Talking about DSP and SDR [Was: On tunnel.py]

To move on in this topic, Alex brings up an interesting point. One of
the reasons for this mailing list is to discuss signal processing,
SDR, algorithms, etc. Hopefully in a way that helps us improve either
GNU Radio itself of our use of GNU Radio tools and blocks. I would
really like to see more of this.

The problem is that in the past, every now and then, someone tries to
do this. But it never seems to work or end very satisfactorily. The
mailing list structure seems like it doesn’t lend itself to properly
discuss signal processing in a useful way.

To start with, we need to start passing around flowgraphs (preferably
GRC) so that we can all see what’s being discussed. There is also the
gnuradio.org wiki that people could use to build pages that better
describe the problem, idea, or suggestion. That format can handle
images, better embedding of formulas, and file attachments (include
data sets that are too big to mail out).

I would love to see someone take this on as a challenge to try to spur
on more discussion about communications. Over the past few years,
we’ve also gotten much better about using Git and tools like github to
submit patches and fixes. So I would encourage everyone to use the
available tools to help us all better communicate with each other.

Thanks,
Tom

On Sat, Mar 30, 2013 at 9:54 AM, Tom R. [email protected] wrote:

submit patches and fixes. So I would encourage everyone to use the
available tools to help us all better communicate with each other.

As I am here mainly for the algorithm talk, this sounds like a great
idea.
Are you envisioning a new Signal Processing and SDR section to the main
page? Some of the topics I think that would be a good start are:

  1. Understanding the (insert commercial standard here) MAC and PHY
  2. Frequency calibration using your local GSM station
  3. The near-far problem and how hardware still matters with software
    defined radios
  4. Synchronization and frequency offset
  5. Modulations and channel filtering
  6. Forward Error Correction
  7. Simulation within GNU Radio
  8. OFDM Demystified
  9. Diversity in Space, Time, and Frequency
  10. Wireless Channels and how to mitigate their effects

I am pretty sure most, if not all, of the topics are covered somewhere
in
the GNU Radio code base or can be easily explained by some people on the
mailing list.

Also, I am sure most everyone has seen it, but Charan Langton’s website:

http://complextoreal.com/tutorials/

has some excellent, freely available tutorials that explain the theory
and
math of a slew of communication related topics. It would be great to
link
those tutorials with a real implementation in GNU Radio.

I think some good candidates for the first topics might be all the new
OFDM
work as well as the simulation aspect of GNU Radio so more of the user
base
feels empowered by the software without feeling like they need hardware
to
be able to use GNU Radio.

So what do you think should be first? What would you like to see? I
know
you’ve blogged about GNU Radio updates. Should those be topics
discussing
algorithms and pragmatic GNU Radio?

I think you said it best:

“It’s not easy, but communications is not easy. In fact, it’s very,
very
hard.”

Seems to apply to every aspect of communication and is not limited to
wireless.

Brian

On Sat, Mar 30, 2013 at 11:35 AM, Brian P. [email protected]
wrote:

mailing list structure seems like it doesn’t lend itself to properly
on more discussion about communications. Over the past few years,

  1. Frequency calibration using your local GSM station
    I am pretty sure most, if not all, of the topics are covered somewhere in

“It’s not easy, but communications is not easy. In fact, it’s very, very
hard.”

Seems to apply to every aspect of communication and is not limited to
wireless.

Brian

Hi Brian,

Yes, having a wiki space to describe and discuss these issues was
exactly my thinking. I use my personal blog (www.trondeau.com/blog) to
discuss some things that I’ve done with GNU Radio, but having a space
on gnuradio.org where people can pose questions, answers, tutorials,
etc. should be a real boon to the community.

I love the idea of taking those tutorials you pointed to and making
GNU Radio applications that showcase them. I suggested a similar
concept of using the “DSP Tips and Tricks” section of IEEE Sig. Proc.
Magazine a while ago:
http://www.trondeau.com/blog/2011/5/16/dsp-tips-and-tricks.html. They
are small, 2-page papers that show off some small DSP tip or trick (as
the name would imply) that is simple to express and therefore work up
in GNU Radio.

Students learning comms theory and DSP could benefit a lot from either
developing these or having them available for study.

Now the question is: who wants to start working on this concept?

I’d suggest that Brian start formatting the page based on the list he
put above, but I think we might be better served by having someone
come up with a full page for something, like the OFDM model you
mentioned, so we can get a feel for how this will go and how the pages
will eventually look.

Tom

On Sat, Mar 30, 2013 at 3:09 PM, Tom R. [email protected] wrote:

I’d suggest that Brian start formatting the page based on the list he
put above, but I think we might be better served by having someone
come up with a full page for something, like the OFDM model you
mentioned, so we can get a feel for how this will go and how the pages
will eventually look.

I setup a little outline of what I thought might be some of the big
hitter
topics here:

http://gnuradio.org/redmine/projects/gnuradio/wiki/SignalProcessing

They are replicated here for comments and suggestions:

Modulations
Specific types of modulations - analog, digital and diversity in
space,
time and frequency are all welcome!

Filtering
IIR, FIR, adaptive, and even analog - all the techniques.

Synchronization
Carrier, symbol and timing recovery.

Forward Error Correction
Including data integrity! Quite an expansive topic.

Hardware and Physics
The impacts of the physical world on software defined radio. Cascade
analysis, near-far problem, and channel models.

Signal Identification and Classification
Figuring out what is out there, and what it means to you. One mans
noise
is another mans signal. So which is it?

Standards
Dissecting and understanding specific industry standards such as
802.11,
GSM, 802.15.4, etc.

Comments? Thoughts?

I like the idea of starting with OFDM since it has gotten a major
rewrite
and is used extensively. Is the majority of the OFDM rewrite in the
next
branch?

Brian

On Tue, Apr 2, 2013 at 10:08 AM, Brian P. [email protected]
wrote:

I setup a little outline of what I thought might be some of the big hitter
Filtering
analysis, near-far problem, and channel models.

Signal Identification and Classification
Figuring out what is out there, and what it means to you. One mans noise
is another mans signal. So which is it?

Standards
Dissecting and understanding specific industry standards such as 802.11,
GSM, 802.15.4, etc.

Comments? Thoughts?

I think it’s a great start! Easy to evolve from here.

I like the idea of starting with OFDM since it has gotten a major rewrite
and is used extensively. Is the majority of the OFDM rewrite in the next
branch?

Brian

Most of the OFDM updates are on master with a few tweaks to the
interface that are on next.

We identified the main bug that’s preventing it from working
completely last week. We just have to see about the right way to fix
is, which will hopefully be soon.

Hopefully Martin, Ben, or Matt can find time to fill out this section
of the wiki, since they did the majority of the rework (might even be
a good staging ground to an actual published paper at the SDR Tech
Conference or somewhere else appropriate).

Tom