Hi everyone
I m B-tech final year student
i m trying to implement graphical equalizer (real time) with TMS320C6711 kit.
(Final Year Project)
I wanna know that
1)Is there any way to introduce adaptive signal processing to it?
2)also ,is there any standard algo to be followed for the task?
I will really appreciate any kind of inputs in this regard
Thank You
Help Needed Regarding Project Graphic Equalizer
Started by ●December 31, 2006
Reply by ●January 3, 20072007-01-03
The graphic equalizer takes a FFT, and divides the transform into different
frequency bins. It then applies different levels of gain
to the different individual frequency bands, for example if you increase the
gain on the lower frequencies, you will be boosting the bass.
you have to take care that overall gains that you do by multiplying the
individual frequency bands are normalized so that overall energy does
not change. You also have to apply a window in the time domain before you
take the FFT to get better frequency resolution at the expense
of trading off amplitude accuracy.
Regds
JS
_____
From: c... [mailto:c...] On Behalf Of
t...@yahoo.co.in
Sent: Friday, December 29, 2006 3:27 PM
To: c...
Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
Hi everyone
I m B-tech final year student
i m trying to implement graphical equalizer (real time) with TMS320C6711
kit. (Final Year Project)
I wanna know that
1)Is there any way to introduce adaptive signal processing to it?
2)also ,is there any standard algo to be followed for the task?
I will really appreciate any kind of inputs in this regard
Thank You
frequency bins. It then applies different levels of gain
to the different individual frequency bands, for example if you increase the
gain on the lower frequencies, you will be boosting the bass.
you have to take care that overall gains that you do by multiplying the
individual frequency bands are normalized so that overall energy does
not change. You also have to apply a window in the time domain before you
take the FFT to get better frequency resolution at the expense
of trading off amplitude accuracy.
Regds
JS
_____
From: c... [mailto:c...] On Behalf Of
t...@yahoo.co.in
Sent: Friday, December 29, 2006 3:27 PM
To: c...
Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
Hi everyone
I m B-tech final year student
i m trying to implement graphical equalizer (real time) with TMS320C6711
kit. (Final Year Project)
I wanna know that
1)Is there any way to introduce adaptive signal processing to it?
2)also ,is there any standard algo to be followed for the task?
I will really appreciate any kind of inputs in this regard
Thank You
Reply by ●January 6, 20072007-01-06
The graphic equalizer takes a FFT, and divides the transform into different
>frequency bins. It then applies different levels of gain
>
>to the different individual frequency bands, for example if you increase the
>gain on the lower frequencies, you will be boosting the bass.
>
>you have to take care that overall gains that you do by multiplying the
>individual frequency bands are normalized so that overall energy does
>
>not change. You also have to apply a window in the time domain before you
>take the FFT to get better frequency resolution at the expense
>
>of trading off amplitude accuracy.
>
>Regds
>
>JS
Hello Everyone
Thanks alot for replying JS
i would also appreciate answers for the questions i posted.
i am trying to explore the possibilities of adaptive signal processing here.
>frequency bins. It then applies different levels of gain
>
>to the different individual frequency bands, for example if you increase the
>gain on the lower frequencies, you will be boosting the bass.
>
>you have to take care that overall gains that you do by multiplying the
>individual frequency bands are normalized so that overall energy does
>
>not change. You also have to apply a window in the time domain before you
>take the FFT to get better frequency resolution at the expense
>
>of trading off amplitude accuracy.
>
>Regds
>
>JS
Hello Everyone
Thanks alot for replying JS
i would also appreciate answers for the questions i posted.
i am trying to explore the possibilities of adaptive signal processing here.
Reply by ●January 9, 20072007-01-09
Hi,
The FFT method would be very complex and needs processing power. You will have take care of sharp attack at end of block and windowing gain compensation. Another disadvantage is frequency resolution.
There are widely known approaches are using peak filter. The main building block of graphic equaliser is All pass PEAK (or shelf ) filter. You can find design details about PEAK filter in the book by S.K Mitra.
There are agian two topologies: one is parallel and another one serial network of Peak filters. In this approach involves transalating the user gain into peak filter coefficient.
In parallel topologies you will have filter running parallel equal to number of bands. finally output of the peak filters is added and scaled.
In serial topologies number of filters will reduce.
Hope this help you.
All the best
Amaresh
sankaran wrote:
The graphic equalizer takes a FFT, and divides the transform into different
frequency bins. It then applies different levels of gain
to the different individual frequency bands, for example if you increase the
gain on the lower frequencies, you will be boosting the bass.
you have to take care that overall gains that you do by multiplying the
individual frequency bands are normalized so that overall energy does
not change. You also have to apply a window in the time domain before you
take the FFT to get better frequency resolution at the expense
of trading off amplitude accuracy.
Regds
JS
_____
From: c... [mailto:c...] On Behalf Of
t...@yahoo.co.in
Sent: Friday, December 29, 2006 3:27 PM
To: c...
Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
Hi everyone
I m B-tech final year student
i m trying to implement graphical equalizer (real time) with TMS320C6711
kit. (Final Year Project)
I wanna know that
1)Is there any way to introduce adaptive signal processing to it?
2)also ,is there any standard algo to be followed for the task?
I will really appreciate any kind of inputs in this regard
Thank You
The FFT method would be very complex and needs processing power. You will have take care of sharp attack at end of block and windowing gain compensation. Another disadvantage is frequency resolution.
There are widely known approaches are using peak filter. The main building block of graphic equaliser is All pass PEAK (or shelf ) filter. You can find design details about PEAK filter in the book by S.K Mitra.
