Digirig 1.9 Windows 11 TX audio 20 dB too high

,
1

I haven’t used digirig for some time, and this is one of the reasons why I use SignaLink. I am doing testing for the local ARES group, and one of the devices to test is the digirig 1.9 TTL device. I understand that these tests are radio- and possibly OS-specific.

I TX on one radio using 0.05W, and RX on another using an external antenna and my shack 2m radio. Testing w SignaLink went as expected.

Connected the digirig, installed the driver from Silicon Labs, and started with Tune.
Driver (CODEC):
Playback: Levels: mic=0.0 dB spkr=-27.5 dB, Enhancements: disable all, Hardware acc=OFF, 16 bit 48 kHz, allow exclusive=optional, Spatial=OFF

I init had the speaker level higher and was completely blowing the Tune audio wave. I gradually lowered it…to no observable effect. So I turned to fldigi, and started lowering its TX level attenuation. When I got down to -19 dB (0.112% of FS), I finally got the fldigi SIG wave form within the suggested band.

So we have -27.5+(-19) dB of attenuation in the audio stream sent to my radio. WOW. Why on earth does the digirig use so much gain? I’m not using it to drive a stadium speaker ! This represents a huge loss of signal accuracy…assuming a 16 bit ADC with 90% loss from fldigi, we are down from 32768 levels to ~3300. I understand that fldigi’s FFT might handle this, but given the advertised use of digirig, it seems a poor design choice. The driver attenuates this further…

For comparison, with SignaLink Version 1,
fldigi TX level -3.0 dB
No SL driver selected as default comms.
Playback: Levels -10.2 dB, Enhancements: disable all, Hardware acc=OFF, 16 bit 48 kHz, allow exclusive=optional, Spatial=OFF

digirig 1.9
RX: Levels 0.0 dB ??? Scale runs 0 to +24 dB > no gain
Recording: Listen = NO, AGC=NO, Levels +0.0 dB, 1 Ch 48 kHz, allow exclusive=optional, Enhancements=NO
This does result in a 100% copy of a plain text message in fldigi, using radio VOL set to about 20% FR. This is acceptable.

Back to digirig RX/TX.
Both stations (digirig, Kenwood F6A, Windows 11) (SignaLink, Kenwood V71A, Windows 10), send and receive a plain text message at 100%. Haven’t tested other modes yet.

Now that I’ve whined, I’m open to an alternative explanation of what is happening here. Remember, I am running with a lot of attenuation with digirig in both send and receive. I’ll take the win, but not happy with the gain in digirig. I am aware that I can cut the 20 dB ATT jumper, but that is on the RX side, and the huge gain is on the TX side.

2

Congratulations on finding a workaround. Denis @K0TX will probaby comment with the authoritative answer.

I am not a SignalLink user, although I regularly particpated in digital nets with users who do. I do not know the details of Denis’s @K0TX design choice for the default output gain. I’m always happy with more distortion-free gain because it gives more headroom for whatever I am trying to do. Are you annoyed because you want the Windows 11 sound settings to be the same for both SignalLink and Digirig? Are you using Digirig for VOX, or taking advantage of PTT?

I appreciate the Digirig’s gain when I use budget computer headset for voice. That requires more gain than the digital radio modes.

@KI5OPY developed HAM Audio Bridge:

You might also solve the settings issue with a software mixer like Banana to do the additional attenuation before the Digirig TX stage. Then you push the config issue into another piece of software, rather than having to change Windows sound settings when you switch between Signalink and Digirig.

73 Constrainted

3

OK, to reply to the questions:

want the Windows 11 sound settings to be the same for both SignalLink and Digirig?
No. I am quite content if they are different…within reason. 20 dB is not even close to reasonable, in my world. To be fair, I need to swap to a different computer and see what the difference is, if any.

taking advantage of PTT
Yes.

I appreciate the Digirig’s gain when I use budget computer headset for voice. That requires more gain than the digital radio modes.
I guess if I bought digirig to listen to music, that would be different. But I got it specifically to do ham digi comms. Yes, I got what you were saying. But you could perhaps use earbuds, rather than headphones, and those would have a higher impedance, and thus higher “volume” for a given setting. Or maybe I still missed the point…

Ham Audio bridge…I’ll look into this for other reasons, but it seems to be inactive? And I’m not sure I’d benefit from it, because I don’t do a lot of swapping. I use different pieces of software and hardware, but usually only one config at a time.

