DRM Software Radio Forums  

Go Back   DRM Software Radio Forums > DRM Software Radio - User Forums > General Topics
User Name
Password
FAQ Members List Calendar Search Today's Posts Mark Forums Read

Reply
 
Thread Tools Search this Thread Display Modes
Old 29-07-2004, 22:07   #31
carknue
Registered User
 
carknue's Avatar
 
Join Date: Dec 2002
Location: Frankfurt, Germany
Posts: 1,321
That's why I use the 8 khz Filter of the AOR7030 when there are strong AM signals +/5 khz. It helps a lot, but if these AM signal have a strong modulation as well, you have no chance
__________________
regards, Carsten
AOR 7030 + Wellbrook ALA 1530P & DWT + Degen DE31 + TG33

Free DRM Tools for analysing, controling and automating your DRM reception and logging
carknue is offline   Reply With Quote
Old 29-07-2004, 23:41   #32
PaulButler
Registered User
 
Join Date: Jun 2004
Posts: 6
Co-channel and in-channel interference

Here's my two-penn'orth on the filter discussion.

Those getting a strong DRM signal probably need no filtering, as the DRM component is strong enough to win through.

Those getting a weaker signal, however, are prone to losing the lot because of both in-channel interference (another carrier, usually AM, within the DRM signal) and co-channel interference (another carrier at one or both edges of the DRM signal).

I often see a good DRM signal with a whopping great carrier soomewhere within it. The rogue carrier prevents some of the DRM carriers from being resolved and also, I suggest, causes the AGC to turn down the gain. Result - no DRM audio.

Solution - a sliding notch filter which can be placed over the rogue carrier, with variable width so it can be tweaked to kill it without also killing too many DRM carriers.

Conversely, I often see one or two carriers at the edges, 5kHz up or 5 kHz down from the centre frequency of the DRM signal. This is where the bandpass filter in DREAM may help - my initial tests suggest that it does in some cases.

But given the most common co-channel interferer is 5 kHZ away, reducing the bandwidth to 8 kHz would help enormously (except for strong interferers which intrude further into the DRM signal). Someone above has mentioned that using the receiver's 8kHz filer helps in this situation.

Here's another thought - why not reduce the transmitted DRM signal to 8 kHz? This would mean fewer carriers and therefore lower bit-rate or reduced resilience (fewer carriers means you can afford to lose less before decoding is unavailable). But it would enable us to use an 8kHz window at the receiver and get rid of a lot of the energy from the carriers 5 kHz above and below.

Another way to approach it is to leave the transmitted DRM as it stands but establish a DRM DX decoding specification which will allow decoding to take place on fewer carriers? The transmtter could send the full-specification DRM signal but we could adjust the decoding to get a reduced quality signal if interference prevents full decoding, by reducing the bandwidth to 8kHZ, or accepting fewer carriers, or both of these.

For those of us for whom DRM is DX, a DX DRM decoding algorithm would work wonders. I would rather receive slightly reduced quality audio (but still significantly better than analogue AM or SSB) than no audio at all, which is what we get for a lot of the time here. And after all, aren't we talking about long-haul options for shortwave broadcasters?

OK, everyone, what do you think?

Paul VK3DBP Melbourne
PaulButler is offline   Reply With Quote
Old 01-08-2004, 07:48   #33
DigiBC
Registered User
 
Join Date: May 2003
Location: near Cologne, Germany (51°00'N / 6°49'E)
Posts: 36
Thumbs up Dream - more improvements

Sorry to interrupt your filter discussion, but there are some new features added to the CVS version of Dream I'd like to post here:

Now Dream writes an INI file and saves configuration and window placement data.
If you open the Dream.ini in an editor you can add your coordinates (in degrees and minutes). Then Dream will insert these coordinates into your logs. Now the resulting DreamLog.txt file is compatible with the log file created by the DRM Software Radio (though the CRC entry is still missing) and it works with Carsten's reception analyser tool "DRMcalc" perfectly.

