DSDPlus

[Jay911]

I too am running a Win7 64-bit machine. I was using unregistered (as mentioned before) VAC, and SDR# 1172.

I might try later on my laptop (also Win7 x64) with my PSR-800's "tap".

 

[Fizz306]

DSD Plus?

sdr radio is not the right choice for this kind of setup as it filters the audio without the option to change it. the closest thing to a discriminator tap with the rtl is sdr#, which allows you to remove any filters from the audio.

 

[KC1UA]

sdr radio is not the right choice for this kind of setup as it filters the audio without the option to change it. the closest thing to a discriminator tap with the rtl is sdr#, which allows you to remove any filters from the audio.

That's a very good point I have entirely overlooked, and that's why it's always a good idea to describe what you're using for a setup when problems occur. I had become so used to using SDR-Radio I entirely forgot about that. Decoding with SDRSharp is immediately near perfect.

Ironically I've used SDR-Radio as described for near perfect decode of DMR and pretty good decode of P25 using the aforementioned method. I guess DMR must be more forgiving.

Thanks for the virtual dope slap!

 

[Jay911]

Fizz306 - or whoever else can help - would you mind giving me a quick rundown of how SDR# should be used to deliver a good clean signal to DSD+ and other such software? For example, do I want to select "NFM" or "RAW" in the Radio tab? In the Audio tab, I presume "Filter Audio" should be unchecked. Am I right? (I know to set my output to "[MME] Line 1 (Virtual Audio Cable)", and set whatever I'm using on the other end - DSD+/Unitrunker/etc - to listen to that same audio input.)

Any other settings crucial to making this work right, please let me know. Thanks much in advance.

 

[KC1UA]

I haven't tried RAW but using NFM mode with Filter Audio unchecked is doing the job for me. I also have Filter bandwidth set at 12000, filter order at 400, squelch checked and at 60 although that could be unchecked I would think and maybe should be for weak signals. Snap to grid and Correct IQ are both checked. I have AF Gain set at just past the 4th segment, and a sample rate of 48000 sample/sec.

 

[mtindor]

Fizz306 - or whoever else can help - would you mind giving me a quick rundown of how SDR# should be used to deliver a good clean signal to DSD+ and other such software? For example, do I want to select "NFM" or "RAW" in the Radio tab? In the Audio tab, I presume "Filter Audio" should be unchecked. Am I right? (I know to set my output to "[MME] Line 1 (Virtual Audio Cable)", and set whatever I'm using on the other end - DSD+/Unitrunker/etc - to listen to that same audio input.)

Any other settings crucial to making this work right, please let me know. Thanks much in advance.

The most crucial setting by far is to ensure that "Filter Audio" is unchecked.

Mode: NFM

Filter Bandwidth: 9000
Filter Order: 10
Filter Audio: unchecked

One can argue whether anything needs done other than unchecking "Filter Audio", but I'd rather be on the safe side. If I did not uncheck "Filter Audio", then the best Filter Order seems to be 10 [and thanks to whomever on the forum clued me in to that]. So I set it to 10 regardless of whether I have Filter Audio unchecked.

I also modify the following line in the config:

<add key="minOutputSampleRate" value="32000" />

to

<add key="minOutputSampleRate" value="48000" />

I'm not sure that this really accomplishes anything though.

That's what works for me.

And I'd also normally set my AF gain so that when I look in Control Panel --> Sounds --> Recording Devices, my VAC Line one is showing about 50%.

DSD was finicky about audio levels. DSDPlus doesn't appear to be finicky. DSDPlus appears to work equally well anywhere from low to high audio (but I still would not suggest an audio level of 1 or an audio level that obviously too high). I think one has to at least get it "in the ballpark."

Mike

 

[Fizz306]

Fizz306 - or whoever else can help - would you mind giving me a quick rundown of how SDR# should be used to deliver a good clean signal to DSD+ and other such software? For example, do I want to select "NFM" or "RAW" in the Radio tab? In the Audio tab, I presume "Filter Audio" should be unchecked. Am I right? (I know to set my output to "[MME] Line 1 (Virtual Audio Cable)", and set whatever I'm using on the other end - DSD+/Unitrunker/etc - to listen to that same audio input.)

Any other settings crucial to making this work right, please let me know. Thanks much in advance.

Jay, see my attached screen. This is what I have used in the past to decode with reasonable success. Yes, the bandwidth is less than half of the 12.5 kHz one would expect. Cuts out a lot of the noise. Also, Youssef filter type has also worked for me. As others have noted, unchecking Filter Audio is an absolute requirement.

