DSD 1.4 and mbelib 1.2.3 released

[44922035]

Thanks dsdauthor,

This is a great piece of software. I am wondering about vocoder support for D-Star?

I understand the AMBE chip has two modes when operating at 3600 bps, one using block FEC and the other using convolution FEC.

I know D-Star uses the block FEC option, and it may be that the Mototrbo chose the convolution FEC.

 

[dsdauthor]

Thanks dsdauthor,

This is a great piece of software. I am wondering about vocoder support for D-Star?

I understand the AMBE chip has two modes when operating at 3600 bps, one using block FEC and the other using convolution FEC.

I know D-Star uses the block FEC option, and it may be that the Mototrbo chose the convolution FEC.

Unfortunately D-Star uses an older version of AMBE that is not published anywhere that I am aware of. D-STAR is absolute crap anyway, hopefully hams will migrate to something more sophisticated and professional like P25/NXDN/TRBO.

 

[dsdauthor]

is the any help that could help us to understand frame type like :

Sync: +DMR o: 24 mod: GFSK g: 25.000000 inlvl: 2% slot0 [slot1] CSBK
Sync: +DMR o: 24 mod: GFSK g: 25.000000 inlvl: 3% [slot0] slot1 CSBK
Sync: +DMR o: 24 mod: QPSK g: 25.000000 inlvl: 2% slot0 [slot1] CSBK
Sync: +DMR o: 24 mod: QPSK g: 25.000000 inlvl: 2% [slot0] slot1 TLC
Sync: +DMR o: 169 mod: QPSK g: 25.000000 inlvl: 2% [slot0] slot1 RATE 1/2 DATA

what is CSBK , TLC

thanks for answer

Your audio input level "inlvl" is way too low (2%). It should ideally be between 40-70%. If you still don't get good decode try the -xr option to invert the signal. You also should ensure that you have a good signal and your discriminator tap is not restricting the signal bandwidth.

 

[drbecker]

The Wiki Page has been updated with links to the latest versions. Special thanks to js_scan888 for figuring out the pseudorandom bit sequence required to get NXDN working. This is why I love open source software and include code for unfinished formats so others can tinker with them.

I am still waiting on confirmation that the ProVoice EA sync added in version 1.3.1 actually works.

DSD 1.4
New features:
Decodes voice from NXDN 9600 signals
NXDN96 frames enabled by default
Improved resistance to NXDN96 sync false positives

Fixed bugs:
.wav file header updated after playing .imb/.amb data files
.imb/.amb files now have correct tgid in filename

mbelib 1.2.3 Fixed bugs:
Cik array in mbe_decodeAmbe2250Parms was too small

Dear dsdauthor,

BRAVO BRAVO BRAVO!!!
Thank-You SO much for this software!!! I am decoding about 40% of Mototrbo but i think my computer
is to slow. With that being said, can you please let me know what type of computer, sound card(s), cpu, memory, etc that you are using to decode with please? I am going to build a new box and I would like to replicate EXACTLY what you are using. I am currently running your software with the 4gig Scandisk usb
disk. I would like to build a box to your exact specs please. I am going to dedicate the new computer to decoding MOTOTRBO
Thanks Much!
Doc

 

[philacop]

DSD NXDN Digital trunking

Hello all

Does the program decode nxdn 9600 baud digital trunked systems? I tried to use it on a 4 ch trunked system & i get no data or info on screen or no voice decode. i can decode trbo & p25 with no problem.
maybe someone has had sucess with nxdn 9600 trunking channel decoding.
thanks

 

[jcardani]

I know philacop and just to clarify, he is asking if DSD can decode NXDN NEXEDGE 9600 trunked voice channels (not just NXDN conventional voice channels). He is not talking about decodding the trunked control channel.

Has anyone had success doing this?

BTW there is a new casino in Philadelphia that uses a 4 channel UHF NEXEDGE trunked radio system.

thanks!

Joe

Hello all

Does the program decode nxdn 9600 baud digital trunked systems? I tried to use it on a 4 ch trunked system & i get no data or info on screen or no voice decode. i can decode trbo & p25 with no problem.
maybe someone has had sucess with nxdn 9600 trunking channel decoding.
thanks

 

[dsdauthor]

Hello all

Does the program decode nxdn 9600 baud digital trunked systems? I tried to use it on a 4 ch trunked system & i get no data or info on screen or no voice decode. i can decode trbo & p25 with no problem.
maybe someone has had sucess with nxdn 9600 trunking channel decoding.
thanks

Are you using dsd version 1.4.1? version 1.4 did not have the sync types for NXDN repeater output. Also you must use the -fi option if it is 6.25khz system as DSD can't auto-detect that symbol rate. If it is 12.5khz system then you must not use the -fi option.

If it still doesn't work after that then capture the signal with your discriminator tap at 48000hz/16bit/mono .wav file and upload it somewhere that I can look at.

 

[ilgrant]

Disregard

 

[c5corvette]

It is also highly desirable to use a serial control cable and the -R option to skip over the channels used for GPS data.

Can you elaborate?

 

[c5corvette]

Under Ubuntu, type 'aplay -l' to see a list of the audio devices available. ( 'l' = lowercase L )

The list will include something like
card 0 <information>
card 1 <information>
card 2 <information>
.... card n <information>.

Usually there is enough information listed for each card that you can tell which one is which.

