[chirp_devel] Mode designator for YSF?
Hi all,
I'm not a Yaesu guy and thus not very YSF-literate. It looks to me like the YSF radios call their digital mode "DN" when enabled (as opposed to "FM"), like D-STAR has DV. Is that right and does it make sense to use that in chirp's mode list for editing these channels? I had originally thought "YSF" like "DMR" would make sense, but now I'm not sure after looking at a radio. Yaesu's documentation doesn't seem to be super clear on this.
It also seems like maybe the convention is to actually set the radio to Auto mode because of the dual FM/digital nature of YSF repeaters, is that right? If we add support for this, does that mean "Auto" is really what we should be setting in most cases?
There's also a "VW" mode, but it seems to be sort of an additional flag, more so than just a mode. So does it make sense to set a channel to FM/DN (or Auto) and then have VW be a separate flag via extra/properties?
If there's someone that is YSF-literate that might be able to test some stuff in the coming week or so, speak up :)
--Dan
On 2023-01-10 02:32, Dan Smith via chirp_devel wrote:
Hello
I can try but don't want to claim correctness on this, nor to know how things should be done best. And I don't want to fight over this :-). The radio that I played with could be set to one of: FM, DN, VW, AMS. This is the interactive setting used when operating; I am not sure whether this also corresponds to the choice available when programming (I'm not using any of those digital modes). AMS is the automatic mode that picks the "right" one of the other tree. I find this to be a sensible choice for multimode repeaters, but not everybody agrees: Some people only want to hear the digital traffic, others only the analog traffic and they would set their radio accordingly. AMS is also a bit annoying since it changes your transmit mode to whatever was last used on the repeater. This is cool because you are automatically responding in the same mode that somebody used for calling, but then again if you want to call, you always have to check your settings first to ensure you are on the desired mode. There is, however, a setting that allows you to set the default for that. (E.g. RX AMS, TX always FM). I don't remember if this is per-channel but that works for me.
As for the DN vs. VW thing: These are different ways the voice is encoded into a representation of the speech model and further treated with the usual techniques (whitening, interleaving etc.). In DN, a voice data "packet" and a metdata "packet" containing the call sign, GPS coordinates and other stuff are transmitted in an alternating fashion. IN VW, there is some kind of metadata pilot package in the beginning and then it's just voice data all the way, thus more of the channel's bandwidth is used for voice data -- in fact, all of it vs. half, which allows them to use a better speech model and the audio quality is supposed to be better. The tradeoff is that a receiving station doesn't see the call sign on late entry or if that pilot packet got crippled. Hence, I believe, the names "Digital Narrow" and "Voice Wide".
All that said, I'd have to check what happens if somebody has his radio set in DN and some VW is incoming, or vice versa. I guess it would still decode but not change the current TX setting? But in any case, I always thought of the VW and DN to be two modes that are more or less separate from each other. Offering them as individual choice would reflect that. With the flag/extra property instead, somebody's though process while programming the radio would go like 'Oh yeah, I want digital on that channel of course... ah yeah, and I want that to be in "Voice Wide" [with the trade-off mentioned above]'. That would also work, I guess. Thus the options/schemes are:
['FM', 'AMS', 'DN', 'VW'] or ['FM', 'AMS', 'YSF'] with the extra flag for improved 'VW' voice. That one would work according to the aforementioned thought process, but I don't like it because:
- Using a checkbox for DN vs. VW imposes some kind of default and I don't see why chirp should suggest one over the other. - In "Fusion", the "F" in "YSF" refers to fusing digital and analog into one thing. This was one of the marketing lines when the system was introduced, allowing for a smooth transition to digital and all that. As such, it was meant to describe the whole "digital plus analog together" idea. Using that term to discriminate the digital part from the analog part thus sounds wrong... But one could use 'Digital' instead of 'YSF' in the second scheme.
I hope that helps somewhat
Matt
Mathias Weyland via chirp_devel wrote:
All that said, I'd have to check what happens if somebody has his radio set in DN and some VW is incoming, or vice versa. I guess it would still decode but not change the current TX setting?