That's how the Dream.ini looks like:
Code:
[Logfile] frequency=9800 latitude=51°00'N longitude=6°49'E startlog=0 [Receiver] filter=0 flipspectrum=0 mlciter=2 muteaudio=0 snddevin=-1 snddevout=-1 [Window geometry] analdemhsize=419 analdemvis=0 analdemwsize=704 analdemxpos=316 analdemypos=316 mainhsize=490 ...
BTW: Yesterday new FAAD libraries have been released so I hope we can expect Dream to decode parametric stereo soon...

Regards - DigiBC
DigiBC is offline   Reply With Quote
Old 01-08-2004, 15:19   #34
Garf
Registered User
 
Join Date: Feb 2004
Posts: 6
I can confirm that the FAAD2 update included complete parametric stereo capabilty, as well as improved handling of signals with bit errors.

Overall DRM performance should be much improved now.
Garf is offline   Reply With Quote
Old 01-08-2004, 23:21   #35
dk8cb
Roland
 
Join Date: Mar 2004
Location: Munich, Germany
Posts: 2,748
Hi,

also, if one defines USE_SSE, decoder.c uses the undeclared functions "apply_scalefactors" and "apply_scalefactors_sse" which used to be defined in specrec.c but now aren't defined anymore.

volker.dsp is also not found.

dk8cb
dk8cb is offline   Reply With Quote
Old 02-08-2004, 01:44   #36
Garf
Registered User
 
Join Date: Feb 2004
Posts: 6
I never tried the CVS version.

Some minor edits will be needed because the API of FAAD2 changed, but they are pretty simple.

'DRMCH_SBR_LC_STEREO' -> DRMCH_SBR_PS_STEREO

faacDecInitDRM( -> faacDecInitDRM(&

You can do without USE_SEE and volker.dsp...
Garf is offline   Reply With Quote
Old 02-08-2004, 12:51   #37
G8JQW
Registered User
 
Join Date: Dec 2003
Location: Birmingham, UK
Posts: 12
@ tactis-ms

I’m also getting the same first block of errors as you. I guess that the “update faad2 include filename and function” errors was caused by you copying both faad.h and neaacdec.h from the include directory, only neaacdec.h is required as there is already a faad.h in the drm/libs.

73s, Roger
G8JQW is offline   Reply With Quote
Old 02-08-2004, 15:37   #38
Garf
Registered User
 
Join Date: Feb 2004
Posts: 6
There are some bugs remaining that may cause faad2 to crash with bad reception. They'll be fixed very soon.
Garf is offline   Reply With Quote
Old 02-08-2004, 16:49   #39
DigiBC
Registered User
 
Join Date: May 2003
Location: near Cologne, Germany (51°00'N / 6°49'E)
Posts: 36
First many thanks to Garf and the rest of the Nero team!

I haven't compiled the new FAAD2 code yet, but Volker (corrados) has just released modified versions of AudioSourceDecoder.cpp and AudioSourceDecoder.h (which won't work with the old FAAD library now!).
For more information visit the following thread of the Dream forum at SourceForge.net: http://sourceforge.net/forum/forum.p...orum_id=242204

@ tacitus-ms
The first block of compiler warnings is resulting from the Settings.cpp which was added 5 day ago. According to Volker they are normal but not displayed usually...

Regards - DigiBC

Last edited by DigiBC : 02-08-2004 at 16:55.
DigiBC is offline   Reply With Quote
Old 03-08-2004, 22:43   #40
tacitus-ms
Registered User
 
tacitus-ms's Avatar
 
Join Date: Apr 2004
Location: 52°04'N 7°42'E = Münster / Germany
Posts: 40
Quote:
Originally posted by Garf
There are some bugs remaining that may cause faad2 to crash with bad reception. They'll be fixed very soon.

Here is another bug, obviously caused by the new stereo feature, perhaps this description may help someone to analyse it. Therefore I posted this bug already on the forum of sourceforge.net

I am using the last version of Dream including the last faad2, which is able to decode p-stereo. (downloaded from the cvs of sourceforge.net on 03.08.04 about 12:00 ) (libfaad was compiled in debug mode).
Listening to DW world from Sines (on 5980 kHz, about 21:30 GMT) today I realized that three times the right audio channel fell silent. DW did not use stereo but 17,46 kbit/s AAC with SBR in mono. The S/N was very good (about 23 dB). Every time the right audio channel had fallen silent the audio signal of the right channel did not return until I tuned to another station (RTL 5990 kHz) or I restarted the software. After tuning to the other channel the signal returned and I could tune back to DW world. But a few minutes later the same happened again.

