Just trying DSD--Screenshot included--What do I need to improve?

[davidmc36]

Just trying it out for the first time.

Just garble so far with a word or two here and there.

Is this mostly due to poor signal or crummy tap?

[kma371]

first thing I always tell people is invert the audio -xr is the command

[davidmc36]

Same or worse (by ear)

[mtindor]

Something very odd about those screenshots is mostly near-zero INLVLs but spurts up to the 20s. Usually the INLVL would remain fairly consistent.

At any rate, I'd probably increase the audio input so that the average is around 25 INLVL. Right now, even on the highest INLVL figures it barely reaches 25.

You can increase that by increasing the level on your Recording setting for the input device your using.

What does your CPU usage look like during a decode?

Mike

[davidmc36]

Originally Posted by mtindor

Something very odd about those screenshots is mostly near-zero INLVLs but spurts up to the 20s. Usually the INLVL would remain fairly consistent.

At any rate, I'd probably increase the audio input so that the average is around 25 INLVL. Right now, even on the highest INLVL figures it barely reaches 25.

You can increase that by increasing the level on your Recording setting for the input device your using.

What does your CPU usage look like during a decode?

Mike

I was trying the audio fairly low as I saw something in a tutorial about not overloading it. Pumped up the volume a bit but still getting it starting at 25ish and settling out around 5-7.

CPU shows 50-70% during decode. May have to give this a shot on my dual core machine.

Just fiddling around so far to see if I could get any idea of how this works. Will have to set up something with, what do you use--Unitrunker?--to control the tapped scanner. Just scrolling through freqs looking for digital is a major pain. Makes it hard to work on fine tuning things too. Need signal more often to work with.

[mtindor]

Originally Posted by davidmc36

I was trying the audio fairly low as I saw something in a tutorial about not overloading it. Pumped up the volume a bit but still getting it starting at 25ish and settling out around 5-7.

CPU shows 50-70% during decode. May have to give this a shot on my dual core machine.

Just fiddling around so far to see if I could get any idea of how this works. Will have to set up something with, what do you use--Unitrunker?--to control the tapped scanner. Just scrolling through freqs looking for digital is a major pain. Makes it hard to work on fine tuning things too. Need signal more often to work with.

I can only compare a T5750 Core 2 Duo and an 3rd gen i7. My T5750 Core 2 Duo could not keep up with decode of P25/DMR (I don't have Provoice nearby). I don't think it is related so much to # of cores or hyperthreading (not sure DSD makes use of either), but rather the CPU rated speed / architecture.

But if it were me, I'd definitely try a multicore / faster CPU machine if I had one. 70% isn't 100%, so that's good. But it's obvious that it is working the machine a bit on decode.

As easy as it is to adjust audio, get the audio up so that it settles on 25 INLVL, not 7 INLVL. Then go back and forth between -xr and not using -xr. I'm not sure if inversion even makes a difference on Provoice.

Can you make anything out at all, or is it competely garbled all the time? If its compltely garbled all the time and you can't make out a single word, maybe it is a tap issue.

mike

[davidmc36]

Originally Posted by mtindor

Can you make anything out at all, or is it competely garbled all the time? If its compltely garbled all the time and you can't make out a single word, maybe it is a tap issue.

mike

Maybe one word or two on a 10 second or so transmission. I am leaning toward the tap since the only thing I used it for previous to a couple days ago was Motorola VHF Control Channel and in another thread we were discussing my decode of some DMR/TRBO and one guy said it looked like the log showed poor tap performance. He said the data on the Motorola VHF was much easier to decode and suggested a voltage divider. That's why I wanted to give the Provoice a shot and I think this confirms the tap is crummy.

Nothing I do seems to bring it higher than 7 or so.

Thanks for the advice on this.

Am having a bit of fun with this now and my curiosity is peaked so I think I will dive into the guts of the scanner and rebuild the tap.

[AZScanner]

I'm experiencing similar issues with DSD on my laptop and I think the problem is it's an old tired laptop with an old tired soundcard and it's not really up to the task of properly decoding P25. That and my ancient BC895 that I've been using for this is ready to become a fossil.

I'm ready to take the next step and get a couple of those SDR dongles I keep hearing about and setting up a dedicated desktop machine for the task. Rather than muck up your thread, I will start my own to seek advice for the project but it's my experience/guess that your machine just isn't up to the task - that's the problem with mine.

-AZ

[davidmc36]

Originally Posted by AZScanner

I'm experiencing similar issues with DSD on my laptop and I think the problem is it's an old tired laptop with an old tired soundcard and it's not really up to the task of properly decoding P25. That and my ancient BC895 that I've been using for this is ready to become a fossil.

I'm ready to take the next step and get a couple of those SDR dongles I keep hearing about and setting up a dedicated desktop machine for the task. Rather than muck up your thread, I will start my own to seek advice for the project but it's my experience/guess that your machine just isn't up to the task - that's the problem with mine.

-AZ

Hey, no sweat joining in.

I was just tinkering with another tapped scanner and have got the inlvl up in the high 30's and even 40's. Getting a bit of understandable audio. When I do start to hear some good audio the CPU usage shoots up to 95% or more, so I think the same as you, need more power.

Funny thing is that when I tried this tap it was even worse on DMR than the first one, go figure.

Fun messing around with it though.

[AZScanner]

Originally Posted by davidmc36

Hey, no sweat joining in.

I was just tinkering with another tapped scanner and have got the inlvl up in the high 30's and even 40's. Getting a bit of understandable audio. When I do start to hear some good audio the CPU usage shoots up to 95% or more, so I think the same as you, need more power.

