Digital Speech Decoder 1.2 and mbelib 1.1 released

[cpuerror]

Well I have been messing with this all evening trying to decode P25P1. I have a disc tap coming off a 246 and feeding into a Signalink USB box. But all I get are little blips and boops and I have the modulation locked in at C4FM. It could just be my setup but it just doesn't want to work for me.

 

[KC1UA]

I'm in the same boat as the above two posters, but I'm impressed. It's a great start, and from the remarks of others very likely my system coming up short. I'm using both an IC-PCR1000 and AOR AR5000A+3 discriminator tap into a Creative Labs Soundblaster Live! 24 USB device and running Ubuntu Linux 9.10 on a pretty fast machine with a lot (4 GB) of RAM. Unfortunately I'm restricted to monitoring a few conventional P25 stations here on Cape Cod, nothing trunked, and one of the two simplex in nature (Massachusetts Military Reservation).

Very intriguing that this is progressing, and very encouraging. I continue to play with it and watch this thread with great interest. Thanks for your efforts.

 

[gary123]

I have had some success with DSD. I have found that having the input level about 2 steps or 20% above zero on the mic input on the computer is best. I have been checking against Fleetnet and 2 conventional systems. I am getting about 60% recovered audio. I know that DSD has some auto levelling incorporated into it, perhaps an option to have a pane open that allows adjustment of the inputs levels (actual input level, and C4FM detect/skew) while showing the output would help in tweaking the aplication.

Another issue on the recovered audio might be that the sound card is having trouble upconverting the audio to the 48000/16k sample rate that DSD outputs. Maybe another pane that shows the current outputs available on the soundcard so that the user can select for the best recovered audio?

I am more of a windows person so Im not sure how much of this is available in linux. I do think that this is one of the best software programs going and am willing to contribute where I can to its developement.

 

[woodpecker]

I'm in the same boat as the above two posters, but I'm impressed. It's a great start, and from the remarks of others very likely my system coming up short. I'm using both an IC-PCR1000 and AOR AR5000A+3 discriminator tap into a Creative Labs Soundblaster Live! 24 USB device and running Ubuntu Linux 9.10 on a pretty fast machine with a lot (4 GB) of RAM.....

I tried a SB Live! 24 USB with Ubuntu but couldn't control the input volume, although I had a mic input slider it would do nothing, any tips how to get it going, I am not too good with linux.

At the moment I am back to the on-board sound, I've found a trbo DMR signal that's quite active, I am managing to understand about 50% of the voice, the rest sounds garbled, I notice the Demod mode is flicking between GFSK and C4FM a bit.

I've found that signal strength is quite critical as is the input volume signal level, I am wondering if I could get the SB Live! 24 USB going that it might improve a bit a bit more, not sure if soundcards make much difference?

Overall though I am highly impressed with the software and the ability to decode trbo is a fantastic achievement, thanks to DSDAuthor!

 

[scanpprcn]

The past four or five posters, I feel your pain...

I've played around lightly with linux esp, Ubuntu, so this has been an 'effort' for me. What I found helps is to adjust the volume input gain using the following:

1) Open a terminal Window
2) Type "alsamixer" <return>

Use the tab key to jump between Input and Output and All (both). It operates much like the Windows mixer that we're probably familiar with.

To the experimenters (like cpuerror) who are not recovering audio that is understandable at all, the input adjustment seems to be critical. Like gary123 said (and dsdauthor) the gain should not saturate the input of the sound card.

I actually put together a brand new higher performance machine for this decoder project only to find a week later that dsd runs just fine with xubuntu on a 866MHz PIII with the built in Intel motherboard sound.

 

[gary123]

another simple way to adjust the audio input is to select system (top toolbar) then preferances then sound. then go to the input tab and check the level there. You can adjust the level while DSD is running.

I have the best operatation when the input indicates 65-75%.

 

[woodpecker]

The past four or five posters, I feel your pain...

I've played around lightly with linux esp, Ubuntu, so this has been an 'effort' for me. What I found helps is to adjust the volume input gain using the following:

1) Open a terminal Window
2) Type "alsamixer" <return>

Use the tab key to jump between Input and Output and All (both). It operates much like the Windows mixer that we're probably familiar with.<snip>

I downloaded and compiled the latest alsa drivers, lib and utils. I have 2 USB soundblasters, (trying one at a time) one is the MP3+ SB0270 the other a SB Live ! 24 SB0490, I am struggling with both, the on-board sound works but is currently disabled. With the Live ! 24 alsamixer will let me adjust the mic volume ok, but I cannot get any playback sound to the headphone jack, the only option for playback in alsamixer is PCM, which I presume is the optical out? So how do I get audio from the headphone jack?

