D-Star Party 2012

sebphoto

http://www.icom.co.jp/world/d-starparty2012/

watty

D-Star: Icom's obsolete, patent encumbered proprietary Digital system designed to sell Icoms. Avoid. Even the internet version needs an expensive overpriced dongle.

GekkePrutser

Just noticed this topic, I'm one of the few here in Ireland experimenting with D-Star.

It's true that it's restricted by its proprietary codec, but in Icom's defense there was nothing available that was open source when D-Star was designed. DVSI's AMBE codec was the only thing that could produce something intelligible with the limited bandwidth available. The open-source codec2 is still being developed today, D-Star is from 2004. The codec is the only thing that is patented (and unfortunately not available as software, only on a DSP chip).

These days you can get a lot of custom hardware, like DVRPTR boards for much less than Icom's hardware, and even my first D-Star handy from Icom only cost me 299, which is not bad for an A-brand 5W dualbander.

Yes it is a bit behind the times in a technical sense, but it's the only real choice for digital voice right now. Unfortunately Yaesu's C4FM standard isn't much better, besides a slightly more modern modulation it still uses a proprietary codec from DVSI and it doesn't have a worldwide linked repeater network which D-Star does have. I'm happy to experiment with it, although I wish it was a bit more widely used in Ireland. On the continent it is quite big and in the UK they've even started rejecting repeater licenses because there's already too many of them.

It certainly keeps me interested in the hobby anyway. It offers a few advantages over FM, especially with the newer radios such as the ID-31.

watty

There is nothing defensible about the product. It exists solely to sell Icoms and is of almost no value.

There are 2 or 3 really good open source choices and a few very much cheaper and better non-open codec solutions.

DVSI's AMBE codec was the only thing that could produce something intelligible with the limited bandwidth available.

Untrue. I've been working with Codecs since 1978 and developing custom hardware with audio codecs since 1986

For the same EIRP it's got poorer range than FM. Nearly useless for Mobile as in poorer areas it just mutes but the FM just gets noisier.

2006 to 2009 working on 4G family VOIP and Data.

martinedwards

its another thing to tinker with.

if we all wanted the best possible output, there would be no-one operating QRP just for the challenge of it.

personally?

on my limited budget?

I'll not be joining in any time soon.

watty

1W PSK31 on the FT817ND is something I must do more often.

Connected to laptop I can do any digital mode including voice that fits on the SSB bandwidth.

http://www.speex.org/

http://www.voip-info.org/wiki/view/Codecs


If you use QPSK or QAM or APSK then there is more bits per symbol than bit rate

QPSK is 2 bits. So a symbol rate of 2k is 4kbps data. But really having FEC is best. So though QAM16 is lower noise margin, less robust, you can have 4 bits per symbol. Thus at 8kps data you can have 50% FEC for 4K voice codec.

Or you can use COFDM. If you use 8K FFT that allows about 6700 carriers. So symbol rate per carrier is 8000/6700 bits per second. You'd want a more exotic FEC that is across the carriers and time.

There is a low bandwidth version of DRM for Amateur use (which is available free and can drive a suitable radio from PC/laptop). This does AAC codec and COFDM with FEC.

LOTS of digital voice options that work with many radios, even an FT101ZD.

I'm looking at budget all-in-one solutions too with a small LCD and doing the job without a PC.

GekkePrutser

watty wrote: »

There is nothing defensible about the product. It exists solely to sell Icoms and is of almost no value.

Of course D-Star was invented just to sell Icoms, I don't doubt that.. But if you're looking for a VHF/UHF digital voice mode at the moment there is no alternative that has any way near the amount of users and repeaters.

The range is indeed worse than FM, but the benefit to me is not range. What I like about it is that you can see who's calling and from where. You can use that data to view who's been calling, even when you weren't around. And the piggyback message on every call gives it a nice personal touch.

I'm from Holland myself and I use it to keep in touch with my friends back home. It has some advantages in that it auto routes to the repeater you last worked. I often check the lastheard logs of my favourite repeaters on my phone to see if anyone I know is active.

The one thing I don't really like with D-Star is that they didn't include a data mode where you can use the entire bandwidth just for data, at the moment you have to transmit empty voice packets just to be able to use the 1200 bps that are piggibacked on top of every voice call. Which is pretty useless and nobody ever uses it. And the other way around, it would be good if those 1200 bits could be used to improve the voice quality such as FEC bits when they're not in use for anything else.

I agree D-Star is behind the times technically, but it works and it's very widespread globally. I wish there was some more activity in Ireland though.

watty wrote: »

There are 2 or 3 really good open source choices and a few very much cheaper and better non-open codec solutions.

Untrue. I've been working with Codecs since 1978 and developing custom hardware with audio codecs since 1986

For the same EIRP it's got poorer range than FM. Nearly useless for Mobile as in poorer areas it just mutes but the FM just gets noisier.

2006 to 2009 working on 4G family VOIP and Data.

I read somewhere that they tried different codecs, such as iLBC but they didn't perform as well (and even the AMBE2020 doesn't work all that well, a lot of HAMs are taken back when they hear D-Star the first time). But I'm not a codec guy I'm more interested in the networks behind it. I'm not saying D-Star is the best ever, but what I like about it is that you can take the radio out of the box, turn it on and talk to someone. But there is loads of room for improvement (and the lack of FEC also annoys me a lot, especially the scrambled text messages).

By the way it sounds like you have a lot of experience to develop a digital mode that would work better, if you have something I'm always interested, I like digital modes. I think there'd be enough uptake on a new mode if one was developed.

watty

iLBC needs low packet loss TCP/IP. Speex is developed to be more robust without packet loss.

For speech you don't want the overhead of TCP/IP on RF link. There are other problems with TCP/IP too. So best approach over RF for Data is to "spoof" TCP/IP and have a kind of software "transparent proxy/router" at each end like Satellite Internet connections use.

You absolutely need FEC over RF. DRM, DRM+, DAB, DAB+, DMB, DVB, DOCSIS etc ALL use FEC.

DOCSIS is a good choice for the RF link as it's multiplexed DVB MPEG2-TS on downlink so 1 to many clients in two way or broadcast mode with or without TCP/IP can be used on one end (either a Peer to Peer "master" or a Repeater Node). The uplink is a choice of CDMA (bad and not used anymore) or QPSK, QAM16, QAM32 QAM64 in timeslots per client. So if one connection you can have all bandwidth. The timeslots can implement "virtual channels" so uplink can be say be QPSK for DX or QRP clients and QAM64 for a high power or closer client.

Downlink is DVB, so only can use same QAM or QPSK for all, QPSK or QAM16 to QAM2048

I have demoed commercial VOIP + Data prototypes using narrow band DOCSIS, rather than the normal 6MHz or 8MHz channels. You can actually use ANY symbol rate on uplink or downlink (thus bandwidth), though the Commercially deployed systems on RF (rather than normal Cable Broadband) are 2, 6, 7 or 8MHz downlink and 0.8, 1.6, 3.2 and 6.4MHz uplink. Channels can be bonded.

So the SAME protocols and codec etc can be a single user at a time in 3KHz, 12.5 or 25kHz channel, (4kbps to 50kbps after FEC) or for multiple simultaneous Voice + data on up to 5 x 8MHz channel (say GHz bands) and aggregate bonded 360Mbps.