Unfortunately, this was a big guessing game for me and took a while of playing around to get it working well. This all has been discussed at length on the DSD thread in the RR forums (it takes HOURS to read through all the pages to find a lot of this useful information). Read, play, experiment, rinse, and repeat. Over. And over. And over... and eventually, it MIGHT work. Maybe. If you're lucky.

If only there was a scanner that would do DMR.

Merry Christmas!

EDIT: it's worth noting that this is what I have used with DSD 1.6. YMMV with this new DSD+...

 

[tylerwatt12]

I usually use DSD for DMR decoding but I'm trying this version on a fairly strong P25 conventional signal and it just doesn't want to work. This same system uses the older DSD and the -f1 -mg options to get pretty good decode of the same signal.

I'm using an R820T dongle into a registered version of VAC and the latest beta of SDR-Radio. It's a Windows 7 Pro 64 bit machine. One thing I do notice is that it's doing a far better job of showing the NAC than the older version...but the audio is atrocious regardless of the input level. It does a nice job on DMR voice though, but so did the older version. I'll try it with a discriminator tapped BC898T later but I am definitely not having much luck with the dongle on P25.

Try setting WFM, and max out your AF gain

Sent from my Nexus 5 using Tapatalk

 

[Fizz306]

I also modify the following line in the config:

<add key="minOutputSampleRate" value="32000" />

to

<add key="minOutputSampleRate" value="48000" />

I'm not sure that this really accomplishes anything though.

Mike, this is in an attempt to get SDR# to force a clock rate of audio sampling within SDR#, if memory serves me correctly. The idea here is to get the software to sample the audio at 48kHz from a raw point of view before it sends it into VAC (or whatever sound output you use) at 48kHz. SDR# is a compromise from this point of view. The best way to really get data into DSD is to use a disc. tap from a scanner into a quality sound card that samples directly at 48kHz, and then into DSD. This was my very limited understanding of this. Maybe someone else here has a better idea...

Think of SDR#'s sampling like this. It would be like taking a 128 kbps MP3 and re-encoding it into a WAV. Just because you up convert it to a higher quality bit rate doesn't mean that it will be higher quality because you're starting with a lesser quality to begin with.

Some of this may be a little inaccurate, as I understand audio very well, but I am by no means a programmer.

I can't wait to get home later with my dongles and give this new flavor of DSD a shot. Super exciting!

 

[Fizz306]

Try setting WFM, and max out your AF gain

Sent from my Nexus 5 using Tapatalk

Not sure this is the best idea. WFM is introducing many magnitudes of excess bandwidth into the picture, introducing a ton of excess noise. Also, maxing your gain will overload the front end of the dongle, and will "blow out" your signal, making the signal unusable for DSD.

 

[moonbounce]

The filter band width on my computer is not a fixed value, it changes as the frequency changed. Is this normal?

 

[mtindor]

Mike, this is in an attempt to get SDR# to force a clock rate of audio sampling within SDR#, if memory serves me correctly. The idea here is to get the software to sample the audio at 48kHz from a raw point of view before it sends it into VAC (or whatever sound output you use) at 48kHz. SDR# is a compromise from this point of view. The best way to really get data into DSD is to use a disc. tap from a scanner into a quality sound card that samples directly at 48kHz, and then into DSD. This was my very limited understanding of this. Maybe someone else here has a better idea...

Think of SDR#'s sampling like this. It would be like taking a 128 kbps MP3 and re-encoding it into a WAV. Just because you up convert it to a higher quality bit rate doesn't mean that it will be higher quality because you're starting with a lesser quality to begin with.

Some of this may be a little inaccurate, as I understand audio very well, but I am by no means a programmer.

I can't wait to get home later with my dongles and give this new flavor of DSD a shot. Super exciting!

Thanks for the explanation, Rob. I hadn't started using that config modification until recently. I can pinpoint any better/worse performance for doing it. I had just seen that somebody recommended it, so I did it.

I get what you're saying though -- basically you can't polish a turd

Mike

 

[tylerwatt12]

Not sure this is the best idea. WFM is introducing many magnitudes of excess bandwidth into the picture, introducing a ton of excess noise. Also, maxing your gain will overload the front end of the dongle, and will "blow out" your signal, making the signal unusable for DSD.

When I tried it with regular DSD, it worked wonders. WFM is naturally quieter than NFM, I set the bandwidth to 10khz, set my PPM, then set a very small shift of a few Hz. When I moved from DSD to DSD+, the only thing that got better was the output audio filtering from DSD+. my decode was already pretty damn good with WFM

Sent from my Nexus 5 using Tapatalk

 

[mtindor]

When I tried it with regular DSD, it worked wonders. WFM is naturally quieter than NFM, I set the bandwidth to 10khz, set my PPM, then set a very small shift of a few Hz. When I moved from DSD to DSD+, the only thing that got better was the output audio filtering from DSD+. my decode was already pretty damn good with WFM

Sent from my Nexus 5 using Tapatalk

As much as this seems to defy logic, it works pretty damned well. I just tried WFM myself, and of course I increased the audio level quite a bit. My decodes in DSD are the same as using NFM and quite possibly even a little better. DSDPlus sounds fine in either setting as well. I think I'll try some TRBO using WFM.