With the MP3+ I get playback sound ok but the mic slider will not adjust the input at all, at 0 it is muted, from 1-100 it is just at 100%. If I try and set the line in as the capture I cannot any sound in at all.

I would like to get one of these USB sound cards working, any help would be appreciated.

TIA

 

[scanpprcn]

Gary123, your findings are interesting. What chipset or sound card are you using? With the four or five sound cards that I've tried, (VIA, SB!, Intel, Crystal) on three machines (Athlon X2, P4, PIII) my best performance for recovered audio has been around 30%-40% C4FM.

woodpecker, on the first machine I built I put in a SB card to use as an input and attempted to use the motherboard audio as an output. dsdauthor indicated that the loading would be minimized using two cards. As you are now experiencing, I found it difficult to find the correct 'mixing' to make this happen. I also tried disabling the on-board audio from BIOS and simply running from the SB card. I actually found that the motherboard chipset had better audio.

I know I'm probably not any help with the audio issue/question, but for me it would be like the blind leading the blind. If I got on your machine I could probably play around in the mixer settings till I got it, but...

What I did do, which I now offer, is to advance all of the sliders in alsamixer (input and output) up to the first or second red block since I didn't know what affected what. Then, I went to the obvious input selectors (line1, line2, mic) and varied those to note any changes. Also, I toggled MIX setting between its various selections till I was decoding. I did find out that in the alsamixer, the "m" key mutes whatever you're selected on and the "space bar" selects an object to be acted on.

Also, when using alsamixer you have to put a switch in the command line to operate on the correct SB card, if you have more than one.

Example:
alsamixer -c2

To know which cards are available and what switch use, look in the /dev/ folder. There should be a icon labeled "sound" or "sound2" (or both). You can and should run two instances of alsamixer if you have two sound cards. So, open two separate terminal windows. In one window type "alsamixer" which will operate on the "sound" icon. In the other terminal window type "alsamixer -c2" to operate on the "sound2" device.

 

[K8TEK]

I tried a SB Live! 24 USB with Ubuntu but couldn't control the input volume, although I had a mic input slider it would do nothing, any tips how to get it going, I am not too good with linux.

At the moment I am back to the on-board sound, I've found a trbo DMR signal that's quite active, I am managing to understand about 50% of the voice, the rest sounds garbled, I notice the Demod mode is flicking between GFSK and C4FM a bit.

I've found that signal strength is quite critical as is the input volume signal level, I am wondering if I could get the SB Live! 24 USB going that it might improve a bit a bit more, not sure if soundcards make much difference?

Overall though I am highly impressed with the software and the ability to decode trbo is a fantastic achievement, thanks to DSDAuthor!

I'm not trying to be rude, but the sound blaster live is junk. Anything made by Creative is garbage.

 

[KC1UA]

I've been happy with the Soundblaster Live's for scanner feeds, but apparently this application is not thrilled with them. I'm back to trying my built in sound. I also have an M-Audio 8 port card installed but Linux doesn't see it properly (yet). I also have the el-cheapo USB sound devices that look like a flash drive, may try one of those for the heck of it if Linux will see it. Another issue for me is the lack of consistently strong P25 signals to work with. Continuing to plug away here. Fun stuff!

 

[woodpecker]

I'm not trying to be rude, but the sound blaster live is junk. Anything made by Creative is garbage.

What is the basis for this statement?

What USB soundcard do you recommend for Linux then?

Just to say they are junk is not helpful at all!

This article says the SB Live 24 is quite good on noise performance:-

M0KGK SDR soundcards page

 

[gary123]

Gary123, your findings are interesting. What chipset or sound card are you using?

for DSD im using a del laptop with a pentium "M" at 1.6gig, on board memory is 512Meg. The sound card is a Intel 8280 using a sigmatel stac9750 processor.

I have a buffered input to the sound card from a bc796xlt. The systems I am monitoring are all 3600baud. One is fleetnet. others are conventional. I am right on top of a fleetnet site (2 miles away) and right under one of the conventional transmitters (1/8th mile). I do monitor another conventional system (20 miles) and have good decode on it even though it is only 1/2 half signal strenth on the 796.