Funny thing is that when I tried this tap it was even worse on DMR than the first one, go figure.

Fun messing around with it though.

I've started a thread for my project, so feel free to chime in on that one.

Something odd I noticed was that when I have an inlevel above 10% I can't get DSD to decode voice very well - maybe a word or two once in a while but that's it. However, if I lower my microphone level to 3 or lower, it suddenly starts decoding like magic. It's still hit or miss but I can hear complete sentences if the inlevel shows around 3 to 5%. This is on a simulcast CQPSK system though, not DMR. DMR sucks no matter what I try - again I think it's just the fact that I'm trying to use a computer that just can't handle it and a scanner that's older than dirt.

You're right though, it is fun to play with. Can't wait to see what I can do with decent hardware.

[awattam]

I am having the same issues with an old Uniden 800xlt and a 6 year old laptop. I didnt use a resistor at the output of the scanner tap and was going to spend some time at the weekend rebuilding the tap so see if I can get any improvement.

I was toying with the idea of adding a USB sound card to seperate the input and output hardware. I dont want to waste any more money or time on this unless I would get results. Any thoughts?

[AZScanner]

Originally Posted by awattam

I am having the same issues with an old Uniden 800xlt and a 6 year old laptop. I didnt use a resistor at the output of the scanner tap and was going to spend some time at the weekend rebuilding the tap so see if I can get any improvement.

I was toying with the idea of adding a USB sound card to seperate the input and output hardware. I dont want to waste any more money or time on this unless I would get results. Any thoughts?

I can tell you that adding a resistor to my tap didn't really change much in terms of results. I went from getting a few syllables now and then to getting a few syllables more often. That was about it. Dedicated sound cards for input and output are probably far more likely to get you better results. I would also set up a desktop machine for this purpose as laptops are notoriously cheap in the soundcard department and often have sound cards that rely on the CPU for processing rather than having their own dedicated onboard processor.

[davidmc36]

Originally Posted by AZScanner

....This is on a simulcast CQPSK system though, not DMR....

Actually the stuff I was playing around with DMR was completely different. Not sure exactly the details of this system.

Ottawa Public Safety Trunking System, Ottawa, Ontario - Scanner Frequencies

[davidmc36]

Originally Posted by awattam

I am having the same issues with an old Uniden 800xlt and a 6 year old laptop. I didnt use a resistor at the output of the scanner tap and was going to spend some time at the weekend rebuilding the tap so see if I can get any improvement.

I was toying with the idea of adding a USB sound card to seperate the input and output hardware. I dont want to waste any more money or time on this unless I would get results. Any thoughts?

I have been trying this with a USB XFI. The laptop only has a mic-in and when I hook it to the tap the first scanner basically stops receiving period. I'm guessing due to the poor quality of the tap. Looks like the second scanner will be the one to play with for now.

[jhampton2000]

>I was toying with the idea of adding a USB sound card to seperate the input and output hardware.

I have tried 5 or 6 different USB sound cards over a period of time.....the most important aspect is sampling rate: the higher the better. The Griffin iMic was the best that I've found so far at an affordable price: DSD really needs its 48kHz sampling rate. 44.1khz doesn't really cut it. But it still can't better a proper PCI-based sound card with sampling rates of 96kHz+.

Jim

[davidmc36]

Getting a bit of action now. Tried it on my better laptop and the processor only shows about 20% usage while decoding. Will have to try fiddling with the sampling rate and or try a desktop with PCI card.

[Unitrunker]

Originally Posted by davidmc36

when I hook it to the tap the first scanner basically stops receiving

You need a resistor in series. Try values 4.7 k to 47 k.

[Unitrunker]

Originally Posted by mtindor

Something very odd about those screenshots is mostly near-zero INLVLs but spurts up to the 20s. Usually the INLVL would remain fairly consistent,

Agreed.
The OP may want to verify the advanced microphone options. Make sure AGC, noise cancelling , and other microphone enhancements are turned off.

[davidmc36]

I get pretty much nothing with mic input, even with the second scanner with what seems to be the better tap. Starting to get some success with line-in on USB sound device. Not sure what the resistor is on the first scanner right now, would have to pull it apart to find out.

I get constant inlvl now with the tap on the second scanner. Right around 20 to 25 seems to get some decent voice. I think the worst problem now is inconsistent signal. The scanner always has static on the system even when analog trunking. I am quite a distance from town. Will give it a shot next time I am on night shift and will be close by the transmitter.

I really need to get something set up so I can track the TG's though so I can listen for a while straight ahead instead of scrolling through the frequencies manually. I just wanted to give it a bit of a try first to see if I would have any luck.

I guess two tapped scanners and Unitrunker to control the voice one is what's reqd if i don't have like a PSR 600 with the 1.0 bootloader?

[davidmc36]

Got the setup in to town today. WOW! Decoding at 90-95% understandable. And that is with the first scanner that the tap seemed crummy on. Showing about inlvl 5% works quite well. Just got tired of hitting the scan button all the time. Gotta get something else tapped if I want to mess around much more.

It appears that the decoding is the best when there is nothing showing up on the right side of the DSD window. The fewer double dashes and "R"'s the better it sounds. Does that sound right?

Not sure if I will try to get a better antenna up here at home or whether it is worth it since I am about 35 km from the source and the terrain is not very good for 800mhz. I am at about the same altitude as the transmitter but there is a fair raise in terrain between me and town.