Digital modes are all the rage these days in amateur radio — hams are using protocols like WSPR to check propagation patterns, FT8 to get quick contacts on many bands with relatively low power, and MSK144 to quickly bounce a signal off of a meteor. There’s also digital voice, which has a number of perks over analog including improved audio quality. However, the major downside of most digital voice modes, at least those in use on UHF and VHF, is that they are proprietary with various radio brands having competing digital standards. To get above the noise a more open standard can be used instead.
The M17 standard, originally created by [Wojciech Kaczmarski] aka [SP5WWP], uses Codec 2 to convert voice into a digital format before it is broadcast over the air. Codec 2 is an open standard unlike other audio codecs. M17 also supports reflectors, which can link individual radios or entire repeaters together over the Internet. While you can make purpose-built modules that will interface with most standard radio inputs, it’s also possible to modify existing hardware to support this standard as well. The video below from [Tech Minds] shows this being done to a radio with only a few hardware modifications and the installation of a new firmware.
For anyone who has been frustrated that there’s no real universal standard for digital voice in VHF and above, M17 could be a game-changer if enough people get tired of their friends being on other proprietary digital systems. There’s plenty of supported hardware out there that most hams probably already have already, including a number of TNC devices like the Mobilinkd and the DigiRig, so it shouldn’t be too hard to get started. If you’re more into networking over radio, though, take a look at this method for sending high-bandwidth IP networking over the UHF band.
While i appreciate the technology, to my ears the digital voice is less intelligable than standard FM or SSB with a bit of noise.
Is there any benefit to using digital voice if you’re working with a bandwidth of no more than 3khz? Is the digital signal able to punch through more noise than SSB?
With a width of 12khz like with NBFM, you can fit in a DRM signal pretty well and DRM at that bandwidth sounds quite decent, allowing 30something kbps AAC, sounding slightly better than 12khz wide AM on shortwave with fading.
I believe digital wins when it comes to signal-to-noise ratio. Also digital can transfer.. well, digital information such data and when in QSO your call name, call group etc.
M17 is designed for VHF and up, with a channel width of 9kHz. The Codec 2 bit rate is 3200 for ‘full’ voice wide, or 1600 for simultaneous voice and arbitrary data.
Check out this video that compares analog FM with DMR, I think you will see how digital is clearly superior to analog.
https://www.youtube.com/watch?v=NZRYqEv8Abc
Common kinds of digital voice aren’t as good as the best. The better digital modes definitely punch through more noise than SSB, by the numbers. The difference between analog noise and imperfect digital voice is practice; heavy analog noise deafens my ears but with digital noise reduction or digital audio, I can actually listen and interpret what I hear.
Correct me if I’m wrong.. but I think that digital modes that involve plugging a TNC into an analog FM radio are going to have basically the same range as analog FM because you are basically turning your digital signal into audio and wrapping that audio in analog FM. It’s like running PSK over a repeater, good for practice and maybe fun but not with the same benefits enjoyed on FM.
Modes where a TNC goes into an “analog” SSB transciever on the other hand are basically (and I am simplifying here) just shifting the frequency of the baseband from the TNC into RF and dumping that on the air. So you get to benefit from the longer range and better noise immunity that come with lower bandwidth.
Besides all that though, there is more here than just a digital voice codec. There seems to be some sort of framework with this for sending data and linking to the internet. That could incluodesome fun stuff to play with.
What you’re missing in the use of digital modulation (over FM and over SSB) is that the recoverable information is superior to the native modulation, mostly because of the inclusion of Forward Error Correction. At the point a “native” FM or SSB signal is unusable, a data mode over that same signal is often usable because the computer (data mode) can reconstruct the original information by making use of the redundant bits that were transmitted.
How do you think things like DMR work? You have a magic chip that turns normal audio into digital audio then sends that to a standard RF section. This has the added benefit of being about to bypass the digital section to receive standard FM.
