Page 1 of 1

Poor audio quality on 10M FM receive

Posted: Mon Jan 23, 2023 7:24 pm
by w-u-2-o
With 10M FM DX starting to show a lot more life, I was, for the first time for me, doing some 10M FM listening. Unfortunately, even large signals were received with quite a bit of nasty crackling sounds.

There are not a lot of adjustments available in FM mode. I tried a variety of settings in Settings > DSP > Options with no improvement.

Any 10M FM aficionados out there using Thetis successfully?

Re: Poor audio quality on 10M FM receive

Posted: Mon Jan 23, 2023 8:07 pm
by ea3aqr
Same issue here.

Re: Poor audio quality on 10M FM receive

Posted: Mon Jan 23, 2023 9:20 pm
by Cosmo1stgen
Scott,

Not exactly the same but I have been on thetis with a transverter down converted to 10m for years on fm repeaters. It worked fine. Haven’t tried it with the latest thetis release so maybe something has changed? I do remember something like this when I first tried protocol 2 years ago did a data base reset.

Ryan

Re: Poor audio quality on 10M FM receive

Posted: Mon Jan 23, 2023 10:50 pm
by KA1GXR
Try wider receive bandwidth filter. For US should be at least 5Khz for wide deviation .It may default to 2.5 khz wide .
Tom
KA1GXR

Re: Poor audio quality on 10M FM receive

Posted: Tue Jan 24, 2023 12:42 am
by w-u-2-o
5KHz sounds better than 2.5KHz but it is still horrendously bad.

Re: Poor audio quality on 10M FM receive

Posted: Tue Jan 24, 2023 2:21 pm
by W2PA
Never tried it until today, Scott.

I'm now hearing the KQ2H repeater on 29.610 (somewhat local) at S8 with a bunch of EU stations using it. I'm not hearing the effect you describe - it sounds fairly good. Of course the various users have better or worse audio but the repeater's signal itself sounds clean.

Edited to add my Options settings:
Buffer size 64, Filter size 2048, Linear Phase, BH-7

Re: Poor audio quality on 10M FM receive

Posted: Tue Jan 24, 2023 3:10 pm
by w-u-2-o
Not at home right now but I will make a recording and post it.

Re: Poor audio quality on 10M FM receive

Posted: Tue Jan 24, 2023 5:56 pm
by w9ac
FM audio sounds good here. Of course, it's bandwidth-limited but I don't hear any obvious distortion, artifacts, nor data breakups. The signal source is an Icom 7610 in FM mode transmitting into a dummy load. Icom audio input source: USB 3.0 sound interface and a studio-grade mic.

Thetis FM DSP settings:

Buffer: 256; Filter Size: 1024; Filter Type: Linear Phase; Filter Window: BH-7

Thetis 2.9.0.6 FW: P2 v2.1.18

Paul, W9AC

Edit: The IC-7610's maximum FM deviation is 4.4 kHz and 3 kHz is the maximum FM modulated audio frequency. Carson's Rule for required FM channel bandwidth results in 7.4 kHz * 2 = 14.8 kHz of significant sideband spectrum.

Thetis' actual Rx bandwidth at the 5 kHz setting is 16 kHz, or 8 kHz either side of center. So, Thetis should be able to demodulate typical NBFM modulation without issue if the sending station's audio bandwidth and deviation is limited to the normal NBFM standard.

Re: Poor audio quality on 10M FM receive

Posted: Wed Jan 25, 2023 9:28 pm
by w-u-2-o
I've been doing some more listening. It seems that I need to see a minimum of 18dB (3 S-units) of SNR before I get full quieting. Anything less than that I get a lot of noise. Worse, the quality of the noise is extremely unpleasant.

I'll publish some recordings when I get time (after this next massive snow storm is cleaned up, perhaps Friday).

Also, I've noticed that none of my audio metering seems to work, from Mic all the way to COMP.

Re: Poor audio quality on 10M FM receive

Posted: Wed Jan 25, 2023 9:38 pm
by w-u-2-o
And it would appear that I've got no modulation on transmit. I'm using VAC. Does FM work with VAC?