Then, if your input card turns out to be card 2 and your output card is card 1, enter

dsd -i /dev/audio2 -o /dev/audio1

The default is card 0, so if one of your cards is card 0, let's say the output card, then just enter

dsd -i dev/audio2

DSD will default to -o dev/audio0, so you don't need to type that.

To use DSD, I assume only one sound card is required and that I can use line-in for the discriminator audio and I can use line out for powered speakers (or anything else I may want to direct the synthesized speech into) - correct?

(I assume people are using mutiple sound cards in cases where they are also using linux to stream the output, or do something else that requires another sound card?)

 

[offsite]

Using two soundcards

Hi C5,

Dug back thru the history of DSD releases and found an early post by dsdauthor. This is probably the best rationale for using separate soundcards for input and output instead of just the onboard audio chipset for both:

http://forums.radioreference.com/tr...ital-speech-decoder-mbelib-2.html#post1282110

Hope this is useful.

-rb-

 

[dsdauthor]

Can you elaborate?

To scan systems that use X2-TDMA you need to scan the channels in conventional mode. A Trunktracker will not recognize the TDMA trunking extensions and automatically switch to those channels.

When scanning in conventional mode you will want to skip non-voice channels like control channels and channels used for P25 packet data. Otherwise the scanner may stay on an unwanted channel for an extended period of time and you will miss voice transmissions on other channels.

Loudoun (and possibly PG) have enabled GPS data transmissions from certain radios (portables only in Loudoun from what I heard). These updates use the channel assigned to data transmissions. When there are no data transmissions the repeater transmits continuous "idle" frames (P25 TDULC) so there is always something happening on the data channel. The problem is the data channel assignment changes roughly every 30 seconds and rotates between all available channels that can be used for either voice or data.

This makes systems like Loudoun essentially unscannable without some method to skip over the data channel. DSD has the ability to send the scanner a "resume scan" command using a computer serial port, scanner control cable and scanner with PC control port.

The -R option for DSD enables this feature, -B lets you set the baud rate of the serial port, -C lets you set the serial port. From the help page (dsd -h) and the README file:

Scanner control options:

-B <num> Serial port baud rate (default=115200)
-C <device> Serial port for scanner control (default=/dev/ttyUSB0)
-R <num> Resume scan after <num> TDULC frames or any PDU or TSDU

-R10 is usually works well. If you set -R too low you may resume scan on voice channels between voice transmissions when you don't want to.

 

[IanWraith]

Hello

I have been playing with the code in my Java port of the DMR parts of DSD. From what I have read 4FSK demodulators usually include a root raised cosine filter for pulse shaping so have been experimenting today. For digital filter design I use this excellent web site ..

Interactive Digital Filter Design

with the following characteristics ..

// filtertype = Raised Cosine
// samplerate = 48000
// corner = 2400
// beta = 0.7
// impulselen = 21
// racos = sqrt
// comp = no
// bits =
// logmin =

The resulting Java code can be seen here ..

gist: 621038 - GitHub

I tested my code against a recording of data from the discriminator of my AOR3000. Prior to adding this filter my code could detect 20 DMR frames and after adding the filter it could detect 23 frames. So not a huge improvement but I think a worthwhile one.

Note that while the code can easily be converted into C from Java it can only be used for 4800 baud 4FSK and would need switching off when demodulating other modes.

I hope the code is of use to someone and if anyone can improve on it please post to this forum.

Regards

Ian

 

[KA1RBI]

// beta = 0.7

Just curious, where did this number come from ? AFAIK, the proper value of excess bandwidth for C4FM / P25 use at 4800 is 0.1 ......

Note that while the code can easily be converted into C from Java

That is a great site - it generated C code for me directly - would seem as though something like this would be a worthy addition to DSD..... It might be interesting to allow user choice of "beta" from a short list of options...

73

Max

 

[IanWraith]

Just curious, where did this number come from ? AFAIK, the proper value of excess bandwidth for C4FM / P25 use at 4800 is 0.1 ......

Nothing scientific I'm afraid. I looked everywhere for a correct value for beta but couldn't find anything. I then went through every value of beta between 0.1 and 0.9 in 0.1 increments. The value of 0.7 gave the best improvement when run against my test recording so I used that.

That web site is very useful though. Sadly the academic at York University who created it died 10 years ago ..

Tribute to Tony Fisher

Regards

Ian

 

[MSM_Maria]

About RRC and alpha parameter

Hello KA1RBI

Maybe You are wrong, and proper value is 0.2 or very near about this.

In general, if You know what type of filter RRC, You can get the alpha value is very close to the truth.

Please look images.

A good visible NXDN has alpha not 0.1 but very really alpha = 0.2 or near.

We write small article about problem with output from discrimanator and program decoding/pocessing An output from discriminator and program processing. I hope it was not in vain.

 

[KA1RBI]

Hello KA1RBI

Maybe You are wrong, and proper value is 0.2 or very near about this.

Right you are! RD-LAP at 9600 baud was the 0.1 one.

 

[MSM_Maria]

Hi, KA1RBI

Maybe, I just do not know. But it is easy to verify, just need an record RD-LAP in IF or I/Q mode, and check it. It is not difficult.

 

[Comint]

Sorry to have to ask, but what program do you use to put the comments on your images?

--
Comint

 

[MSM_Maria]

Hello Comint.

Welcome to the high-quality signal analysis. => Signals Analyzer`s (SA) description in English
Our blog => Signals Analysis

p.s. Sorry for small off