DMR is “proprietary” (I don’t think really) but considering how widespread its use is and how many open source implementations are for it, it isn’t closed at all.
And the most imporant: there are plenty of commercial radios supporting DMR (the big mass).
I believe it is proprietary in that it requires a DMR decoder chip that’s single source (and a black box). You can’t run a DMR transceiver on a free software stack on a general purpose CPU for example like you can with the open source codecs mentioned.
If DMR was open source you would see it as an option in all the cheap China radios because it would be a simple software coding job to add support. Instead it requires they buy the chip to support DMR and that raises the price a fair bit.
It is not black box. DMR is described well in ETSI and doesn’t need any specific hardware to decode.
Proprietary isn’t always closed source.
You need an AMBE2+ Codec chip for DMR. Which is a black box and totally proprietary and closed source.
My PC can decode DMR just fine by software alone. If there is any chip requirement then it must be for the encoding side.
You only need ambe+2 from dvsi if you want to be compatible with existing radios. The (open) ETSI standard for both DMR and dPMR does not specify any specific voice codec, its a commercial choice of the manufacturers to use ambe, there are plenty of alternatives out there, including codec2.
To add any software that’s used for decoding DMR without an AMBE2+ codec chip is using a reversed engineered version of the codec, which is a legal gray area since its not properly licensed like if you bought the chip. Still totally proprietary.
No you don’t. See other comments too.
The Chinese DMR radios don’t use the AMBE chip, the codec is in the firmware.
The Chinese wrote their own implementation of the codec which works ok but is not quite the same as the proprietary codec chip.
Jason VK7ZJA ( sadly now silent key), compared the audio when using the proprietary chip in a Motorola radio, with the cheap Chinese radios and found the Chinese implementation yielded better quality audio
There is an open source implementation of the codec on GitHub which is used in various Android and iPhine apps, as well as various PC applications
Also. As has been posted by other people. The ETSI DMR standard is free for anyone to use and does not specify the audio codec
The reasons people continue to use DMR with the AMBE codec is because all the major DMR networks will only pass traffic encoded with the AMBE codec.
I heard that the FreeDMR network considered supporting non AMBE audio, but no one every got around to implementing this
Did Jason ever publish the results of his comparison anywhere?
Isn’t there a digital radio standard for broadcast in Europe?
DRM? Digital Radio Mondiale? Yes.
It’s interesting and reading news stream (Journaline) with the free software is possible.
*But* the voice codec isn’t free. That’s why official releases of Dream don’t have it. You need to compile it yourself if you want to hear something. Or find an unofficial binary online.
Same situation with D-STAR. It was invented by some Japanese dude (?) and made completely free. Except for the voice codec. Which is proprietary, of course, because people were lazybback then and oriented on the commercial digital voice standards. That’s an universal excuse, by the way (“but but it was normal to use commercial available chips for that. We didn’t do anything wrong. Everyone was doing it.”) *sigh* 🙄
Some people simply don’t live up to the ideals of amateur radio. They think they’re real hams, but betray our ideals.
Let’s imagine FM was patented and required a license fee to be used. Or if FM chips were made by one company only. It would have caused an up cry.. It wouldn’t be free and violate the principles of ham radio.
In case of FM, hams would have had at least tried to develop something compatible, but free. The counterpart would have been PM here, in our thought experiment.
Unfortunately, something like this wasn’t even considered when DMR or D-STAR were designed. They have no option for an alternative codec, because their fathers were so limited in their thinking back then. They didn’t even consider proprietary voice codecs to be a problem. They blindly obeyed industry. Such a shame. Hams should lead technical progress, not be mindless consumers. 😔
DMR was designed by ETSI as a standard for commercial radio systems. So using a propietary standard wasn’t a problem. Then some HAM started to reprogramming the commercial radios to work in amateur radio bands. Using DMR has some quirks because is geared to private commercial radio uses.
