On 07/11/2016 17:09, Evgeny Filatov wrote:
> On 07/11/2016 00:08, abhi5491 wrote:
>> In "Adaptive filter theory" and "Advance signal processing and Noise
>> reduction" they have directly come up with the term gain without stating
>> how they got it.
>>
>> In adaptive filter theory they have jumped from equation 13.16 to
>> 13.18.Same in advance signal processing equation 7.74 to 7.76.
>>
>> https://drive.google.com/drive/folders/0Bz-qIQBWiVixdWlGR0J3d1lELVU
>>
>> ---------------------------------------
>> Posted through http://www.DSPRelated.com
>>
>
> [offtopic]
>
> Haha, and I'm trying to make sense of Eq. 27 in a 1993 paper about a
> certain deconvolution algorithm:
>
> http://ftp.kapitza.ras.ru/people/kosarev/cpc_1993.pdf
>
> Other than that I believe Eq. 27 is formulated with a typo (the fraction
> inside the sum should be multiplied by P_{ij} for it to make sense), I

Apologies, meant to say P_{ik}...

Gene

On 07/11/2016 00:08, abhi5491 wrote:
> In "Adaptive filter theory" and "Advance signal processing and Noise
> reduction" they have directly come up with the term gain without stating
> how they got it.
>
> In adaptive filter theory they have jumped from equation 13.16 to
> 13.18.Same in advance signal processing equation 7.74 to 7.76.
>
> https://drive.google.com/drive/folders/0Bz-qIQBWiVixdWlGR0J3d1lELVU
>
> ---------------------------------------
> Posted through http://www.DSPRelated.com
>

[offtopic]

Haha, and I'm trying to make sense of Eq. 27 in a 1993 paper about a 
certain deconvolution algorithm:

http://ftp.kapitza.ras.ru/people/kosarev/cpc_1993.pdf

Other than that I believe Eq. 27 is formulated with a typo (the fraction 
inside the sum should be multiplied by P_{ij} for it to make sense), I 
wonder what kind of a deconvolution algorithm that is. It's neither 
steepest descent nor Richardson-Lucy, although it DOES look similar to 
the both, and the authors claim their algorithm converges to the maximum 
of the likelihood function.

I would totally love if there were a bit more of a discussion about 
various possible options to do a deconvolution iteration.

Gene

On Monday, November 7, 2016 at 10:08:22 AM UTC+13, abhi5491 wrote:
> In "Adaptive filter theory" and "Advance signal processing and Noise
> reduction" they have directly come up with the term gain without stating
> how they got it.
> 
> In adaptive filter theory they have jumped from equation 13.16 to
> 13.18.Same in advance signal processing equation 7.74 to 7.76.
> 
> https://drive.google.com/drive/folders/0Bz-qIQBWiVixdWlGR0J3d1lELVU
> 
> ---------------------------------------
> Posted through http://www.DSPRelated.com

It's the step-size and nothing more - mu. We engineers tend to mix our terminologies sometimes.

In "Adaptive filter theory" and "Advance signal processing and Noise
reduction" they have directly come up with the term gain without stating
how they got it.

In adaptive filter theory they have jumped from equation 13.16 to
13.18.Same in advance signal processing equation 7.74 to 7.76.

https://drive.google.com/drive/folders/0Bz-qIQBWiVixdWlGR0J3d1lELVU

---------------------------------------
Posted through http://www.DSPRelated.com