You might also solve the settings issue with a software mixer like Banana to do the additional attenuation before the Digirig TX stage.
Don’t really need another layer. Just need the ones I use to work responsibly.

I’d be interested to hear why all the gain, so I can await a reply from Denis.

Most hams I interface with regarding fldigi don’t use the FFT or SIG displays, to their detriment. Both are extremely helpful, but require some learning to be that useful. Though I have been a ham for decades, I had not used digi modes until the past few years. When I was first introduced to fldigi, the digi modes were presented as magic or an art. I eventually learned that this POV stemmed from ignorance. It’s just science, and at the user level, not really hard science, just not as simple as turning on the radio.

This addresses my point of view:
Consider an fldigi modem that adds 2 sine waves to send a signal. Assume both have the same amplitude. The transmission will have a peak amplitude of 2. Now imagine that you are transmitting that with fldigi, and you use the TX level attenuator to decrease the amplitude of the wave packet by 50%. You have now lost vital information needed to reconstruct the signal. You send that signal.

The fldigi FFT will not, I //think// reproduce the original frequency spectrum.
But I made an assumption here that may not apply to fldigi, so fldigi may be better at analyzing its own signals, even when they are greatly attenuated.

Of course, I don’t really know how the different fldigi modes construct a wave packet, so I don’t know how applicable this scenario is.

The max digi wave generated by fldigi contains the most data. Anything we do to attenuate that chips away from the decode success rate.

I could do some FFT analysis and|or capture some fldigi transmission sent via digirig. I could compare the RX data to fldigi’s output. That would tell me if I am losing data, or just annoyed in principle :wink:

Of course, the question stands…why is the gain so high in digirig?

4

Your methods of analysis are sophisticated, and really take advantage of fldigi’s capabilities. When you finish your analysis, I encourage you to update us here in the forum.

I want to respond, so someone switching from Signalink to Digirig, or starting digital modes from scratch with Digirig can get going quickly.

For HF, set up the Digirig with your computer and radio. Find a Web SDR on the internet at the appropriate range for your band. Find an open frequency where you have privileges, and tune and then TX until you have a “good” audio level. If you can hear distortion, use the software (fldigi or Windows) to attenuate the signal. Even better is to listen to the audio levels in “on-the-air” signals like MSK, Vara, Packet, etc. to set your audio level to match. In my experience 90% of the time, “good” is good enough, to then communicate with stations. If you can’t reach a WebSDR, find a friend, or use another HF rig with a dummy load connected to the antenna input, or start with a Winlink RMS node.

For VHF, use an HT with the antenna disconnected ( or with a dummy load) as monitor, and set the audio level to be “good” and that will be good enough to get started.

73 Constrainted

5

I wrote:

So come on, “headset and voice” means I am talking about the mic.

You had some great posts about using the Digirig with tm–71a , yet:

Since you don’t use the Digirig anyway, why don’t you stick to the Signalink, and recommend the Digirig to the ARES group?

73 Constrainted

6

Oops. forgot the smiley face. I meant the comment about music as a joke, so I apologize for seeming to be snarky.

OTOH, I missed the point that you were using the headset to use the mic. Not sure how that works with digirig. It seems this would be a good way to get a headphones+mic on a laptop that doesn’t have as TRRS input. Care to add a few comments about this? I will also browse the web. This seems like a good use of digirig I’d like to explore.

Why I haven’t used the digirig for a while:
Well, the SignaLink works fine for me.
I wanted to get the digirig and test it so that I could be more informed when I talk with other local hams.
I don’t like the coiled cords. They were all that was available when I purchased. I grew tired of carrying a 2 kg weight to hold my HT in place while I used the digirig . Of course I didn’t do that, but I did add a TRRS patch cable to my bag so that I could put the radio where I want it, and not right next to the laptop. I still don’t understand those who seemingly prefer the coils. Maybe they work better with heavier radios. My V71A is bolted down, so it doesn’t move when I attach the digirig coiled cable. I will say that I carry more gear to connect to the Icom radio in the EOC, but that is because I use a switch inline with the mic so that I can talk and do digi on the same radio. My V71 has 2 VFOs, so I don’t need the switch there, but the Icom does not. I’d need that with the dr cables, too, I guess. The dr cable is designed to replace the mic in the Icom. I have a sealed Pelican-like case (70x120x280 mm) for the EOC gear, so it is convenient.