There are radios more ham friendly, but they still have a commercial mode and are approved for that use.
DRM it’s a digital LW/MW/SW standard used by some broadcasters, but isn’t widely used, anoother thing.
FM _was_ patented, and did require a license. It’s just that it expired long ago. Most of your radio and all its component parts were patented, and many of them (many of the IC’s) still are made under a bunch of licensed patents.
Most of this is hidden from you within the purchase price of the components. It’s still there.
Is it still radio if you need several billion dollars of connected infrastructure between the participants?
Said about the internet.
M17 has both RF and over the internet. You can setup a repeater and also have simplex communication on it.
Internet infrastructure not absolutely required for DMR. I use RF to get to the local DMR repeater. Talk with other users on the same repeater who access with a direct RF link.
Yes it is still radio it is just the advancement of technolegy. Some of the ham community are still stuck in the vaccume tube and spark gap. There are plenty of modes using the internet, I am so tired of the excuse what if the internet goes down, show me when that has happened ever.
Just because it uses IP doesn’t mean it requires the Internet. You can link repeaters or hotspots with whatever you want, including additional radio links or a patch cable.
There is no requirement for billions or event thousands of dollars of infrastructure.
Hams are ALLOWED to use proprietary encodings? I’m honestly surprised. Seems contrary to the ethos and contrary to the other ways they’re restricted.
100% agree. I’m a ham and was excited about D-STAR when it first came out … until I learned about the propriety voice codec. That totally killed my interest and I haven’t been tempted by any similar digital radio since. M17 may finally pull me in.
When D-STAR was developed, there were no open source codecs capable of generating usable, low bit rate voice to data streams. An inexpensive, commercial, and yes proprietary, solution was available. At the time, anyone could purchase an AMBE chip for less than $20 and roll their own D-STAR equipment.
M17 is a terrific experimentter’s project. CODEC 2 is a great achievement allowing more amateur and commercial development in digital voice. For the time being, none of that is available as an off the shelf, ready to go communication system for amateur radio.
That may happen one day. The M17 project is happy to provide all their specifications to anyone who wants to use it for commercial development. Anyone choosing that will have a steep climb to compete with the three other, incompatible, digital voice modes currently in popular use.
I’d love to see it. I’m not interested in the experimental aspect myself, but I’m glad other people are.
K4AAQ
It’s cheating, I think. The people responsible do argue with complying to accepted standards. But it’s really like using a Flash plug-in in a 100% open source web browser and call it open source.
As long as there is a reasonable way for Amateur Radio transmissions to be received / decoded, yes. It’s called experimentation and it’s one of the many, many cool things about Amateur Radio. In the case of encoding voice using the DVSI chip(s), you can easily receive / decode transmissions by using another inexpensive DVSI chip. Another example is the increasing use of LoRa in Amateur Radio – LoRa is proprietary, but LoRa transmissions are easily decoded by using another inexpensive LoRa radio. Yet another example is the use of Pactor 2, 3, and 4 – all proprietary encode / decode, but can easily be monitored with off-the-shelf Pactor 2, 3, and 4 units.
If you were using a encoding technique, and did not publicly disclose that technique, or it wasn’t easily obtainable, that’s called obfuscation, and that is not legal in Amateur Radio. But if you do publicly disclose the encoding technique, and anyone can use that information to receive / decode your transmission, great. 30 years ago a network that I was part of figured out that we could pass more traffic faster if we compressed our data prior to transmitting it. If you received those transmissions, but did not have access to the compression algorithm, that’s (effectively) obfuscation, even though that wasn’t our intent. But once the compression technique is publicly disclosed, no problem.
A recent example of this approach is the NinoTNC implemented a new, unique forward error correction system called IL2P. If that was used with no public disclosure, uh oh. But IL2P is very publicly disclosed – https://tarpn.net/t/il2p/il2p.html. So, no problem.
“… uses Codec 2 to convert voice into a digital format before it is broadcast over the air.”