Last edited by tacitus-ms : 03-08-2004 at 22:49.
tacitus-ms is offline   Reply With Quote
Old 04-08-2004, 06:11   #41
DigiBC
Registered User
 
Join Date: May 2003
Location: near Cologne, Germany (51°00'N / 6°49'E)
Posts: 36
Maybe this bug is solved now.

There's a new version of hcr.c available!
Download from CVS: http://cvs.sourceforge.net/viewcvs.p.../libfaad/hcr.c

Regards - DigiBC
DigiBC is offline   Reply With Quote
Old 04-08-2004, 06:33   #42
DF1PAW
Registered User
 
Join Date: Jul 2003
Location: Gebhardshain JO30VR
Posts: 1
Release or Debug?

Do I need to compile Debug code? Or does the release code work now?

Do you know when (if ever) SSE support will be re-enabled?


Regards,
Andreas Weller
DF1PAW is offline   Reply With Quote
Old 04-08-2004, 11:14   #43
Garf
Registered User
 
Join Date: Feb 2004
Posts: 6
Release mode will work.

There's still a small bug that one channel may disappear with mono transmissions. I'm gathering a few more testcases to be sure the fixes work.

My first worry now is making sure everything works well, and improving performance with bad receptions even further. After that I may have a look at SSE support but it is certainly not a priority right now.
Garf is offline   Reply With Quote
Old 22-08-2004, 00:20   #44
PY2PLL
Registered User
 
Join Date: Aug 2004
Location: S. B. Campo, SP (Brazil)
Posts: 3
Filters ... to test

Hi ...

Yesterday I place a 10KHz FIR filter between my product detector and sound card. My PC is an 800MHz one so I can't use built in software BPF.

Hard to say if any improvment. The filter is running on a DSP56002EVM (Motorola), 48KHz sampling, etc etc ...

OK, the spectrum is much clear now ... but most of the problems still receiver's AGC trigger. Next step: I'll derive the AGC control voltage from this filter output (or use one channel as 10KHz BW filter and the other one as 4.5KHz BW AGC sampler.

Another thing I want to do: add 3 fixed notches, may be 20Hz wide each at F0-5KHz, F0 and F0+5KHz.

Does anybody know how to do it on such platform? I'm nil on DSPs and programming and untill now I can't belive that I make this 10KHz BW filter work :o)
__________________
73
Marcus
PY2PLL/PY3CRX
http://py.qsl.br
PY2PLL is offline   Reply With Quote
Old 22-08-2004, 14:58   #45
dk8cb
Roland
 
Join Date: Mar 2004
Location: Munich, Germany
Posts: 2,748
Hi,

you can do an "off-the-air" experiment to test the effectiveness of such a filter combination.

1. Record a 12 kHz IF signal with interference as a wav-file with 48 kHz sampling rate.

2. Run this wav-file through DL4YHF's Spectrum Lab program (freeware) and apply the filter function that you want to test. Let it write the output to a new file. Spectrum Lab may be found here
http://homepages.compuserve.de/WoBue.../spectra1.html

3. Run Dream using the commandline parameter -f once for the original file and once for the new file and compare results.

I have done so using the latest Dream CVS version and a notch filter on the centre frequency, but it did not improve decoding. The result may however depend on the kind of interference encountered. (It did improve reception with the last version of Dream, but not with the current one, which has a better channel estimation.)

On a sufficiently fast system, it might even be possible to do this all in real time by cascading both Spectrum Lab and Dream by means of a program named 'Virtual Audio Cable' which can be found at
http://www.ntonyx.com/vac.htm but so far, I have not tested this.

Edit: I am still searching for a freeware program that allows to cascade audio applications like 'Virtual Audio Cable' does, but I haven't found one yet. If anybody knows of one then please tell me.

73, Roland, dk8cb

Last edited by dk8cb : 22-08-2004 at 16:31.
dk8cb is offline   Reply With Quote
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is Off
HTML code is On
Forum Jump


All times are GMT. The time now is 14:44.


Powered by vBulletin Version 3.5.3
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.