DN/VW is a transmit setting only; as long as one of them is set, on receive the radio automatically switches to the correct mode (whilst keeping the transmit setting intact).
I can try but don't want to claim correctness on this, nor to know how things should be done best. And I don't want to fight over this :-).
Hah, I don't want to fight, I just want to make a good decision about how we can support this without having to change course later. Thanks for the brain dump. Even though I'm not sure we agree on the conclusion, I think the discussion is good to make sure we don't make a decision that we have to change later. Please interpret the below as critical design discussion and not fighting :)
The radio that I played with could be set to one of: FM, DN, VW, AMS. This is the interactive setting used when operating; I am not sure whether this also corresponds to the choice available when programming (I'm not using any of those digital modes).
Yeah, this seems really confusing to me. I have an FT3 to play with, and I see that it lets you select FM, DN, or Auto for the transmit mode on the radio. Similarly in Yaesu's software you can select these three for a given channel. However, in their software for the FT2 and FT1 (which are basically identical in memory layout) there is no such option. The manuals for those radios seem to indicate that the behavior on the radio itself is the same, but I can't figure out why their software won't let you choose that. And, even though I can select VW in a separate place in the radio, the FT3 software seems to have zero mention of it.
I've decoded the bits for AMS, FM, DN but because of the above discrepancy I'm not sure if I should expose those to the FT1 and FT2, or keep it in the FT3 subclass. Is it possible that Yaesu just didn't implement knobs for the digital settings in their new (at the time) flagship digital radios?
AMS is the automatic mode that picks the "right" one of the other tree. I find this to be a sensible choice for multimode repeaters, but not everybody agrees: Some people only want to hear the digital traffic, others only the analog traffic and they would set their radio accordingly. AMS is also a bit annoying since it changes your transmit mode to whatever was last used on the repeater. This is cool because you are automatically responding in the same mode that somebody used for calling, but then again if you want to call, you always have to check your settings first to ensure you are on the desired mode. There is, however, a setting that allows you to set the default for that. (E.g. RX AMS, TX always FM). I don't remember if this is per-channel but that works for me.
This (the TX preference) is also not set-able in their software. I haven't checked to see if it's a per-channel setting when stored on the radio though.
['FM', 'AMS', 'DN', 'VW']
I guess I'm not sure this is the best plan. Seems to me that "AMS" is really a behavior of the radio, in that it will change your mode between FM and DN based on the last received signal. Further, "VW" doesn't seem like a mode to me. It's still the same modulation, but with more voice data and less data data, more akin to transmit bandwidth on an HF radio or something. And especially since you don't control whether you receive VW or not (right?) it also doesn't seem like a mode to me.
I'd also hate to add "AMS" as a mode in chirp's master list, since it's a behavior. Some other radio could use AMS to mean "FM or D-STAR" and Yaesu themselves could do something like add DMR to these radios and AMS would switch between all three depending on the received signal. The modes need to be pretty stable in their meanings, as that's how we determine if you can copy/paste a memory between two radios. If they both consider "AMS" to be a mode, but one means "FM or DN" and the other means "FM or AM", that's not going to help anyone.
- Using a checkbox for DN vs. VW imposes some kind of default and I
don't see why chirp should suggest one over the other.
Well, VW seems to be off by default in the radio itself, and Yaesu's own software doesn't even appear to let you select it. I assume that means that it's a sort of demo-able feature, perhaps to make the audio sound good for presentations. How prevalent is the real-world usage?
- In "Fusion", the "F" in "YSF" refers to fusing digital and analog
into one thing. This was one of the marketing lines when the system was introduced, allowing for a smooth transition to digital and all that. As such, it was meant to describe the whole "digital plus analog together" idea. Using that term to discriminate the digital part from the analog part thus sounds wrong... But one could use 'Digital' instead of 'YSF' in the second scheme.
Yeah, fair enough. I'd be more inclined to use "YSF" to mean "AMS" just because it's less generic. However, users looking at their radio, manuals, and Yaesu's own software will likely not intuitively know what it means.
