Hi,

?

I’ve ben doing a bit more reading on amateur radio digital voice modes used on HF.

?

To focus on one possibility for a digital voice mode that may be compatible with nonlinear RF finals, I discovered that D-Star is being experimented with by some hams on HF.

?

D-Star uses GMSK, Gaussian Minimum Shift Keying on VHF/UHF. Best I can tell the modulation mode is not changed in the HF world.

?

If so, this might be a way to add a voice mode to digital only modes transceivers, with simple nonlinear RF output stages, like the QDX.?

?

D-Star uses a proprietary codec, so what I’d suggest is use the frequency shift (GMSK) modulation method but swapping out the D-Star codec for the FreeDV codec.?

?

Any major flys in the ointment here? Other than getting ham radio operators interested in yet another voice mode?

?

There are, also, multiple frequency shift modes that might offer some advantages. But GMSK seems the likely starting point for experimentation or at least for this conversation.

?

I know there are folks that have gained much more knowledge of this area than my beginnings at sorting through this area of knowledge. Hopefully some of those folks will be able to add some practical additional knowledge on the subject.

?

Tom, wb6b

?

I believe section?97.307(f) of the FCC Rules still places a maximum symbol rate of 300 baud on emissions below 10 meters.

On Oct 15, 2021, at 10:19, Tom, wb6b <wb6b@...> wrote:

?

Hi,

?

I’ve ben doing a bit more reading on amateur radio digital voice modes used on HF.

?

To focus on one possibility for a digital voice mode that may be compatible with nonlinear RF finals, I discovered that D-Star is being experimented with by some hams on HF.

?

D-Star uses GMSK, Gaussian Minimum Shift Keying on VHF/UHF. Best I can tell the modulation mode is not changed in the HF world.

?

If so, this might be a way to add a voice mode to digital only modes transceivers, with simple nonlinear RF output stages, like the QDX.?

?

D-Star uses a proprietary codec, so what I’d suggest is use the frequency shift (GMSK) modulation method but swapping out the D-Star codec for the FreeDV codec.?

?

Any major flys in the ointment here? Other than getting ham radio operators interested in yet another voice mode?

?

There are, also, multiple frequency shift modes that might offer some advantages. But GMSK seems the likely starting point for experimentation or at least for this conversation.

?

I know there are folks that have gained much more knowledge of this area than my beginnings at sorting through this area of knowledge. Hopefully some of those folks will be able to add some practical additional knowledge on the subject.

?

Tom, wb6b

?

The d-Star mode is open source, the chip that the two radio
manufacturers use is proprietary, but it is only out of convenience
that Icom and Kenwood use the AMBE vocoder. Just as Yaesu and Alinco
use the same chip to implement C4FM and DMR

Of course, one would need to learn Japanese to read the D-Star
specifications, but they are public, were developed by JARL, and are
available for anyone to implement without license fee.


<>

Personally, the AMBE chip (at about $20 in single qty) seems like a
bargain compared with other methods of implementing D-Star, DMR, C4FM,
etc., but opinions vary.

I can buy an AMBE board for about $150 and 'operate' D-Star on my
windows PC w/o a radio (but what fun is that?):

<>


<>

That $150 board costs so much because it not only has the $20 AMBE
chip, but also a fair amount of other technology to support serial,
USB, WiFi AND Ethernet connectivity.

Ken, N2VIP

On Oct 15, 2021, at 10:19, Tom, wb6b <wb6b@...> wrote:

?
Hi,
?
I’ve ben doing a bit more reading on amateur radio digital voice
modes used on HF.
?
To focus on one possibility for a digital voice mode that may be
compatible with nonlinear RF finals, I discovered that D-Star is
being experimented with by some hams on HF.
?
D-Star uses GMSK, Gaussian Minimum Shift Keying on VHF/UHF. Best I
can tell the modulation mode is not changed in the HF world.
?
If so, this might be a way to add a voice mode to digital only modes
transceivers, with simple nonlinear RF output stages, like the QDX.?
?
D-Star uses a proprietary codec, so what I’d *suggest is use the
frequency shift (GMSK) modulation method but swapping out the D-Star
codec for the FreeDV codec.?*
?
Any major flys in the ointment here? Other than getting ham radio
operators interested in yet another voice mode?
?
There are, also, multiple frequency shift modes that might offer some
advantages. But GMSK seems the likely starting point for
experimentation or at least for this conversation.
?
I know there are folks that have gained much more knowledge of this
area than my beginnings at sorting through this area of knowledge.
Hopefully some of those folks will be able to add some practical
additional knowledge on the subject.
?
Tom, wb6b
?

