View Full Version : Perseus SDR and DRM

23-03-2008, 08:06
Hi all,

since last week I own the Perseus SDR. This is a fantastic radio for analogue reception but I had a lot of troubles to use it for DRM. To help others and to exchange informations I started this thread. Here my informations gathered the last days:
The demo version of Virtual Adio Cable (VAC) adds every 5 to 10 seconds the word "trial" into the datastream which influences Dream badly - only the full version works.
The best settings of VAC I found until now can be seen in the attachment
With this settings the input spectrum in dream (start parameter -c 3) is very clear and DRM-decoding is stable with a SNR better than 30 dB.
With these settings the SNR below 20 dB is a little better compared to my DRT1, above 20 dB SNR the DRT1 is up to 5 dB better.

Bernd, DF9RB

23-03-2008, 11:10
Hi Bernd!

Thats great to hear!

Most of the guys owning a perseus by now don't like DRM and so there wheren't very much infos about perseus performance on drm. Please keep us informed.


25-03-2008, 15:14

until two days ago I thought transfering data digitally from (nearly) the antenna to the loudspeaker output is the best way to decode DRM. Now I know this is a wishful thinking concerning I/Q-signals. The Perseus receiver delivers has SR (sampling rate) of 31250, Dream works with SR of 48000. VAC transferes and resamples the data stream from the Perseus to Dream. When I configure VAC with a SR from 31250 to 48000 the Input Spectrum in Dream shows a nearly unsuppressed mirror image (see 31250_48000.png), and the SNR is not better than 26 dB. With these settings resyncs of Dream only happen when I produce a high CPU-load (average CPU-load approx. 30 % when running Perseus and Dream). When I force Perseus or Dream to an other SR by VAC the mirror image is much better suppressed (48000_62500.png) or perfect suppressed (62500_62500.png). The SNR is up to 33 dB then - but Dream resyncs every 5 to 20 minutes even when there was no high CPU-load!!!!!

Has anybody ideas what to try for solving this problem???

Bernd, DF9RB

25-03-2008, 16:00
Hi Bernd,
what version of Dream are you using, do you get the resyncs also with 1.9.6 and older or only with the latest version?

25-03-2008, 20:05
Hello Simone,

I tested mostly with 1.10.7 but I did also some tests with older versions and
I had the resyncs also with older versions of Dream. I often run Dream (also version 1.10.7) with the DRT1 in parallel to identify possible TX-failures or propagation problems. With the DRT1 I had no resyncs over an hour from DW Sines but more than 5 with the Perseus.... I presume the problem are the different sampling rates of Perseus and Dream.


27-03-2008, 11:47

I am very interested in the discussion. The receiver seems to be a marvellous progress. But till now, I did not buy one. Probably I will do it.

The problems related to DRM are downsampling of an audio stream to 48 kHz, the normal working frequency of DRM. To be forced to use as an example VAC by the owner and have experiences seems for me not to be the right way of presenting and developping a receiver.

There should be ? by hardware an output for quite right downsampled audio-streams for as an example DRM. Waiting for DRM+ says too, that there will be a new sampling frequency for this purpose. ... and one should be able, to have an down-sampled output with the possibility to down-regulate (or absolute ) the sample rate.

A lot of users will have a second problem too connected with special demodulations programms, which need an additional (second) output.

I like VAC and WAVECLONE. But PERSEUS should at some time be ready and not a test-bed for the profi.

Wolfgang Hartmann, Nürnberg, 49N28, 11E03

27-03-2008, 15:55

after many hours of testing I found out a setting of VAC, in Perseus and my computer which produces less resyncs together with Dream.
- I disabled "cool and quiet" in BIOS of my computer
- setting the parameters in VAC according to the attachment
- I increased the sampling rate in Perseus from 125 kS/s to 500 kS/s
With these settings I have resyncs between 30 minutes and 1 hour
The "cool and quiet" function adapts the PC-clock according to the CPU-load.
Presumable every clock-change affect the resampler badly. The sampling rate of the Perseus should have no influence to the datastream to DRM but only change the visible frequency range - but how knows?
Wolfgang the Perseus has a SR of 31250 in the mode "User" (I found this in a description) and seems to have the same SR in DRM mode. When I configure a SR range from 11025 to 48000 VAC displays 31250 when I start Perseus.
Perhaps a SR of 48000 from Perseus could solve my problem.
I observed everything concerning the Persus SDR in internet for weeks before I ordered one. And one reson to order was the continuous improvents done by the developers. So I am quite confident that also the DRM mode will work in some weeks.