One of the major ways I support ARES is by manning an EOC that uses an Icom radio. I happen to have a SignaLink jumper block for that radio, so I use SL there. I didn’t buy a $30 ICOM RJ-45 Cable for Digirig, just kept using the SL. And I’d def need a TRRS extension in the EOC, because I share the space with 5 other systems, and need to be careful not to disrupt them. (Might not be clear why…) And a switch (see above).

In closing, I read where you mentioned using 2 radios without antennas, or perhaps a dummy load. I thought that a good idea, but def a dummy load. Then I realized that I have 2 radios that can go as low as 0.05W (Kenwood F6A, D72A), so I gave that a run last night, and it worked great. I had been using the V71A specifically because it “normalizes” to some degree RX and TX audio over the DATA port. That is, RX audio is boosted or reduced to keep it in a voltage window; TX audio from the source is boosted or reduced before sending to the mixer, so that it is a “normalized” audio signal when transmitted. All this is on FM, of course. I wanted to use this feature while testing SL and digirig so that I could observe the effect of changing the TX ATT in fldigi. Very instructive. It also provided a way to look at the effect of changing the CODEC gain during RX, knowing that I was receiving a predictable signal. But for just testing, the two HTs are sufficient.

I don’t have a preference for either digirig or SL at this point. The space savings of the dr, one of the features that attracted local hams, is eaten up by the volume needed for the coiled cables (and for me, the 3.5 mm TRRS extension cable), and the apparent cost savings is reduced by the need to have a radio class-specific cable. My goal was to test and report, not to recommend. Neither of the devices is “PnP”, as my testing shows, but they are not difficult to configure once all the audio controls are understood. I was surprised to see that with digirig I needed to use the fldigi TX ATT with one radio, and a gain of zero with another, but, again, now that I have the notes (shared with other local hams), I am good to go. I see hams looking for help with configuration problems with both devices, and I opine that the confusion stems from a poor understanding of all the devices in the audio stream and how they affect the signal. Not included in this thread are several pages of my notes regarding these audio streams.

Thanks for sharing your ideas.
Kind regards,
~R~

7

my radio to digirig connection list is limited for sure but here is what i have used:
ic-736 (i have multiple)
ic-821h
g90
x5105
uSDX+ v2
htx-10

the above radios (i know, very short list compared to what is out there :slight_smile: ) have worked without issue concerning the input/output gain of the digirig mobile.

the trace on the digirig can be cut for 20db attenuation but i’ve not needed it so far.

i used my HP 8924c with the above rigs as well for testing.
(decades adjusting actual hardware TNCs)

good luck
:slight_smile:
kb0wlf

8

Uh, you guys were busy here…

Regarding Digirig’s output level: not sure what is going on in @TheDustyKey’s case, but the levels are set to satisfy most of the use cases including the sensitive mic inputs on HTs and older rigs (e.g. Kenwood TH-F6A) as well as less sensitive line-level connections through accessory ports (e.g. Kenwood TM-V71A). Between the level/volume control on the computer’s side and gain/sensitivity control on radio’s side there is always a way to get the things where the full dynamic range of signals is used without the distortion. If there is a challenging case, we can look at the specific including the amplitudes supplied and expected.

Were all for combinations tested: DR/SL with TH-F6A/TM-V71A?

9

Hi Denis,
Thanks for chipping in.

Background: As a physicist and engineer, I tend to get techie about everything :wink: I want to have enough understanding to be able to help other hams intelligently, without resorting to “magic”.

Background: ISTR that I was unhappy with the digirig gain months ago when I first started using it. Just like this time, not enough to not use it, just enough to be aware that, for me and my configurations, the gain was/is an issue. But it’s not that simple, as the records below show.