The d-Star mode is open source, the chip that the two radio manufacturers use is proprietary, but it is only out of convenience that Icom and Kenwood use the AMBE vocoder. Just as Yaesu and Alinco use the same chip to implement C4FM and DMR

?

Of course, one would need to learn Japanese to read the D-Star specifications, but they are public, were developed by JARL, and are available for anyone to implement without license fee.

?

?

AMBE is just the audio codec (source coding), it has nothing to do with
the actual channel coding (modulation), right?

And indeed, AMBE patents already expired some time ago... and there are
some open source implementations hanging around for some time, eg:


For eg., freedv uses codec2 or LPCNet for audio coding.

The D-Star spec might be public, but it all seems rather murky.
As you say, it is implemented by buying something like that proprietary ABME chip,.
FreeDV, on the contrary, has code available that can run on a low end ARM processor.
That seems more in line with what Amateur Radio should be.
Clearly, opinions vary.

From Wikipedia:??

Proprietary codec? ?Like other commercial digital modes (,?,?,?,?,?), D-STAR uses a closed-source proprietary voice codec () that's patented by Digital Voice Systems, Inc. (DVSI)?because it was the highest quality and only codec available in silicon when the system was released. Amateur radio operators do not have access to the specification of this codec or the rights to implement it on their own without buying a licensed product. Amateurs have a long tradition of building, improving upon and experimenting with their own radio designs. The modern digital age equivalent of this would be designing and/or implementing codecs in software. Critics say the proprietary nature of AMBE and its availability only in hardware form (as ICs) discourages innovation. Even critics praise the openness of the rest of the D-STAR standard^[]?which can be implemented freely. As of 2017 the patents have expired, as announced by Bruce Perens, K6BP at the 2017 ARRL/TAPR DCC in his State of Digital Voice talk.

Would be interesting to see what Bruce Perens thinks about D-Star these days.
He's one of the original forces behind the FreeDV codec (didn't write it, but encouraged its development).

Jerry, KE7ER

On Fri, Oct 15, 2021 at 12:02 PM, Ken Hansen wrote:

The d-Star mode is open source, the chip that the two radio manufacturers use is proprietary, but it is only out of convenience that Icom and Kenwood use the AMBE vocoder. Just as Yaesu and Alinco use the same chip to implement C4FM and DMR

?

Of course, one would need to learn Japanese to read the D-Star specifications, but they are public, were developed by JARL, and are available for anyone to implement without license fee.

?

?

On Oct 16, 2021, at 10:01, Tom, wb6b <wb6b@...> wrote:

So are the folks experimenting with D-Star on HF just kind of flying under the radar?

Tom, wb6b

On Oct 15, 2021, at 15:29, Jerry Gaffke via groups.io <jgaffke@...> wrote:

The D-Star spec might be public, but it all seems rather murky.
As you say, it is implemented by buying something like that proprietary ABME chip,.
FreeDV, on the contrary, has code available that can run on a low end ARM processor.
That seems more in line with what Amateur Radio should be.
Clearly, opinions vary.

This is all interesting information. I'm back in the reading mode exploring the things that have come up in the posts.?

One would think that "advanced multi-band excitation coding" (AMBE) being invented so long ago would be easily improved upon. But, sometimes the fundamental design is well done and not a compromise that was realizable with the technology of the time (circa 1990), that it stands the test of time. Certainly would not need the expensive custom codec chips; now that it can be implemented in software on current microprocessor chips.?

One thing that occurred to me might be a sort of "earth-to-mars" mode. Where each operator's voice, in a conversation, is stored/buffered and sent at a significantly lower bit rate to the destination. This could allow for having voice conversations under significantly reduced signal to noise ratios, facilitating better DX communication. Albeit with a turnaround delay for each side to pick up on their side of the conversation. Maybe would give each side some time to think a bit before they start speaking again to reply. Would also produce sort of an intriguing feeling of suspense, waiting for the data stream, in the noise, to end so you can listen the reply from the distant other location.?