Bernd, DF9RB

03-04-2008, 09:27
Hallo Bernd,

I have the impression, that one needs for DRM decoding a sample rate of 48 kHz normally. There is no profit to use 96 kHz or 192 kHz.

To have a different samplerate, You derive the signal for DRM decoding in programms as DREAM, SoDirRa and others, it is a special calculating to achieve this task. This could be done - I think easily - if one has to DOWN-sample from higher samplerates. The other side: UP-sampling is connected with a lot of work (cpu and programming) and possibly with a lot of minor mistakes.

All these factors are connected with I/Q-working, a very special demodulation type, which is very sensible to amplitude and phase-differences.

The Perseus should because of this factors, use generally downsampling for DRM. ... and: have an output for this purpose!

Thinking further: DRM+ will be coming with a samplerate of 100 kHz - maybee-. For this case, one would need an output samplerate of 100 kHz.

And: The overall question is the procedure to optimize amplitude and phase of a soundstream. Phase and amplitude are changing not only by means of VFO but also from outside signal coming in. As far as I see with experiments with the RxDis Software, they are changing during receiving.

The marvellous Perseus - as far as I see the receiver - has to improve soft- and hardware - to optimize all features of receiving.

Interesting times coming!


03-04-2008, 19:50
Today I tried WindRad 1.31 for DRM with the Perseus receiver - and what shall I say: it works perfect!!!! As long as I observe now no resyncs of Dream and better SNR than with the DRT1. Logs will follow.

Bernd, DF9RB

03-04-2008, 20:14
That's a very promising undertaking, Bernd !

Alan Johnson
04-04-2008, 03:25
I look forward to see what settings you are using. I assume that VAC is still required?


04-04-2008, 14:19
Hi all,

here a short log of the Perseus-receiver with Winrad vs DRT1 (DW from Sines with an excellent signal). The SNR from the Perseus moves a little up and down as I switched off the AGC which gives a little higher SNR at optimized level settings. To answer Alan´s question: yes also Winrad needs VAC. The big difference to the original Perseus-software is that SR from Winrad is 24000 instead of 31250 from the Perseus-software which seems to fit much better to the 48000 with which Dream works. Settings in Winrad are: Input: Perseus HF-receiver, Output: Virtual Audio Cable, Mode: DRM - That´s it! Now for an excellent receiver an excellent software exists!

Bernd, DF9RB

05-04-2008, 11:41
Hi Bernd!

I've got a question: In Winrad I've got to calibrate Phase and Amplitude on I/Q by myself and its optimized for just one frequency in the spectrum.

As far as I know the perseus (and his software) optimizes aliases by himself. Is it done by the Software (PC Software) ore more hardwareside (Perseus CPU)?

If it would be first methode, the Pereus would behave itself in Winrad like a standard I/Q Receiver (Optimized only at one Frequency, Images appear besides Optimum-Frequency).

If optimizing is done in the Perseus itsself, it would behave like a very wideband non-I/Q Receiver (without any images in the spectrum).



05-04-2008, 13:42
Hi Stefan,

I am not a specialist for I/Q-receivers but I presume that the Perseus-hardware optimizes aliasis itself. When I change amplitude, phase or delay in Channel Skew Calibartion in Winrad I see images in the spectrum. The optimum is always amplitude = 1, Phase/Delay = 0 between MW and 15 MHz where I tried. In "Input Spectrum" in Dream is a image but it seems not to influence Dream (see attachment).