I'm kinda surprised - this seems like a bit of a trainwreck, and more confusing than a digital-only D-STAR repeater to me. There is zero activity on my local YSF repeaters, but based on this, I kinda wonder if that activity would be a cacophony of people in both modes not really realizing when they're digital or not. Is it actually better in practice?
I definitely get your argument that assuming a default for AMS is imposing a preference on people, although it seems like that's Yaesu's intended default. So I guess it seems like if we add a DN mode to mean "Yaesu's C4FM format" and then flags for AMS or VW (with a default to Yes and No respectively), that's the right behavior, and keeps DN-supporting radios speaking the same language. If people cry foul at being defaulted to AMS, we could add a preference of some sort to control that. Even if that needs to change in the future, I think adding DN now so we can support repeaterbook YSF entries and general control of these radios is not something we'd have to remove/change later.
Does that seem okay, at least to start? Anyone else have information or an opinion on this?
--Dan
On 2023-01-11 00:13, Dan Smith via chirp_devel wrote:
Hello
Reading what I just wrote during the past hour or so, there is a lot of technical details that may be boring to say the least... Thus you may want to scroll to the end where I tried to address the solution that you proposed, which is quite good IMO.
The manuals for those radios seem to indicate that the behavior on the radio itself is the same, but I can't figure out why their software won't let you choose that. And, even though I can select VW in a separate place in the radio, the FT3 software seems to have zero mention of it.
That seems odd indeed. Then again, Yaesu made quite a number of very odd choices when it comes to Fusion... What happens if you select and store VW on the radio and read out the memory? Does it have the VW flag? Another idea may be to look at another third party programming software made by RT Systems and check if they added the functionality...
Is it possible that Yaesu just didn't implement knobs for the digital settings in their new (at the time) flagship digital radios?
That is possible in the light of what I wrote above... There is also the following tidbit that may have affected such decisions: VW and DN differ from each other in three fundamental ways. One is the way the metadata and the digital voice is multiplexed, as mentioned in my previous E-Mail. The second one is how the digital voice frames are processed and turned into the bit sequence that is actually transmitted, e.g. how and what kind of redundancy is added, how the digital voice bits are scrambled, block-interleaved etc. (although "DN" by itself has two ways of doing that...), and by the way the voice is encoded into the speech model representation. While both DN and VW use the same fundamental speech model, they differ in the number of bits that are used to encode the various parameters of the model. These models and encoding schemes are proprietary and you can buy chips to transcode from PCM and back. Two details are interesting about this: (a) The DN type is actually the same as the one used by DMR (but not stipulated by the DMR standard). This is why it is easy to write software that bridges between DMR and the YSF DN flavour. Also, there is no loss of quality due to transcoding, since you don't have to decode to PCM and re-encode. "VW" may have dropped out of favour because it is (was?) not supported by those bridging solutions. (b) the common chips that you can buy to convert back and forth actually do not support the "mode" used in VW. Yaesu licenses some DSP IP to do that, but this is not something you can do in low quantities, thus there is no easy/legal way for us to homebrew receivers that fully support YSF. While I don't think that Yaesu cares about that, they may have come to the conclusion that it was not smart to use that "special" thing for VW and backtracked a little? Who knows...
Seems to me that "AMS" is really a behavior of the radio
That makes a lot of sense.
Further, "VW" doesn't seem like a mode to me. It's still the same modulation, but with more voice data and less data data, more akin to transmit bandwidth on an HF radio or something.
I would disagree with that statement but without drama. My disagreement is because of the technical detailes outlined above. The differences are much more substantial than just a matter of bandwidth. Yes both are 4FSK, the air packets have the same preamble and header, and indeed they use the same fundamental speech model, but everything else is different. In an odd way, DMR and DN are more similar to each other than DN is to VW. Does that make DMR and DN the same mode? Then again, my point is based on the technical implementation of those things and that may not be what people care for or what makes for the best implementation decision in chirp.
And especially since you don't control whether you receive VW or not (right?)
Correct, that was confirmed by Charlie earlier tonight.
