DSD4Win hardly detect P25

[ronenp]

Hi
I have played a bit with DSD4Win and tried to decode the P25 with SDRSharp and RTLbased SDR
on one PC the CPU raised to 100% when the DSD started to decode and by that I could hear a word here and a word there but it was impossible to understand the conversation

Yesterday i tried it on another PC and the CPU was only on 50% peak but same problem
is there any special setting needed to be done Such as Audio level ? or VAC parameters ? if yes how do i ajust it correctly ?
anyone has a success in decoding reasonable the P25 ?
Please Advice
Thanks Forward
Ronen - 4Z4ZQ
Ronen Pinchooks (4Z4ZQ) WebSite

 

[n4yek]

While I no nothing about SDR radio just curious as to where are you getting your audio feed from that you feed the DSD Software?
I know on a scanner you can't take it from a headphone jack or speaker jack, you need to pull your audio from the discriminator in order for DSD to be able to decode.
Just a thought.

 

[ronenp]

The SDR oudio is getting directly to the Virtual Audio Cable and from there to the DSD
It is like output from Discriminator

 

[KenK]

Limited luck here too

I have only got a word here and there with the same set up as you .

I would be curious to know what is a successful setup to copy APCO
with DSD.

Is DSD the only free software that tries to decode APCO?

Ken
N1KK

 

[michaeldim]

It could be simulcast distortion, that's messed me up on both my scanner, and DSD.

I use the same setup as you, it'd seem. In SDRSharp, Make sure that the audio filter is switched off, of course. Also make sure your using the latest version.

Then mess around with the gain settings, I was able to get it to decently decode a good sentence or two.

DSD still isn't brilliant at decoding, mind. Even working it's best it sounds broken and watery.

 

[mtindor]

There are simply too many variables to really suggest why you might not be able to copy P25 well. A few of them are:

1. Audio Levels

You need to make sure you have proper audio levels into DSD from SDRSharp.

2. SDRSharp settings

When I installed SDRSharp, for use with a Funcube Dongle, filtering was enabled by default. If you are going to run with Filtering enabled, you better make damned sure you have fiddled around for quite some time determining proper filter settings. The default filter settings in SDRSharp, at least in my case, wouldn't allow me to decode anyother other than Moto Type II.

What I did was uncheck "Filter Audio" under the Audio section. Then I saw that I was decoding everything fairly well. At that point I realized that the default filter settings were nowhere near optimal for decoding P25 / DMR.

3. CPU resources.

SDRSharp is pretty light on CPU usage, compared to other SDR software. But, the combination of SDRSharp, DSD and some other software can overburden a machine that doesn't have multiple cores or doesn't have hyperthreading.

I'd guess though that your problem is likely a combination of improper audio levels along with having "Filter Audio" checkmarked with suboptimum filter settings.

Being that I have only used the Funcube Dongle [and thus have never used an RTL-based SDR device], I can't say how your audio settings should be. With the Funcube Dongle though, for best decoding across the board, I need to lower the audio level (under Recording Devices) for the actual Funcube Dongle to "1" or "2". Anything more and there is too much gain being piped into the SDR software. On the Funcube Dongle, anything over a "1" or "2" for the audio level on the device degrades performance. This may or may not be an issue with RTL devices.

Audio levels into DSD also are semi-critical. I've found that an INLVL of 20-30 in DSD seems to be best for my setup. If you have an INLVL significantly below 20 or significantly above 30, I'd suggest adjusting your AF Gain in SDRSharp to a level that provides you with a DSD INLVL of 20-30.

If you still have problems after all of this, then I guess CPU could be an issue. It might help if you'd give us an idea of what type of CPU you are using -- ya know, the good stuff -- make / model / speed / HT or non-HT / multicore, etc.

Mike

 

[ronenp]

Thank you for the brief explain
The Audio of course id with filter off
but regarding the level is there any tip how to adjust the level ? is it only by listening and adjusting till the best quality of decode come out ?

as for the CPU it is a notebook with pentum 1.6GHz Hope that kind of CPU should be inough

Any suggestions welcome
Regards
Ronen - 4Z4ZQ
Ronen Pinchooks (4Z4ZQ) WebSite

 

[mtindor]

Thank you for the brief explain
The Audio of course id with filter off
but regarding the level is there any tip how to adjust the level ? is it only by listening and adjusting till the best quality of decode come out ?

Brief? Not sure how much more lengthy you would want the explanation to be rofl.

