I managed to get gr-rds built tonight, and was able to test it planted
inside my FM receiver.
Even for a local station, there are lots of errors, and more distant
stations get even more.
I’m going to play around a bit with filter bandwidths, etc, to see if
I can improve it.
Looking at the modulation technique, it’s no big surprise–they use a
bi-phase baseband coding
(basically like differential Manchester), that drives a DSB-SC
modulator with a carrier at
57kHz. But that signal is generally about 20dB below the other audio
sub-channels on the
FM demodulator output. So the SNR is generally terrible.
I’m rather surprised, after looking at it, that anyone bother deploying
RDS. It seems utterly
crappy. It’s generally broadcast with fairly-low SNR, and there’s no
FEC built in, only CRC.
I re-jigged the XML for the demodulator part so you can reset the
demodulator state machine
from within a GRC flow-graph (there was a reset() method on the demod,
but it wasn’t exposed
within the XML framework for GRC).
Also, I may re-jig the decoder portion–it currently injects decoded
frames into a message queue and
to stdout. Might be useful to do other things with those messages.
Like is there a simple way to
have a message dequeuer within a GRC flow-graph so that one could stuff
messages into a variable?
–
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
I think the problems with SNR is with the FM Demod, my car stereo can
get RDS as the station fades 40 miles away, but with my USRP has
trouble picking up the stereo pilot from a 50kw station one mile away
( -10dBm ). FM demod is hard for software ( I think GNUradio just uses
a zero-crossing detector ).
Do you have the code, I would love to test it.
Actually, I’m using the blks2 WBFM block, which I believe is a PLL
demodulator.
Both of my “test” stations are actually roughly 100km
away from me, broadcasting from a high point in the Ottawa Valley.
One
station comes in about 10dB stronger than the other, although both have
an ERP of about 100kW, and both are located on the same transmit tower.
I’ll send you my ‘stuff’ once I’m more comfortable with it. The audio
quality is actually reasonably good, so I think the FM demodulator
portion is just fine.
-Marcus
On Fri, 24 Feb 2012 11:12:28 -0500,
Andrew D. wrote:
I think the problems with SNR is with the FM
Demod, my car stereo can
get RDS as the station fades 40 miles away,
but with my USRP has
trouble picking up the stereo pilot from a 50kw
station one mile away
( -10dBm ). FM demod is hard for software ( I
think GNUradio just uses
a zero-crossing detector ).
Do you have
the code, I would love to test it.
On Thu, Feb 23, 2012 at 11:33
PM, Marcus D. Leech wrote:
I managed to get gr-rds built tonight, and
was able to test it planted inside my FM receiver. Even for a local
station, there are lots of errors, and more distant stations get even
more. I’m going to play around a bit with filter bandwidths, etc, to see
if I can improve it. Looking at the modulation technique, it’s no big
surprise–they use a bi-phase baseband coding (basically like
differential Manchester), that drives a DSB-SC modulator with a carrier
at 57kHz. But that signal is generally about 20dB below the other audio
sub-channels on the FM demodulator output. So the SNR is generally
terrible. I’m rather surprised, after looking at it, that anyone
bother deploying RDS. It seems utterly crappy. It’s generally broadcast
with fairly-low SNR, and there’s no FEC built in, only CRC. I re-jigged
the XML for the demodulator part so you can reset the demodulator state
machine from within a GRC flow-graph (there was a reset() method on the
demod, but it wasn’t exposed within the XML framework for GRC). Also, I
may re-jig the decoder portion–it currently injects decoded frames into
a message queue and to stdout. Might be useful to do other things with
those messages. Like is there a simple way to have a message dequeuer
within a GRC flow-graph so that one could stuff messages into a
variable? – Principal Investigator Shirleys Bay Radio Astronomy
Consortium http://www.sbrac.org [1]
_______________________________________________ Discuss-gnuradio mailing
list [email protected] [2]
Discuss-gnuradio Info Page [3]
Links:
[1] http://www.sbrac.org
[2]
mailto:[email protected]
[3]
Discuss-gnuradio Info Page
[4]
mailto:[email protected]
It would be great to have the rds code inside the gnuradio tree instead
of
in a separate cgran project!
; )
I don’t know why, but when I tried the cgran rds code, the grc example
worked much better then the python one.
Best regards,
Rafael D.