Bernd, DF9RB

17-04-2008, 11:21

an information by Alberto says:

"when Winrad is set to DRM, it changes its output sampling rate to 24 kHz, so to be able to transfer a band of 12 kHz
to the output. The program DreaM is perfectly capable to use this sampling rate, both directly, using a second audio
card (as I do), or via VAC (Virtual Audio Cable), which I have and have experimented with."

I think resampling from the Perseus sampler rate of 31,.. kHz to 48 kHz is much more unsafe then to give the 24 kHz directly to a decoding program as DREAM or SoDiRa.

So far the good news about the new Winrad version related to DRM has its answer from programming skills of Alberto.

Wolfgang Hartmann, Nürnberg

Alan Johnson
20-04-2008, 21:20
I just realized that I have two sound cards in my computer - an onboard Realtek chip (currently disabled in BIOS) and a PCI-e ASUS. I wonder if I could use them instead of Virtual Audio Cable and get better results with Perseus and Dream? Can someone tell me how to configure them? Do I set up one as recording device and the other for playback in the Audio tab of Sounds and Audio Devices in Control Panel? I'm running XP Pro SP2.

Thanks in advance,
Alan Johnson

21-04-2008, 10:54
Hallo Alan,

just install/enable your Onboard-Soundcard and try to work with this situation.

If You call RECORD or PLAY-Mixer, You are able to choose the soundcard, You want to engage for a special purpose.

I try to append a screen from Winrad 1.31, where You could see, how it feels to have two soundcards. Sometimes it is a little complex, because all soundcard related programms try to optimize themself and change the fixing of the soundcard.

Take the weaker soundcard for output.

In some situation, the onboard-card is better than the additional card for decoding purpose.

Wolfgang Hartmann

PS: Till now, I do not use a PERSEUS

07-05-2008, 15:13
Hi all,

I found this rather interesting article by David Lipets of Tadiran Communications Ltd, concerning SDRs. Read all about it : http://www.rfdesignline.com/206902442 and also the Related Articles at the bottom of the third page.

11-02-2009, 22:32
Hi all,

after some tests over a few days with the PERSEUS, which one I got from a friend , I resumed, PERSEUS is a nice small receiver with quite good results,
but not perfect at all, particularly on DRM....

- 800 Euro , this is a lot of money, if I have a look on performance and
- I didn´t get any stable signal ( on DRM !) with the receiversoft
( perseus 1.21d) , VAC and DREAM. There was a resync every
few minutes, and I don`t no why. I tried a lot of parameter in VAC,
different DREAM versions,but nothing helped...
- with the free software Winrad 1.41 everything is still working
( ..and in my mind , with a better performance!)
- and, btw. ,- high pc performance is needed for high Output Sample Rate..


12-02-2009, 17:46
Hello Michael,

my Perseus is running very stable in DRM-Mode after a lot of try and error adjustments.
The most important settings are:
SR 62500 to 62500 and "cable format" in VAC and
Audio Latency 8 (minimum) in Perseus.

With these settings I have no resyncs in Dream for hours (under good condx)

The only little disadvantage of the Perseus is the not optimized AGC - but now the developer (Nico) started to redesign the AGC for Dream. He promised me a test version the next days.

Bernd, DF9RB

12-02-2009, 17:55
Hi all,

800 Euro , this is a lot of money, if I have a look on performance and

....and only up to 30 MHz (40 MHz wo preselector)...

12-02-2009, 18:38

Winrad has a special management of Samplerates.

You are allowed to be free, to use a lot of different samplerates for input. I see for output only 8,0 kHz, 11 kHz and 24 kHz. Dream seems to prefer 48 kHz as other programs like SoDiRa.

I think, because of that feature, Dream has to change the samplerate while the signal is sent. This procedure could take some more processor time.

The AGC optimization for DRM is dependent from the velocity of AGC. The faster agc is working, the worser for decoding a signal. The whole signal in one period contains a lot of - let me say - subcarriers, which are combined by the digital process. It needs a certain time and withtin this time there should no remarkable changing of amplitudes of signals even there are only a few peakes to be seen.

