On Sun, Mar 17, 2019 at 09:50:51AM +0530, Anirban Banik wrote:
> Hi Felix, Marcus,
> I apologise for my delayed reply. I have meanwhile gone through most of
> the tutorials on the GNU Radio website. I am interested in the Filter
> Design Tool Enhancements.Â
> As suggested by Felix, I would be obliged if you could provide me with
> more information on the possible outcomes of the project, Marcus, and how
> I need to approach, and begin working on it.
Anirban,
one important thing to keep in mind is that you have about 3 months of
coding time; students often overestimate what they can achieve in that
timeframe. So, a very important part of writing a proposal is having
a good and realistic idea of what you can achieve. Have a solid
base-level goal, and then add stretch goals. Make it so that you will
finish the base-level goal with a 95% confidence within the given
timeframe, and then add some stretch goals such that the last stretch
goal will be completed with a 5-10% probability.
Applying this to the filter design tool: Go run the current filter
design tool with the latest 3.8 branch. You will immediately find that
certain things don't work well, who knows if the tool works at all.
First step would thus be to polish the tool and clean it up, modernize
it. I'm not sure about its state w.r.t. no longer using pyqwt. Maybe
that's a place to start.
Then, take a look at its extensibility. Last time I checked, it couldn't
generate half-band filters, for example. Instead of just adding those,
though, think about how the tool could be written that adding filter
types is trivial.
-- M
No comments:
Post a Comment