The CGRAN RDS code seems abandoned, I’m thinking about cloning it to
github so it can be converted to UHD and Cmake and fixed up.
It would be nice, but it could be fixed up and converted to cmake in the
current svn repo also…
I don’t know about having it inside the main codebase, but
converting it to use the CMake build system would be good, since it
didn’t really understand how to place its python files into a filesystem
that uses /usr/local/lib64–so I had to move them manually.
-Marcus
On Fri, 24 Feb 2012 10:23:48 -0800, Rafael D. wrote:
It would
be great to have the rds code inside the gnuradio tree instead of
in a
separate cgran project!
; )
I don’t know why, but when I tried
the cgran rds code, the grc example
worked much better then the python
one.
Best regards,
Rafael D.
Actually, I’m using the
blks2 WBFM block, which I believe is a PLL demodulator. Both of my
“test” stations are actually roughly 100km away from me, broadcasting
from a high point in the Ottawa Valley. One station comes in about 10dB
stronger than the other, although both have an ERP of about 100kW, and
both are located on the same transmit tower. I’ll send you my ‘stuff’
once I’m more comfortable with it. The audio quality is actually
reasonably good, so I think the FM demodulator portion is just fine.
-Marcus On Fri, 24 Feb 2012 11:12:28 -0500, Andrew D. wrote:
I think the problems with SNR is with the FM
Demod, my car stereo can
get RDS as the station fades 40 miles away,
but with my USRP
has style=“padding-left:5px; border-left:#1010ff 2px
n-left:5px;
width:100%”>( -10dBm ). FM demod is hard for software ( I think GNUradio
just uses I managed to get gr-rds built tonight, a
here are
lots of errors, and mo
ations get even more. I’m going to play
around a bit with filter bandwidths, etc, to see if I can improve it.
Looking at the modulat
ester), that drives a DSB-SC modulator
with a carrier at 57kHz. But that signal is generally about 20dB below
the other audio sub-channels on the FM demodulator out
NR is
generally terrible. I’m rather surprised, after looking at it, that
anyone bother deploying RDS. It seems utterly crappy. It’s generally
broadcast with fairly-low SNR, and there’s no FEC built in, only CRC. I
re-jigged the XML for the demodulator part so you can reset the
demodulator state machine from within a GRC flow-graph (there was a
reset() method on the demod, but it wasn’t exposed within the XML
framework for GRC). Also, I may re-jig the decoder portion–it currently
injects decoded frames into a message queue and to stdout. Might be
useful to do other things with those messages. Like is there a simple
way to have a message dequeuer within a GRC flow-graph so that one could
stuff messages into a variable? – Principal Investigator Shirleys Bay
Radio Astronomy Consortium http://www.sbrac.org [1] [1]
_______________________________________________ Discuss-gnuradio mailing
list [email protected] [2] [2]
Discuss-gnuradio Info Page [3] [3] Links:
------ [1] http://www.sbrac.org [4] [2] mailto:[email protected]
[5] [3] Discuss-gnuradio Info Page [6] [4]
mailto:[email protected] [7]
_______________________________________________ Discuss-gnuradio mailing
list [email protected] [8]
Discuss-gnuradio Info Page [9]
_______________________________________________ Discuss-gnuradio mailing
list [email protected] [10]
Discuss-gnuradio Info Page [11]
Links:
[1] http://www.sbrac.org
[2]
mailto:[email protected]
[3]
Discuss-gnuradio Info Page
[4]
http://www.sbrac.org
[5] mailto:[email protected]
[6]
Discuss-gnuradio Info Page
[7]
mailto:[email protected]
[8] mailto:[email protected]
[9]
Discuss-gnuradio Info Page
[10]
mailto:[email protected]
[11]
Discuss-gnuradio Info Page
On Fri, Feb 24, 2012 at 2:58 PM, Philip B. [email protected]
wrote:
On 02/24/2012 02:45 PM, Rafael D. wrote:
It would be nice, but it could be fixed up and converted to cmake in the
current svn repo also…
But most of us have forgotten how to use subversion 
For those who want to have their cake and eat it too:
http://subgit.com/
Brian
On 02/24/2012 02:45 PM, Rafael D. wrote:
It would be nice, but it could be fixed up and converted to cmake in the
current svn repo also…
But most of us have forgotten how to use subversion 
Philip