Results of a lot of practice and changing agc-factors in experimental situations.

There is a program, able to change samplerates (RATEMONKEY), which could be usefull for the situation.

If Winrad will be changed in future to higher samplerates as 48 kHz for output the whole program would be dependent of more cpu power.

I hope, I gave right hints from my practical life in DRM.

Wolfang Hartmann, Nürnberg

13-03-2009, 13:01
The only little disadvantage of the Perseus is the not optimized AGC - but now the developer (Nico) started to redesign the AGC for Dream. He promised me a test version the next days.

Bernd, DF9RB

Hi Bernd,
any news about the AGC yet?

13-03-2009, 16:10
Hi Simone,

there is no new software available yet but I am very confindent that Nico will keep his promise as he always did in the past.
I will give more informations and test results as soon as the new software is available.

Bernd, DF9RB

31-05-2009, 19:11
Hello all,

the new software for the Perseus (Version 2.1f) was available for some time.
Now I found the time to test the new AGC for DRM. The result is amazing.
With a synthesized DRM signal the SNR is now up to 51 dB!
Over the air test showed always a better SNR than my good old DRT1 and a few less audio drop-outs under critical conditions.

Bernd, DF9RB

01-06-2009, 08:32

Perseus and DRT1 sharing the same antenna. Signal with strong and fast fading. 1st Perseus, 2nd DRT1

Bernd, DF9RB

01-06-2009, 08:56
Over the air test showed always a better SNR than my good old DRT1... Bernd

Hi Bernd,
How about give VoR from Bolshakovo a try for SNR? 9.73 MHz topped 40.96 here on my FRG-7700 one day and this morning (so far) 39.98 dB.

01-06-2009, 12:18
Hi all,

@Bernd, excellent... give it a try and specify HW & SW setup, antennas, preselector if any, filters....

01-06-2009, 14:58
Hello Terje, John,

Terje, I think during you typed your message I was listening to VoR.
But today was deep fading and a PLC-modem was active (I couldn t find out until now from which neighbor it comes from....). So the maximum SNR today was "only" 38 dB.
John I used a CD with an DRM-Signal, fed to a mixer (NE602) following the Perseus tuned to 477 kHz. This signal was for sure no optimal but i did not have something better.
The On-Air test was with my EWE antenna and a simple T-Plug (nothing matched!) deviding the signal to the Perseus and DRT1.

Bernd, DF9RB

01-06-2009, 15:08
Hi all,


...I used a CD with an DRM-Signal, fed to a mixer (NE602)...

I see... the old fashioned way...

21-01-2011, 18:17
I like to check condx with the PERSEUS. :)
Outstanding Radio Taiwan International signal on 3955 kHz in AM. :D

23-01-2011, 07:26

a new software (Perseus40a) is available now. With this new software the Perseus SDR can act as a server or client. Over a fast network even DRM is possible. I tested this yesterday over WLAN between my PC and my netbook. It worked for hours without interruptions.
In the screenshot (badly fragmented because of the 100 k limit) you can
see 3 Perseus in parallel. One is at home, the other over remote desktop from our clubstation (Perseus server software not yet installed) and a third Perseus from a server near Munich.
On http://microtelecom.it/map/PerseusServers.html you can see the worldwide active Perseus server.

73s Bernd, DF9RB

23-01-2011, 08:16

here a better screenshot.
Perseus Server, Perseus Client and Dream running on the same PC.
(CPU load around 5 % on a quad-core)
Datarate is around 640 kbps in LAN_Mode for DRM.
In ADSL_Mode the datarate is around 150 kbps with good
audio quality, 70 kbps in GPRS_Mode with a little reduced
audio quality.

Bernd, DF9RB

24-01-2011, 18:05

today I tested DRM with the Perseus over internet.
The Perseus_Server at our clubstation was configured in "LAN-Mode".
The Perseus_Client and Dream running at home.
The datarate is around 640 kbps.
Some resyncy because of interrupted internet traffic.

Bernd, DF9RB