Use the AF Gain slider in SDRSharp to adjust the audio that is being sent from SDRSharp. Do this while watching DSD and noting the INLVL in DSD. Everyone will have a differing opinion on what the best INLVL is. I would recommend that you adjust the AF Gain in SDRSHarp so that the INLVL in DSD shows between 20-25.

as for the CPU it is a notebook with pentum 1.6GHz Hope that kind of CPU should be inough

Notebooks typically don't have the same kind of performance [because they have power saving features built in] that a desktop does. You didnt'say _exactly_ what CPU model you have. "Pentium 1.6" isn't really a CPU model. But I'll go out on a limb and suggest that your laptop probably is not fast enough for optimum performance.

But, IF your laptop isn't up to the task, there isn't much you can do about that. So you need to make sure that you have properly adjusted audio levels. You might also want to use various command line switches in DSD to force P25 optimizations. If the system you are monitoring is simulcast, you might have inherent problems with multipath distortion. If the P25 signal that you are listening to is weak, and if you aren't using DSD with DMR Filter, then you might want to try using the version of DSD with hte DMR filter.

Mike

 

[ronenp]

Hi again
I have made some more research trying to figure out where the problem is
I saw that while the software running on the notebook and the DSD try to decode the CPU raise to 100% so installed a spectravu which take less CPU and then when the DSD decode the CPU reach almost the 100% but still no good detect
so i install it on a Desktop with core 2 extreme Processor with 4 core of 3.4GHZ (X9770) and CPU now was 5 Percent on every CPU even while the DSD decode but same problem a word here and there and nothing else the volume set to 2-% decode on the DSD i even tried the new DSD and same problem in P25 and MotoTRBO

I have noticed a strange behave with this setup that when i press the "Play" Button on the SDRsharp after few seconds it start to behave strange like the software run in pulses of 1 second i.e the refresh of the spectrum anlyzer is done every 1 or 2 second interval and not as usual also then the DSD stop decoding because probebly it dont get the full streem only bursts of it .. if i close the DSD the SDRsharp start to work properly again and if i press the stop and play again in the SDRsharp (without colsing the DSD ) after a second or two same problem occure mabe it is a buffer problem in the DSD or so ...
Im frustrated
what else can be done ?
Any advice is welcome
I use SDRSharp latest release VAC 4.1 licenced (not the demo) and tried with latest DSD 1.6 but also with the famous 1.4
The receiver is the RTLSDR dongle ...
and the OS on the PC is winXP SP3

 

[mtindor]

Hi again
I have made some more research trying to figure out where the problem is
I saw that while the software running on the notebook and the DSD try to decode the CPU raise to 100% so installed a spectravu which take less CPU and then when the DSD decode the CPU reach almost the 100% but still no good detect
so i install it on a Desktop with core 2 extreme Processor with 4 core of 3.4GHZ (X9770) and CPU now was 5 Percent on every CPU even while the DSD decode but same problem a word here and there and nothing else the volume set to 2-% decode on the DSD i even tried the new DSD and same problem in P25 and MotoTRBO

I have no doubt thta 100% CPU will cause problems. However, you said that you still can't get a decent decode even when using an X9770 desktop @ 5% per CPU. So the underlying problem isn't CPU resource starvation.

I'd question the incoming signal. Are you sure it is clean? Is this a simulcast P25 system? If it's a simulcast system you might be suffering from multipath distortion. Could you be suffering from strong signal overload?

What I would suggest next is that you set up Unitrunker to see if you can sit on the control channel of the P25 system and get a solid control channel decode (90+ health). If you cannot, then either you don't have a clean signal, or you have the wrong filtering settings, or you have too much or not enough audio gain into DSD.

Make sure that your INLVL in DSD 1.4/1.5 is no more than about 25. I think Woodpecker changed how INLVL is rpeorted in DSD 1.6, so I'm not sure yet what the INLVL should be in 1.6.

In your SDRSharp make sure you don't have any filtering enabled. With no audio filtering in SDRSharp, you should try to decode the control channel of any EDACS or P25 trunked system to see if you're getting good decode rates when running Unitrunker.

Also, you might want to consider recording a P25 audio sample (using the VAC Line In as your audio source, when you are running SDRSharp), and then sending it to somebody else [me if you want] so that somebody else can check the quality of the audio that is being fed into DSD.

Mike

 

[ronenp]

Thank you for your quick reply
Few more details
the system are Trunked P25 and another is regular mototrbo rptr no simulcast
as for unitrunker the unitrunker decode excelent the signal even in much lower signal then the one i test
As for recording im willing to do it but dont know how
may you tell me what should i do in order to record so i will record it and send it to u
Thanks Forward