There are agian two topologies: one is parallel and another one serial network of Peak filters. In this approach involves transalating the user gain into peak filter coefficient.
In parallel topologies you will have filter running parallel equal to number of bands. finally output of the peak filters is added and scaled.
In serial topologies number of filters will reduce.
Hope this help you.
All the best
Amaresh
sankaran wrote:
The graphic equalizer takes a FFT, and divides the transform into different
frequency bins. It then applies different levels of gain
to the different individual frequency bands, for example if you increase the
gain on the lower frequencies, you will be boosting the bass.
you have to take care that overall gains that you do by multiplying the
individual frequency bands are normalized so that overall energy does
not change. You also have to apply a window in the time domain before you
take the FFT to get better frequency resolution at the expense
of trading off amplitude accuracy.
Regds
JS
_____
From: c... [mailto:c...] On Behalf Of
t...@yahoo.co.in
Sent: Friday, December 29, 2006 3:27 PM
To: c...
Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
Hi everyone
I m B-tech final year student
i m trying to implement graphical equalizer (real time) with TMS320C6711
kit. (Final Year Project)
I wanna know that
1)Is there any way to introduce adaptive signal processing to it?
2)also ,is there any standard algo to be followed for the task?
I will really appreciate any kind of inputs in this regard
Thank You
Reply by ●January 10, 20072007-01-10
Amaresh-
> The FFT method would be very complex and needs processing power. You
> will have take care of sharp attack at end of block and windowing gain
> compensation. Another disadvantage is frequency resolution.
>
> There are widely known approaches are using peak filter. The main
> building block of graphic equaliser is All pass PEAK (or shelf ) filter.
> You can find design details about PEAK filter in the book by S.K Mitra.
>
> There are agian two topologies: one is parallel and another one serial
> network of Peak filters. In this approach involves transalating the user
> gain into peak filter coefficient.
> In parallel topologies you will have filter running parallel equal to
> number of bands. finally output of the peak filters is added and scaled.
>
> In serial topologies number of filters will reduce.
Very accurate comments. Using 5502 on this card:
http://www.signalogic.com/sigc5502.htm
we implemented the serial topology, using both peak and notch filters. We used
64-bit fixed-point precision which allowed accurate, stable IIR filters down to 5 Hz
at 48 kHz sampling rate (24-bit audio data). We could run 60+ filters in real-time.
-Jeff
> sankaran wrote:
> The graphic equalizer takes a FFT, and divides the transform into different
> frequency bins. It then applies different levels of gain
>
> to the different individual frequency bands, for example if you increase the
> gain on the lower frequencies, you will be boosting the bass.
>
> you have to take care that overall gains that you do by multiplying the
> individual frequency bands are normalized so that overall energy does
>
> not change. You also have to apply a window in the time domain before you
> take the FFT to get better frequency resolution at the expense
>
> of trading off amplitude accuracy.
>
> Regds
>
> JS
>
> _____
>
> From: c... [mailto:c...] On Behalf Of
> t...@yahoo.co.in
> Sent: Friday, December 29, 2006 3:27 PM
> To: c...
> Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
>
> Hi everyone
> I m B-tech final year student
> i m trying to implement graphical equalizer (real time) with TMS320C6711
> kit. (Final Year Project)
> I wanna know that
> 1)Is there any way to introduce adaptive signal processing to it?
> 2)also ,is there any standard algo to be followed for the task?
>
> I will really appreciate any kind of inputs in this regard
> Thank You
> The FFT method would be very complex and needs processing power. You
> will have take care of sharp attack at end of block and windowing gain
> compensation. Another disadvantage is frequency resolution.
>
> There are widely known approaches are using peak filter. The main
> building block of graphic equaliser is All pass PEAK (or shelf ) filter.
> You can find design details about PEAK filter in the book by S.K Mitra.
>
> There are agian two topologies: one is parallel and another one serial
> network of Peak filters. In this approach involves transalating the user
> gain into peak filter coefficient.
> In parallel topologies you will have filter running parallel equal to
> number of bands. finally output of the peak filters is added and scaled.
>
> In serial topologies number of filters will reduce.
Very accurate comments. Using 5502 on this card:
http://www.signalogic.com/sigc5502.htm
we implemented the serial topology, using both peak and notch filters. We used
64-bit fixed-point precision which allowed accurate, stable IIR filters down to 5 Hz
at 48 kHz sampling rate (24-bit audio data). We could run 60+ filters in real-time.
-Jeff
> sankaran wrote:
> The graphic equalizer takes a FFT, and divides the transform into different
> frequency bins. It then applies different levels of gain
>
> to the different individual frequency bands, for example if you increase the
> gain on the lower frequencies, you will be boosting the bass.
>
> you have to take care that overall gains that you do by multiplying the
> individual frequency bands are normalized so that overall energy does
>
> not change. You also have to apply a window in the time domain before you
> take the FFT to get better frequency resolution at the expense
>
> of trading off amplitude accuracy.
>
> Regds
>
> JS
>
> _____
>
> From: c... [mailto:c...] On Behalf Of
> t...@yahoo.co.in
> Sent: Friday, December 29, 2006 3:27 PM
> To: c...
> Subject: [c6x] Help Needed Regarding Project Graphic Equalizer
>
> Hi everyone
> I m B-tech final year student
> i m trying to implement graphical equalizer (real time) with TMS320C6711
> kit. (Final Year Project)
> I wanna know that
> 1)Is there any way to introduce adaptive signal processing to it?
> 2)also ,is there any standard algo to be followed for the task?
>
> I will really appreciate any kind of inputs in this regard
> Thank You