How prevalent is the real-world usage [of VW]?
I have no idea :-). Also there may be regional differences so any particular answer may be biased. Still the solution you proposed below seems like a sensible one to me...
However, users looking at their radio, manuals, and Yaesu's own software will likely not intuitively know what it means.
This is a good point. Everything I wrote is based on the technical details and not necessarily on how people use their radios and software. I should add that I was involved in reverse-engineering the missing parts of the YSF protocol when it was launched. My interest in YSF was mostly in this regard and I have not used any of that in practice...
trainwreck
... aye ...
I kinda wonder if that activity would be a cacophony of people in both modes not really realizing when they're digital or not. Is it actually better in practice?
The difference is pretty clear by the presence of noise vs. the robotic nature of digital voice. There have been issues with people not having AMS turned on and transmitting on top of ongoing contacts because they failed to realize that the frequency was in use. Also, at least some radios disable AMS when you establish a networked connection ("WIRES-X") which adds to the confusion. I practice, YSF has been mostly a novelty thing around here, people try it after purchasing a new radio and then quickly lose interest, for the most part.
So I guess it seems like if we add a DN mode to mean "Yaesu's C4FM format" and then flags for AMS or VW (with a default to Yes and No respectively), that's the right behavior, and keeps DN-supporting radios speaking the same language.
I guess that does make sense. What about the following scenarios, though:
- Would there be a difference between 'FM' plus AMS checked and 'DN' plus AMS checked? Because according to this logic, that should be the same setting. How would we deal with people who program FM & AMS and once they read back their programming, chirp displays that as DN & AMS (or vice versa)? Would that cause confusion? Would users know what to select if they actually want the AMS setting? I wouldn't really know if I had to select 'FM' or 'DN' along with the AMS checkbox to get that...
- What about the combination FM, no AMS and VW? That would not make sense. Is there a way to gray out the VW checkbox in that case? Then again, this may be similar to the RX-Tone column when TSQL is not used..
I hope that helps, and apologies of the verbosity of my post
Matt
That seems odd indeed. Then again, Yaesu made quite a number of very odd choices when it comes to Fusion... What happens if you select and store VW on the radio and read out the memory? Does it have the VW flag?
This is what I'm saying I haven't done yet. I decoded more of the unknown flags last night (which I merged a bit ago) which will help cut down on the things I need to look for. But yeah, I will do this. The way the radio presents it, it might be a global thing and not per-channel, I'll just have to see.
That is possible in the light of what I wrote above... There is also the following tidbit that may have affected such decisions: VW and DN differ from each other in three fundamental ways. One is the way the metadata and the digital voice is multiplexed, as mentioned in my previous E-Mail. The second one is how the digital voice frames are processed and turned into the bit sequence that is actually transmitted, e.g. how and what kind of redundancy is added, how the digital voice bits are scrambled, block-interleaved etc. (although "DN" by itself has two ways of doing that...), and by the way the voice is encoded into the speech model representation. While both DN and VW use the same fundamental speech model, they differ in the number of bits that are used to encode the various parameters of the model. These models and encoding schemes are proprietary and you can buy chips to transcode from PCM and back. Two details are interesting about this: (a) The DN type is actually the same as the one used by DMR (but not stipulated by the DMR standard). This is why it is easy to write software that bridges between DMR and the YSF DN flavour. Also, there is no loss of quality due to transcoding, since you don't have to decode to PCM and re-encode. "VW" may have dropped out of favour because it is (was?) not supported by those bridging solutions. (b) the common chips that you can buy to convert back and forth actually do not support the "mode" used in VW. Yaesu licenses some DSP IP to do that, but this is not something you can do in low quantities, thus there is no easy/legal way for us to homebrew receivers that fully support YSF. While I don't think that Yaesu cares about that, they may have come to the conclusion that it was not smart to use that "special" thing for VW and backtracked a little? Who knows...
