like to remove all audible frequencies with a very steep high-pass filter. After this, I plan to apply amplitude modulation to make the ultrasonic content audible, without changing the playback speed or pitch of the sound — only through the lower sideband generated by the AM.
> > Since AM also produces a higher sideband, I’d like to filter this out to prevent any aliasing.
> > So, I think I need both a very steep low-pass and high-pass filter.
> > Any thoughts?
> > 
> > > Am 10.11.2024 um 18:28 schrieb Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE>:
> > > 
> > > First of all, I am not sure why you want a HP filter, since aliasing would happen above a certain frequency not
> > > below. I assume you are talking about a LP filter instead.
> > > 
> > > Having said that, it only makes sense to use an LP filter if you are downsampling the signal at some point,
> > > and for that you can try designing a linear phase FIR with the desired curve. GEN 53 can do this for you
> > > 
> > > https://csound.com/manual/GEN53.html
> > > 
> > > and implement it with ftconv.
> > > 
> > > If however you are not downsampling, LP filtering is not going to help you remove any aliasing.
> > > ========================
> > > Prof. Victor Lazzarini
> > > Maynooth University
> > > Ireland
> > > 
> > >> On 10 Nov 2024, at 17:11, Philipp Neumann <0000119f78f3a4f9-dmarc-request@LISTSERV.HEANET.IE> wrote:
> > >> 
> > >> *Warning*
> > >> 
> > >> This email originated from outside of Maynooth University's Mail System. Do not reply, click links or open attachments unless you recognise the sender and know the content is safe.
> > >> 
> > >> hello everybody!
> > >> 
> > >> I’m searching a way to avoid producing alias frequencies.
> > >> 
> > >> I’m working with ultrasonic sounds and do some AM on it. So i need a way to filter out the higher sidebands
> > >> 
> > >> I thought about using a high-pass filter with a very steep slope. but the filter i found are not that steep.
> > >> 
> > >> do someone has some recommendations for this?
> > >> 
> > >> Greetings,
> > >> Philipp
> > >> Csound mailing list
> > >> Csound@listserv.heanet.ie
> > >> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> > >> Send bugs reports to
> > >>      https://github.com/csound/csound/issues
> > >> Discussions of bugs and features can be posted here
> 
> > > 
> > > 
> > > Csound mailing list
> > > Csound@listserv.heanet.ie
> > > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> > > Send bugs reports to
> > >        https://github.com/csound/csound/issues
> > > Discussions of bugs and features can be posted here
> > 
> > Csound mailing list
> > Csound@listserv.heanet.ie
> > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> > Send bugs reports to
> >        https://github.com/csound/csound/issues
> > Discussions of bugs and features can be posted here
> > Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here 
> > Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here
> 
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here
> Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here