So are the folks experimenting with D-Star on HF just kind of flying under the radar?

Tom, wb6b

Hi,

This is all interesting information. I'm back in the reading mode
exploring the things that have come up in the posts.?

One would think that "advanced multi-band excitation coding" (AMBE)
being invented so long ago would be easily improved upon. But,
sometimes the fundamental design is well done and not a compromise
that was realizable with the technology of the time (circa 1990), that
it stands the test of time. Certainly would not need the expensive
custom codec chips; now that it can be implemented in software on
current microprocessor chips.?

One thing that occurred to me might be a sort of "earth-to-mars" mode.
Where each operator's voice, in a conversation, is stored/buffered and
sent at a significantly lower bit rate to the destination. This could
allow for having voice conversations under significantly reduced
signal to noise ratios, facilitating better DX communication. Albeit
with a turnaround delay for each side to pick up on their side of the
conversation. Maybe would give each side some time to think a bit
before they start speaking again to reply. Would also produce sort of
an intriguing feeling of suspense, waiting for the data stream, in the
noise, to end so you can listen the reply from the distant other
location.?

So are the folks experimenting with D-Star on HF just kind of flying
under the radar?

Tom, wb6b

D-star on HF works, I already tested with a IC-7100. Of course it could
be considered a "wideband" HF mode, as it uses > 6 kHz of bandwidth. The
sad part is that the same IC-7100 does not support such wide bandwidth
in "standard" SSB mode. Anyway, freedv with LPCNet is way ahead in time,
using a much more advanced modem (and smaller bandwidth) and a
state-of-the-art AI-based vocoder.

Rafael

On 10/16/21 5:01 PM, Tom, wb6b wrote:

Hi,

This is all interesting information. I'm back in the reading mode
exploring the things that have come up in the posts.?

One would think that "advanced multi-band excitation coding" (AMBE)
being invented so long ago would be easily improved upon. But,
sometimes the fundamental design is well done and not a compromise
that was realizable with the technology of the time (circa 1990), that
it stands the test of time. Certainly would not need the expensive
custom codec chips; now that it can be implemented in software on
current microprocessor chips.?

One thing that occurred to me might be a sort of "earth-to-mars" mode.
Where each operator's voice, in a conversation, is stored/buffered and
sent at a significantly lower bit rate to the destination. This could
allow for having voice conversations under significantly reduced
signal to noise ratios, facilitating better DX communication. Albeit
with a turnaround delay for each side to pick up on their side of the
conversation. Maybe would give each side some time to think a bit
before they start speaking again to reply. Would also produce sort of
an intriguing feeling of suspense, waiting for the data stream, in the
noise, to end so you can listen the reply from the distant other
location.?

So are the folks experimenting with D-Star on HF just kind of flying
under the radar?

Tom, wb6b

On Sat, Oct 16, 2021 at 05:20 PM, Jerry Gaffke wrote:

sending compressed speech slowly to get around that 300 baud thing

I can see your point. If it turns out that the people running D-Star on HF are not pushing the regulation envelope using GMSK modulation, that could mean the possibility of a digital voice HF mode that is compatible with nonlinear RF finals. Then the slow bit rate, buffered, mode would simply be an additional way to extend the DX possibilities for people who really want to do DX and OK with a delay for the purpose.

Switching out the AMBE codec for something like FreeDV or Codec2 would be a good upgrade to a more modern codec (voice encoder) with, from what I've been reading, a two to one improvement in the bitrate needed to transmit the encoded voice.?

Hi,

If I wanted to experiment with nonlinear RF finals and digital modes with my uBitx, would doing something like changing the bias and drive for the uBitx finials, driving them harder with little or no DC bias, be a practical way to do this?

Would that provide an efficiency gain that would let the uBitx putout more power (or at least generate less heat) with the same finals, without a redesign of the RF output stages?

Likely I'd build a simple digital RF chain and switch it in, close to the final stages, when I want to try these digital modes.?

Tom, wb6b