Background: I was asked to provide some info on fldigi audio streams and digi configurations. All I did was write up several docs about the subject, then start recording settings to give other hams an idea of what they might encounter.

satisfy most of the use cases including the sensitive mic inputs on HTs and older rigs (e.g. Kenwood TH-F6A) as well as less sensitive line-level connections through accessory ports (e.g. Kenwood TM-V71A).

Yes, this is the challenge, because we all use a range of equipment, from computers to radios.
For example
Kenwood D72A & digirig 1.9 TX: fldigi TX level -3.0 dB: Playback: Levels: mic=0.0 dB spkr=-27.5 dB
Kenwood F6A & digirig 1.9 TX: fldigi TX level -19.0 dB !!! Playback: Levels: mic=0.0 dB spkr=-27.5 dB
Unfortunately, this is where I started, and was so surprised at the -19 dB needed.
I didna check the V71A with digirig, partly because I already know that the DATA port normalizes audio levels to some degree. Adjusting fldigi TX level 0.0..-7.0 dB has no effect on the F6A RX SIG, illustrating the V71A audio signal normalizing feature.

F6A & SignaLink V1 TX: fldigi TX level -3.0 dB: Playback: Levels -10.2 dB: Compare to digirig above

I don’t consider the -19 dB a showstopper, just surprising, but there is still on the table the unreviewed issue of how much attenuation can an fldigi audio stream take and still be robust. I discuss that elsewhere, but I am not satisfied that I have all the info I need. For example, with a PSK mode, amplitudes are, as I understand it, relatively constant, so they can handle more attenuation than, say, a mode that has varying audio levels. With the varying audio levels, attenuation need to be low enough that the lowest signals are readable and the highest signals are not clipped. I don’t know that range. That level of technical detail is not in the fldigi docs.

– The audio stream starts w fldigi. With a TX attenuation of 0 dB, I assume that the audio is produced at the maximum “safe” level. That is, given the ADC specs (say, 16 bit), it is likely safe to make any PSK audio peak at 15 bits, assuming 1 sign bit. Signed 16-bit audio has a range of values from -32,768 to +32,767. That’s a lot of detail. But it’s really a lot of “volume”, and each piece of hardware in the stream, from the transmitter to the receiver, including repeaters, if any, get to handle that audio stream.

In every case, my goal was to find settings that resulted in filling the SIG band, not the whole vertical display, to 80-90%. There is a whole discussion about audio signal scaling wrt fldigi that is too long for here. But I have noted in several posts that I was first introduced to all this digi stuff with the remarks that it is magic. Several hams with prior experience offered to “train” me to “hear” proper modes. All nonsense. It’s all plain electronics, and every ham should be able to use tools and apps readily available to understand what is happening here. Being a ham is not just memorizing the answers to an exam, which is so popular today.

Were all for combinations tested: DR/SL with TH-F6A/TM-V71A?
No. I am pasting my results at the end of this post.
V71A & SignaLink V1
F6A & SignaLink V1
D72A & digirig 1.9
F6A & digirig 1.9

I have on my list to check the V71 with digirig, and to then check them all with the mini…odd, I just looked for it in the digirig store and can’t find it. Ah…

I see that v1.2, which I have, has “significantly improved PTT by tone/VOX”.

fldigi provides 3 good tools for analyzing audio signals: waterfall, FFT, and SIG. Each of them looks at the audio a different way.

73 Rich NE1EE

fldigi audio settings 20251109a.txt

Sample setting values to give ops an idea of what to look for.
Note the use of named CODECs to make the fldigi config easier.
These are guidelines; settings are OS- and radio-specific.

2025Q2a
flamp - 2.2.14
fldigi 4.2.06 flarq 4.3.9 included
flmsg - 4.0.23
flrig - 2.0.05
flwrap - 1.3.6

2025-11-07 specs are listed in the order they occur in the audio stream

fldigi 4.2.06, Windows 10
Kenwood V71A & SignaLink V1 TX
fldigi TX level -3.0 dB
SL not selected as default comms.
Playback: Levels -10 dB, Enhancements: disable all, 16 bit 48 kHz, allow exclusive=optional, Spatial=OFF
Adjusting fldigi TX level 0.0..-7.0 dB has no effect on the F6A RX SIG, illustrating the V71A audio signal normalizing feature.