While I’m sure this was meant in the familiar sense, for those unaquainted with amateur radio, this remark bears clarification. Hams don’t “broadcast.” They transmit. There is a regulatory distinction between the two.
Broadcast technically means “from one source to many destinations” which applies here. So yes – hams broadcast.
Except previous poster is making a legal point, not a technical one.
For US hams, part 97.3 (a) 10 defines: “Broadcasting. Transmissions intended for reception by the general public, either direct or relayed.”
Amateur radio operators are not allowed to broadcast, in the meaning defined in the relevant law.
Except if it’s too other hams — but perhaps your point means that already.
For those wondering where you can get that cool radio… it is from OpenRTX. It is a replacement firmware for the TYT and other radios that opens it up to support M17. See more at https://openrtx.org
No, that’s not quite accurate.
Hams transmit a request or a response to a known and actively identified individual station and must not attempt to Broadcast to many listeners (an audience)
Intentionally doing so would be a breach of there licence conditions/terms/rules.
Yes, the nature and physical properties of radio will mean it is a transmission that may be received by many stations.
but the ‘intention’ of the transmission is ‘to’ a specific ‘single’ station rather than to ‘many’.
This intention which is an important distinction means it is not a Broadcast.
Hams are only permitted to transmit to a specific intended station,
The only exception to this would be any initial cQ call,
which by its nature and intention ‘is a transmission for a response from ‘any’ licensed station.
But again, technically, within the terms of the license, this is still not a Broadcast.
As the intention of the transmission is to, specifically, other license holders, rather to ‘any’ listener.
With the exception of say the ARRL broadcasting Morse code practice or beacons that more than one person may use.
If all the fun stuff breaks the rules, why get a license at all?
Maybe you have never called CQ or QRZ?
That’s a broadcast to the entire world, and completely legal.
Naw, not really. While you cannot “broadcast” every time you key doesn’t have to mean it’s intended for only one person. Otherwise nets wouldn’t exist. Everyone on a net is listening to what the person who’s turn it is is saying. You are indeed intending for that to happen. Also, when in a rotation, you often ID and say “in the group” acknowledging that your transmission was intended for the group of people you are taking to. Nothing is in violation here. The violation is the “one way” traffic of intentional broadcasting.
Amateur Radio Operators CAN legally “broadcast” with one key restriction – the information transmitted in the “broadcast” must be applicable / of interest to other Amateur Radio Operators, thus such a transmission is called a “Bulletin”. If memory serves, the FCC definition of a “broadcast” is information intended for anyone / everyone. “Bulletins” by the nature of their information is not intended for anyone / everyone, Amateur Radio Operators.
This isn’t to say that Amateur Radio Operators transmiting “bulletins” hasn’t been abused – check out http://www.arrl.org/news/glenn-baxter-ex-k1man-sk-engaged-in-protracted-enforcement-battle-with-fcc
“With the exception of say the ARRL broadcasting ”
I have a email with Dan Henderson of the ARRL ADMITTING
what they do is ILLEGAL.
But like many in this country, they are above the law :(
My only question is, how will it catch on? This is like the 6th project I’ve seen trying to get open-source/libre digital radio work, and I’ve yet to see one truly take off and replace (or come close to) either FM or any proprietary digital radio system. FreeDV has existed for some time now, yet no one is using it. Then there’s the younger NPR (though it’s for transferring IP, there’s no reason why you couldn’t stream Codec 2 voice via RTP over it), and then a handful of others that I already forgot the name of.
It’s like the old XKCD comic “Standards” (927), in which we keep creating so many incompatible protocols for the same problem that they will hinder each other’s adoption.
All of the projects you mentioned are focused on solving different problems.
M17 is targeting digital voice on VHF and up.
FreeDV is focused on HF which has very different requirements in terms of bandwidth and noise. Using it on VHF+ is totally possible but you would need an all-mode radio which excludes HTs and most mobiles.
