Yes, I know. Speaking from the standpoint of design flow, it would
be helpful if the user doesn't need to touch those lines of code if
they'd like to take advantage of the filter designer.
For example, if you could output the co-efficients to a file, and
your code calls:
coeffs = get_coeffs_from_file(my_coeffs_file.csv)
you'd be able to take advantage of the analytical features of the
filter designer without modifying your code. And lets say someday
you had a FIR filter hardware. A custom script might generate a
file of a specified format, and the device driver could pick up the
file and poke the coefficients. In short, having the ability to
call a custom script allows the user to produce a file of arbitrary
format, or do any other arbitrary operation with the coeffecients.
-John
On 06/26/2012 10:27 AM, address@hidden wrote:
John:
Well, there's already gr.firdes, which can be called directly
from python, and produce coefficients that can be easily
transformed into whatever you need.
Granted, that's only for relatively-simple filters.
from gnuradio import gr
coeffs=gr.firdes.low_pass(1.0,samp_rate,corner_freq,transition_width,widnow,beta)
And then do whatever you want with the resulting coefffs (like
scaling to fixed-point 16-bit values, for example)
-Marcus
On 26 Jun 2012 13:21, John Malsbury wrote:
That
is pretty awesome, Sreeraj. Is the output of the program a set
of co-efficients to be copy-pasted, a compilable block, or
something else? It might be useful to have a feature where the
user can call a custom python script with the co-efficients
passed as arguments. This would allow users to easily generate
formatted files, like those used by Xilinx coregen or some
custom format that might be poked to the FPGA, for example. It
would be neat to make some updates to GRC to allow the user to
open the utility and re-generate co-efficients as they do in
MATLAB.
This is making me think of some other potential projects, like a
similar 'digital modulation designer'. For example, we might
use a single block for all modulation types, select the
modulation type and parameters, run simulations to determine BER
vs. Eb/N0, theoretical discrete constellations, ACPR, PAR, etc.
Or perhaps, a "packet designer" where you can graphically
configure FEC, interleaving, sync words, etc. Obviously, the
complexity on this one could grow real fast.
Just ideas... Hopefully you'll have a chance to present your
approach to this at the GNU Radio conference.
-John
On 06/26/2012 09:35 AM, Ben Hilburn wrote:
Sreeraj
-
I just wanted to echo Tom's thoughts that this look
really, really great! I can't wait to see the final
version! This will really be a significant contribution to
the capabilities of GNU Radio, in my opinion.
Cheers,
Ben
On Sun, Jun 24, 2012 at 7:05 AM, sreeraj r <address@hidden>
wrote:
Hi all,
During the initial phase of SoC, I was
concentrating on improving the GUI for filter
designing. A brief list of some of the major
features added to the filter design tool
(gr_filter_design) is given below
--Support for multiple views (grid and
old tabbed view)
--Stop-band attenuation configuration via
Band-diagram
--Editable pole-zero plots with conjugate
movement support
--Impulse, Step responses, phase delay
plots
--Overlay plots
The tool is not completely tested as some more
bugs need to be patched. A few more feature
additions like add/delete buttons for poles and
zeros, alignment of band-diagram, addition of
button icons etc are still pending. All these
additions will be done with in a few days. In the
meantime it would be really helpful if the
community can provide feedback in terms of feature
additions or bugs so that I can improve the
design.
After finishing this work I will start working
on the actual IIR filter designs as gr-filter has
already been merged to the master.
_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________
Discuss-gnuradio mailing list address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
|