Mike

 

[Fizz306]

DSD Plus?

When I tried it with regular DSD, it worked wonders. WFM is naturally quieter than NFM, I set the bandwidth to 10khz, set my PPM, then set a very small shift of a few Hz. When I moved from DSD to DSD+, the only thing that got better was the output audio filtering from DSD+. my decode was already pretty damn good with WFM

Sent from my Nexus 5 using Tapatalk

very interesting, thanks for the tip. will try later and report back!

 

[KC1UA]

Well this just gets better and better. Listening to DMR, Slot 1 voice comes from the left PC speaker, and Slot 2 comes from the right...simultaneously when both slots are in use.

edit: Whoops, I noticed after posting this that Mike already mentioned the simultaneous decoding, but sent to separate speakers...just too cool.

This is simply fantastic.

 

[mtindor]

Well this just gets better and better. Listening to DMR, Slot 1 voice comes from the left PC speaker, and Slot 2 comes from the right...simultaneously when both slots are in use.

edit: Whoops, I noticed after posting this that Mike already mentioned the simultaneous decoding, but sent to separate speakers...just too cool.

This is simply fantastic.

It sure is. It took me a minute to figure out what was going on when I first encountered it. Given that many TRBO systems have multiple transmitters and sites, you never know if you'll be able to reliably follow a conversation from start to finish. At least now, we know we never have to miss a conversation occurring on a single transmitter we're monitoring.

Mike

 

[moonbounce]

With a discriminator tap DSDplus makes for a 100% improvement in Ap 25.

With SDR Sharp it and my R820t dongle and VB cable it is actually worse than dsd1.6. But I am really happy to have this newest program, thanks redbeard for posting this thread..

Moonbounce

 

[mtindor]

With SDR Sharp it and my R820t dongle and VB cable it is actually worse than dsd1.6. But I am really happy to have this newest program, thanks redbeard for posting this thread..

I would think this would be the exact time for playing around with the advanced tuning. Your tap gives you the best input audio. Your setup (whether it be SDR# settings, sample rate mismatches, the R820t, whatever) probably has audio that is significantly more distorted than that which is coming off of the discriminator tap.

Go through the [tedious] process of performing the advanced tuning, and it may very well help you to get better performance out of the R820t + SDR# + DSD+.

By the way, I run both a discriminator tap and a dongle+SDR# combination, and I'm quite certain that the audio reaching DSDPlus from that combination is much worse than that which is delivered via the discriminator tap. The only problem is that it is the holidays and I haven't really had time to play around with advanced tuning in DSD to eek out the best performance from a less than stellar audio source. I do plan to play around with it though.

If you do attempt the advanced tuning, I would advise you not to skimp. Do it per the directions, which means a lot of repetition. Use an audio sample no more than a minute in length (so that it doesn't take you a day to cycle through all of tuning).

mike

 

[mtindor]

Raw Audio Samples

For those of you who are experiencing better performance using DSD than DSDPlus when using a dongle + SDR# setup, please consider recording that incoming audio (from SDR#) into a 48khz .wav file. You can upload your samples to a site like zippyshare.com for free (no account needed) and then can provide a link to that audio file.

Then the rest of us can scrutinize that audio file and see what is wrong with the raw audio / give you some suggestions for tuning parameters if necessary.

It is very difficult for anybody to pinpoint a problem like that without having access to the raw audio.

How do you get a raw audio sample and upload it, you ask? Well, that is easy to do.

• 1. Run DSDPlus like you normally would, with the audio piping into it from SDR#
  
  
• 2. While DSD is running, press R to start recording a raw audio file
  
  After you have a sufficient audio sample (1 minute is good), then press R again to stop the recording of raw audio.
  
  
• 3. Find the newly created raw audio file in your directory where DSDPlus is.
  
  It will be named something similar to: DSDPlus-Raw-Input_2013-12-25@175957.wav'
  
  If you have multiple files starting with DSDPlus-Raw-Input, make sure you get the correct one.
  
  
• 4. Then go to http;//www.zippyshare.com
  
  Click Browse to browse for the DSDPlus-Raw-Input* file, then click Start Upload. Once the upload is completed, the site will display a URL that you can copy and paste in this thread so that others can download and scrutinize that raw audio.

Mike