Re: Poor audio quality on 10M FM receive

Posted: Thu Jan 26, 2023 2:38 am
by w9ac
w-u-2-o wrote:And it would appear that I've got no modulation on transmit. I'm using VAC. Does FM work with VAC?

FWIW - VAC in FM mode works here with the WDM driver, but not ASIO. However, I am using a modified ChannelMaster.dll file and have full FM audio control and metering with ASIO. I only engage VAC for digital modes with WDM.

The modified ChannelMaster file disables all front panel audio connectivity and digital audio is routed directly to a Presonus ASIO sound interface without engaging VAC.

Paul, W9AC

Re: Poor audio quality on 10M FM receive

Posted: Thu Jan 26, 2023 12:21 pm
by w-u-2-o
It should not require any software mod's.

I'm suspecting a bad database. I probably will not have time to investigate that until tomorrow.

Re: Poor audio quality on 10M FM receive

Posted: Thu Jan 26, 2023 10:14 pm
by w-u-2-o
I made some good progress today on my FM problems.

It was not a bad database.

My transmit problems were operator error.

However, on receive there is something weird about audio scaling inside Thetis when in FM mode. When using VAC as a virtual line output normally RX1 AF (or RX2 AF) should be set to 100. At this setting, when Thetis produces a maximal audio output (0dB) this is the level that is sent out of the VAC interface for processing by external audio applications. This works perfectly for SSB, CW, DSB, AM and SAM. In these modes, with AGC gain set to 10dB above the displayed average noise floor on the panadapter, NR OFF, one sees an audio output of about -15dB peak with just noise. With a signal in the passband, this will climb perhaps as high as -2dB peak (obviously there are some AGC effects in play).

When in FM mode, with RX1 AF set to 100, with just noise the audio output is pinned at 0dB and is badly clipped on negative peaks (but not positive peaks). Obviously any signals that appear in the passband cause the same sort of audio output. However, if I reduce RX1 AF to 30, then the audio level for noise drops to about -3dB peak, negative peaks are no longer clipped, and it sounds OK.

Some oscilloscope traces to follow when I get a chance.

Re: Poor audio quality on 10M FM receive

Posted: Fri Jan 27, 2023 11:23 pm
by N9HLL
Not that it matters, But I also find the FM performance almost unuseable However, The audio levels thru the chain are WAY too high someplace, I just so seldom use it I`ve never tried to figure out where....Lowering the Master AF way down does seem to cure the distortion on receive to a managable level.


Rick (N9HLL)

Re: Poor audio quality on 10M FM receive

Posted: Fri Jan 27, 2023 11:39 pm
by w-u-2-o
I don't find it unusable at all. I had my first 10M FM QSO yesterday and got a glowing signal report, strength and audio, both.

But the problem with receive audio levels is definitely an impediment where weak signal reception is concerned. Set audio gain high enough for the typical low deviation of most stations and the clipping on noise is tremendously disruptive.

It's worth noting that most stations are not driving their transmit audio hard enough for maximum deviation. I don't understand why.

Re: Poor audio quality on 10M FM receive

Posted: Sat Jan 28, 2023 2:36 pm
by w-u-2-o
I had promised some scope traces, so here they are.

I used the Friture app (https://friture.org/) to visualize the digital audio stream. It is quite handy. Here I have visualized both short term and long term audio, and on the right the RMS level is noteworthy (the peak varies too much).

Here is the audio output of noise at 5KHz deviation and RX1 AF of 30:

image.png
image.png (146.31 KiB) Viewed 2686 times


Here is the same thing with RX1 AF at 100. The positive peaks are reaching but not exceeding 1.0 (0dB), i.e. I don't believe there is any positive peak clipping, but the negative peak clipping is obvious and massive.

image (1).png
image (1).png (154.91 KiB) Viewed 2686 times


In contrast, it is impossible to create this condition in LSB, USB or CW modes. For example, I found an S9+20 digital signal. I set mode LSB, bandwidth 3.3KHz (more than enough to encompass the entire signal), NR OFF, AGC Gain max (120), RX1 AF 100. All other settings the same as for FM. The audio output never gets beyond approx. -1.5dB and it is completely symmetrical.

image (2).png
image (2).png (136.62 KiB) Viewed 2686 times


I also took a casual look at AM, SAM and DSB modes. In those modes audio gain is very much reduced from SSB and CW. This is something that has been noticed by others in the past. It would be nice if gains across the various modes could be normalized to that used for SSB and CW. Particulary for FM where it presently hampers weak signal reception.