You could stream Codec2 voice via RTP over NPR, but NPR’s configuration is very heavily IP based and would not work well in it’s current form to roam from node to node or repeater to repeater.
It’s my guess that someone will be able to build M17 into to a software defined radio in a portable radio form factor. There is some proof of concept (hardware) units being developed right now and they’ll be vastly simpler than the complex mess (electronically, software, and user interface) that is a current portable two-way radio (no matter how cheap). Because such a radio is mostly software, it’ll be incredibly hackable for all kinds of experimentation – someone develops a new mode, people download it to their SDR, flash it, and are instantly running that new mode. The current two-way radio, repeater, coordination, etc. paradigm is so fossilized that this new approach will run rings around it. Imagine hackers having their own “private” network of single frequency repeaters on apartment balconies – not encrypted, but most Amateur Radio Operators won’t bother to get M17, so effectively, private. There will be some catalyst hardware, event, influencer, etc. that will cause such a thing to go viral.
As for FreeDV, it’s a Catch-22; without a large base of users, or significant demand from the potential buyers of radios specifically for FreeDV, the (HF radio) manufacturers don’t have much incentive to build in FreeDV, despite it being free of any intellectual property encumberances. Without it being easy to use (not an external modem), it’s a bit of hassle for users to get going and build up a large user base.
FreeDV OUGHT to be easy to implement on FlexRadio units – they’re “mostly software” and FreeDV is software, so it ought to be easy to implement. There was FreeDV add-in support in an earlier version of FlexRadio’s SmartSDR software, but it hasn’t been updated (by ???) for the current version of SmartSDR.
One of the goals of FreeDV with their recent grant from ARDC (https://www.ardc.net/apply/grants/2023-grants/enhancing-hf-digital-voice-with-freedv/) is to get FreeDV “built in”:
Work with commercial HF radio companies to embed FreeDV into at least two commercial radios, greatly reducing set up effort and reducing latency.
The first company to implement FreeDV on the rig gets my next radio purchase.
Actually I bet Xiegu would be willing to add it to the X6100. Apparently the WFView server component is has been added to their upcoming firmware release so they have a history of working with the open source community to add features requested by users.
Didn’t the mcHF have it already?
It Does. The mcHF is a really cool open-source project, and if I was in the market for a QRP rig, I would get one for sure. Unfortunately kits/projects like the mcHF, sBitx, etc… are pretty niche and the overall user base isn’t enough to tip drive demand for a new mode. If icom, Kenwood, or Yaesu added it to a rig, the potential user base would increase dramatically.
I only mentioned Xiegu because WFView recently announced on their Patreon (annoyingly paywalled but you can see the headline here: https://www.patreon.com/wfview) that the upcoming 1.7 firmware for the X6100 integrates their open-source project. This is a completely different attitude from the other manufacturers, especially Yeesu who goes around publicly stating they will void your warranty for programming a radio with CHIRP.
Even if Xiegu would include FreeDV, it’s still probably not a large enough user base to tip the scales, but it would be a start.
In reality almost all on-air presentations of digital encoding are more or less open-source, or at least in the public domain, that is we know how they work and we know how to make them function in various other devices – otherwise the various digital modes could/would not be supported by MMDVM implementations that support transport and bridging of different modes.
C4FM +/ Fusion +/- WiresX +/- NXDN – so far implemented only by Yaesu and ‘interop’ @ MMDVM
D-Star- so far implemented only by Icom and ‘interop’ @ MMDVM
DMR – ETSI open-standard, plus maker/service-specific augmentations
P25
So this works primarily simplex – radio to radio. How would an M17 repeater be implemented? What type of access control rather than mere (undesirable) RF carrier-present? What latency would decoding, assessing, permitting, then retransmitting impose?
If one implements the DigiRig (analog to FM) interface solution… how does one know/ensure only 9KHz of occupied bandwidth?