You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Clemens Novak <cl...@familie-novak.net> on 2012/08/20 17:00:20 UTC
[math] Convolution
Dear all,
I would like to work on some signal processing functions (as indicated
on the wiki WishList) and started with the convolution of 2 sequences
(represented as RealVector). I have completed a first working version
(some error checking code, formatting, unit tests etc are missing); I am
unsure of how to continue: Is the next step to create a new Jira ticket
and upload my final code there (for further discussion/review/...)?
Thanks for your help & kind regards - Clemens
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [math] Convolution
Posted by Bruce A Johnson <jo...@umbc.edu>.
On Aug 23, 2012, at 8:31 AM, Clemens Novak wrote:
> On 2012-08-20 19:25, Luc Maisonobe wrote:
>> Le 20/08/2012 17:00, Clemens Novak a écrit :
>>> Dear all,
>>
>> Hi Clemens,
>>
>>>
>>> I would like to work on some signal processing functions (as indicated
>>> on the wiki WishList) and started with the convolution of 2 sequences
>>> (represented as RealVector). I have completed a first working version
>>> (some error checking code, formatting, unit tests etc are missing); I am
>>> unsure of how to continue: Is the next step to create a new Jira ticket
>>> and upload my final code there (for further discussion/review/...)?
>>
>> Yes, you can do that.
>>
>>>
>>> Thanks for your help & kind regards - Clemens
>>
>> Thanks for your interest and contribution.
>>
>> Luc
>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>
> Dear all,
>
> besides the ongoing discussions in MATH-851, I would like to discuss some extensions of this library in the area of discrete-time signal processing:
>
> (1) convolution of 1D and 2D sequences based on (i) straightforward calculation of the convolution sum and (ii) using the FFT.
>
> (2) filter functions for 1D sequences, i.e. y[n] = \sum_{l=0}^{L-1} b_l x[n-l] - \sum_{m=1}^{M-1} a_m y[n-m]
>
> (3) window functions (such as raised cosine, blackmann...) used in conjunction with FFTs
>
> (4) possibly some basic filter design (such as window method, Parks-McClellan algorithm...)
>
> However, first I want to ensure that you think these topics belong in Commons math - I have been reading the intro (... is a library of lightweight, self-contained mathematics and statistics components addressing the most common problems...) and I am not perfectly sure whether these topics really fit here?
>
> In MATH-851 we shortly touched whether using RealVector or a simple double array is better. What's your opinion here? The double array might be better in terms of performance; when based on RealVector, however, more basic functionality would be provided out of the box. Has anyone made a performance comparison between arrays and RealVector in order to have some guidelines here?
>
> Kind regards - Clemens
>
I use Commons Math extensively in signal processing applications so I, for one, would love to see more features in this area.
cheers,
Bruce
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [math] Convolution
Posted by Clemens Novak <cl...@familie-novak.net>.
On 2012-08-20 19:25, Luc Maisonobe wrote:
> Le 20/08/2012 17:00, Clemens Novak a écrit :
>> Dear all,
>
> Hi Clemens,
>
>>
>> I would like to work on some signal processing functions (as
>> indicated
>> on the wiki WishList) and started with the convolution of 2
>> sequences
>> (represented as RealVector). I have completed a first working
>> version
>> (some error checking code, formatting, unit tests etc are missing);
>> I am
>> unsure of how to continue: Is the next step to create a new Jira
>> ticket
>> and upload my final code there (for further discussion/review/...)?
>
> Yes, you can do that.
>
>>
>> Thanks for your help & kind regards - Clemens
>
> Thanks for your interest and contribution.
>
> Luc
>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
Dear all,
besides the ongoing discussions in MATH-851, I would like to discuss
some extensions of this library in the area of discrete-time signal
processing:
(1) convolution of 1D and 2D sequences based on (i) straightforward
calculation of the convolution sum and (ii) using the FFT.
(2) filter functions for 1D sequences, i.e. y[n] = \sum_{l=0}^{L-1} b_l
x[n-l] - \sum_{m=1}^{M-1} a_m y[n-m]
(3) window functions (such as raised cosine, blackmann...) used in
conjunction with FFTs
(4) possibly some basic filter design (such as window method,
Parks-McClellan algorithm...)
However, first I want to ensure that you think these topics belong in
Commons math - I have been reading the intro (... is a library of
lightweight, self-contained mathematics and statistics components
addressing the most common problems...) and I am not perfectly sure
whether these topics really fit here?
In MATH-851 we shortly touched whether using RealVector or a simple
double array is better. What's your opinion here? The double array might
be better in terms of performance; when based on RealVector, however,
more basic functionality would be provided out of the box. Has anyone
made a performance comparison between arrays and RealVector in order to
have some guidelines here?
Kind regards - Clemens
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [math] Convolution
Posted by Luc Maisonobe <Lu...@free.fr>.
Le 20/08/2012 17:00, Clemens Novak a écrit :
> Dear all,
Hi Clemens,
>
> I would like to work on some signal processing functions (as indicated
> on the wiki WishList) and started with the convolution of 2 sequences
> (represented as RealVector). I have completed a first working version
> (some error checking code, formatting, unit tests etc are missing); I am
> unsure of how to continue: Is the next step to create a new Jira ticket
> and upload my final code there (for further discussion/review/...)?
Yes, you can do that.
>
> Thanks for your help & kind regards - Clemens
Thanks for your interest and contribution.
Luc
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org