Wow, yeah. It sounds to me from the above (correct me if I'm wrong) that maybe DN encompasses two things, sort of layer 1 and layer (or 7). Layer 1 being the lowest protocol and Layer 7 being some of the voice/data/codec stuff , and that VW mode would reuse a lot of layer 1 but swap out more of the Layer 7 stuff than just "twice as many voice packets" (so to speak). Maybe?
Further, "VW" doesn't seem like a mode to me. It's still the same modulation, but with more voice data and less data data, more akin to transmit bandwidth on an HF radio or something.
I would disagree with that statement but without drama. My disagreement is because of the technical detailes outlined above.
Yep, I'm seeing...
The differences are much more substantial than just a matter of bandwidth. Yes both are 4FSK, the air packets have the same preamble and header, and indeed they use the same fundamental speech model, but everything else is different. In an odd way, DMR and DN are more similar to each other than DN is to VW. Does that make DMR and DN the same mode? Then again, my point is based on the technical implementation of those things and that may not be what people care for or what makes for the best implementation decision in chirp.
Right okay. So I still want to call DN the mode (in terms of what CHIRP needs, layer 1). And I guess say that there's also a DN/VW switch for Layer 7. I know, I'm making up layers.
I have no idea :-). Also there may be regional differences so any particular answer may be biased. Still the solution you proposed below seems like a sensible one to me...
Ack, cool.
trainwreck
... aye ...
/me screenshots for later.
I kinda wonder if that activity would be a cacophony of people in both modes not really realizing when they're digital or not. Is it actually better in practice?
The difference is pretty clear by the presence of noise vs. the robotic nature of digital voice. There have been issues with people not having AMS turned on and transmitting on top of ongoing contacts because they failed to realize that the frequency was in use. Also, at least some radios disable AMS when you establish a networked connection ("WIRES-X") which adds to the confusion. I practice, YSF has been mostly a novelty thing around here, people try it after purchasing a new radio and then quickly lose interest, for the most part.
Yeah, I'm sure you can tell the difference. What I mean by cacophony is in terms of the chaos (as you describe) and the wildly different sounds when you're expecting one or the other :)
- Would there be a difference between 'FM' plus AMS checked and 'DN' plus AMS checked? Because according to this logic, that should be the same setting. How would we deal with people who program FM & AMS and once they read back their programming, chirp displays that as DN & AMS (or vice versa)? Would that cause confusion? Would users know what to select if they actually want the AMS setting? I wouldn't really know if I had to select 'FM' or 'DN' along with the AMS checkbox to get that...
Well, first let me say that CHIRP is an abstraction, and as such, we have to make some concessions in the name of compatibility.
But, what I've got going right now is that selecting "FM" will turn off DN and AMS, the equivalent in Yaesu's software to selecting "FM" and "Analog" in the "Analog/Digital" column. Selecting "DN" will enable AMS, the equivalent in Yaesu's software of selecting "FM" and "AMS" in the "Analog/Digital" column. Setting AMS and/or VW when the mode is "FM" or "NFM" or "AM" won't do anything. That way if you copy and paste an FM channel from a repeater directory or another non-YSF radio, you'll get an analog channel that won't try to auto-mode you into anything.
- What about the combination FM, no AMS and VW? That would not make sense. Is there a way to gray out the VW checkbox in that case? Then again, this may be similar to the RX-Tone column when TSQL is not used..
Well, I'm not planning to actually expose these as columns in the main editor, but options in the RightClick->Properties dialog, like D-STAR stuff is now. There's no rules-based stuff in there to grey out certain things when others are set, mostly because that's like a "tell me what you want it to look like" sort of interface. Thus, no tone mode based hiding of fields in there. But that could change in the future.
I hope that helps, and apologies of the verbosity of my post
Don't apologize, verbosity is what I was looking for :)
So what I've got so far is this:
https://github.com/kk7ds/chirp/pull/376
Which allows me to trivially select FM and DN with the above "probably what most people want" defaults, and to import YSF repeaters from repeaterbook. I haven't implemented the toggles for disabling AMS (but keeping DN on) or VW yet.
--Dan
participants (3)
-
Charlie Li -
Dan Smith -
Mathias Weyland