 

[woodpecker]

I've managed to get about 95% voice ok now from a trbo system which is at an airport about 30 miles away, the signal level reads ~-104dBm on my Icom 9500 so its not exactly strong. I am now using the SB Live 24 USB as the input and the MP3+ USB as the output. For trbo I am finding running with the -mc option to force C4FM optimisation works best.

In between the idle bursts I am seeing a lot of Unknown burst type: 1101 messages, anyone know what these are, they seem quite frequent so I don't think it is a decoding error:-

Sync: +DMR mod: C4FM input: 43% 011111111101110101110101 110111111111010101111101011101011101111101011101 slot0 [slot1] Unknown burst type: 1101

 

[dsdauthor]

I've managed to get about 95% voice ok now from a trbo system which is at an airport about 30 miles away, the signal level reads ~-104dBm on my Icom 9500 so its not exactly strong. I am now using the SB Live 24 USB as the input and the MP3+ USB as the output. For trbo I am finding running with the -mc option to force C4FM optimisation works best.

In between the idle bursts I am seeing a lot of Unknown burst type: 1101 messages, anyone know what these are, they seem quite frequent so I don't think it is a decoding error:-

Sync: +DMR mod: C4FM input: 43% 011111111101110101110101 110111111111010101111101011101011101111101011101 slot0 [slot1] Unknown burst type: 1101

The "burst type" field describes what kind of data non-voice frames contain. the DMR spec defines several possible values from 0000 - 1010 (see dmr_data.c for the list). That could be some Motorola proprietary burst type.

 

[woodpecker]

The "burst type" field describes what kind of data non-voice frames contain. the DMR spec defines several possible values from 0000 - 1010 (see dmr_data.c for the list). That could be some Motorola proprietary burst type.

Ok thanks for that, I've found 2 more trbo channels this morning, both are borderline on signal strength, is there any way to improve weak signal readability with some soundcard DSP??

Even more exciting for me in the UK was to find a P25 P1 signal, I had DSD forced in C4FM, not sure if that is the correct modulation for P25 but I still got decode of nac and tg, will see if I can recover some audio from it, it doesn't seem to be very active though.

I am finding with trbo that it seems to need more signal than an FM analogue signal would need to make it readable, is this normal?

 

[K8TEK]

What is the basis for this statement?

What USB soundcard do you recommend for Linux then?

Just to say they are junk is not helpful at all!

This article says the SB Live 24 is quite good on noise performance:-

M0KGK SDR soundcards page

Get a real sound card, not some USB junk. Get something made in the last decade preferably, not something based on 1998 technology. Look at the specs like SNR, higher is better. My point is Creative's "Live" brand is two generations behind.

By the way, your link proves my point. It was published over 3 years ago!

 

[woodpecker]

Get a real sound card, not some USB junk. Get something made in the last decade preferably, not something based on 1998 technology. Look at the specs like SNR, higher is better. My point is Creative's "Live" brand is two generations behind.

By the way, your link proves my point. It was published over 3 years ago!

Just because the interface is USB does not make a card junk, you clearly don't have a clue, there are pro audio cards that use USB.

You've still quoted no recommendation for a good card for use with Linux. All you have done is bashed Creative and added nothing positve to this thread, if you've nothing to recommend then don't bother posting again!

 

[gary123]

Ok thanks for that, I've found 2 more trbo channels this morning, both are borderline on signal strength, is there any way to improve weak signal readability with some soundcard DSP??

You could concider getting a rf preamp. Im using a commercial amp on some bands fo firm up signals. This might be a bit overkill in heavilly conjested rf areas. Typically a good commercial rf preamp will run about $100.

 

[scanpprcn]

I hesitated posting this cause I probably don't know what I talking about, but I use a MotoTRBO radio on a local system. The radios are programmed for repeater TDMA, simplex digital TRBO, and a couple of analog channels.

I haven't tried decoding from the repeater because it's far away from me but I have a handheld at home. I tried the simplex digital TRBO and didn't see any response whatsoever from dsd.

Since I have the C4FM (P25) nailed down I figure I shouldn't be far off from decoding TRBO, but I wonder if the simplex digital mode of TRBO is unlike the repeater 2-time slot TDMA.

I'd have to drag a linux desktop out in range of the repeater, which is a chore. So, before I do that I was wondering if anyone knew if I should be decoding the digital simplex form of TRBO just the same as the 2-time slot TDMA, repeated form?

 

[K8TEK]

there are pro audio cards that use USB.

Then use what the pros use, be prepared to take a big hit in the checkbook.