Re: Poor audio quality on 10M FM receive

Posted: Sat Jan 28, 2023 3:53 pm
by Joe-W4WT
Well, with this discussion I decided to try 10 FM for the first time with my 7000 DLE this moning and like you Scott, find the rx audio noisy on most signals. Adjusted varous things with no real improvement. Heard few that were full quieting out of Europe but maybe that is how it is. Most signals out of Europe on 29.600 were running about an S7 so I assume it takes more than that for full quieting. I'll have to compare it to the 7300 and see if it is the same.

I had no trouble operating VAC on transmit. It works just fine using ASIO from the UMC202HD. I received very nice audio reports using the same profile I use on SSB. Found I had to turn DUP off to see Tx audio in the panadapter. Wasn't sure which deviation mode to use. I noticed it appears most of the Europeans I saw ran 2.5k. Is that considered "standard"?

Didn't operate all that long though due to the Rx noise annoyance.

One other thing I noticed was I had to run the squelch higher at 2.5 dev vs 5.0. Seems like it should be the opposite.

edit: I have now compared the 7000 to my 7300 on various 10m FM signals and they sound almost identical for noise so I guess this is normal.

Joe, W4WT

Re: Poor audio quality on 10M FM receive

Posted: Sat Jan 28, 2023 5:07 pm
by w-u-2-o
I haven't figured out what's what and who's who for 2.5 vs 5KHz deviation. I think you have to roll with that punch. I've been running 5KHz mostly.

Again, VAC or ANAN audio, either works fine, and the SSB audio settings seem to carry over very well. I just use the same transmit profile.

At this point the only real trick is to drop the RX1 AF level to about 30. Any more than that and for large signals squelch tails sound awful, and for weak signals the sound of the clipping really interferes with the human perception of speech in noise.

Re: Poor audio quality on 10M FM receive

Posted: Sun Jan 29, 2023 11:45 am
by DL2XY
Maybe you want to try this workaround:
WDSP.DLL with reduced FM demodulator AF output by 12dB.

wdsp.zip
(890.65 KiB) Downloaded 74 times

Re: Poor audio quality on 10M FM receive

Posted: Sun Jan 29, 2023 5:00 pm
by w-u-2-o
DL2XY wrote:Maybe you want to try this workaround:
WDSP.DLL with reduced FM demodulator AF output by 12dB.

wdsp.zip

That's very cool, Walter, thank you! But -12dB is a bit too much. Would you mind creating a patch version with -6dB, please? I assume (hope) that this is x64, right?

Re: Poor audio quality on 10M FM receive

Posted: Sun Jan 29, 2023 6:40 pm
by DL2XY
w-u-2-o wrote:
DL2XY wrote:Maybe you want to try this workaround:
WDSP.DLL with reduced FM demodulator AF output by 12dB.

wdsp.zip

That's very cool, Walter, thank you! But -12dB is a bit too much. Would you mind creating a patch version with -6dB, please? I assume (hope) that this is x64, right?


Yes, its x64. If someone need x86 gimme a call.

-6dB is already too much under difficult conditions (Weak signal, pulsed QRM, too high deviation or carrier offset), testet with RF signal generator.
But if you like:
wdsp fmd -6db x64.zip
(890.67 KiB) Downloaded 57 times

Re: Poor audio quality on 10M FM receive

Posted: Sun Jan 29, 2023 10:00 pm
by w-u-2-o
Warren came up with a solution that uses a limiter after the demodulator. You can find the patch for Thetis 2.9.0.6 x64 here: https://community.apache-labs.com/viewtopic.php?f=9&t=4486