SRRC pulse shaping question

Hello all,

I have a question applying to everybody that have tried transmitting
symbols in different kinds of modulation (BPSK, QPSK, GMSK etc).
I am using the following configuration in GNU Radio ( random source
block
-> bytes to symbols -> SRRC block -> file sink)

The attached figure plots the generated symbols (+1,-1) with red color
along with the upsampled SRRC waveform (blue)
For some reason that I am not able to understand the last 8-bits do not
have a correspondence with the waveform created. In other words
the waveform is “cut” at the last byte.
Can anyone help me with the above.

Thanks,
-George

Have you taken into account the group delay of the filter? How long is
your filter? Try sending zeroes at the end to flush your filter state?

Brian

On Tue, Dec 4, 2012 at 11:35 PM, George S. <

The filter is consisted of 55 taps.
I tried by sending zeros. Though zeros are interpreted in -1’s from the
modulation mapping block.
Moreover, the zeros will be taken into account as symbols to be shaped,
therefore, the same thing will happen.

-George

On 12/5/12 12:07 AM, Brian P. wrote:

For some reason that I am not able to understand the last 8-bits

_______________________________________________
Discuss-gnuradio mailing list
[email protected] <mailto:[email protected]>
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio


Sklivanitis Georgios
Ph.D Student
Communications Division,
Department of Electronics and Electrical Engineering
University at Buffalo, The State University of New York
North Campus, Office 205, Furnas Hall
Buffalo, NY 14260
www.acsu.buffalo.edu/~gsklivan

You’ll need to figure out how to flush the filter state and send actual
0+0j samples through. Maybe you can set a flag and multiply the samples
coming out by 1 or 0 depending if you are flushing or not?

Brian

On Wed, Dec 5, 2012 at 12:11 AM, George S. <

I think this means creating a new block as I cannot come up with a quick
way of implementing the above.
In any case, thanks much.

-George

On 12/5/12 12:27 AM, Brian P. wrote:

Have you taken into account the group delay of the filter?  How
    Hello all,
    8-bits do not have a correspondence with the waveform
Ph.D Student
Communications Division,
Department of Electronics and Electrical Engineering
University at Buffalo, The State University of New York
North Campus, Office 205, Furnas Hall
Buffalo, NY 14260
www.acsu.buffalo.edu/~gsklivan  <http://www.acsu.buffalo.edu/%7Egsklivan>


Sklivanitis Georgios
Ph.D Student
Communications Division,
Department of Electronics and Electrical Engineering
University at Buffalo, The State University of New York
North Campus, Office 205, Furnas Hall
Buffalo, NY 14260
www.acsu.buffalo.edu/~gsklivan