Print

[AE8GS]

I noticed today while using JS8Call and fldigi (for RTTY) that the receiver audio (from an ICOM 7300) seen by those programs includes the effects of the Noise Reduction filter. It made me wonder whether that is good or bad. If NR is intended for human speech and hearing, then it might not be best for the digital mode modems. In other words, they might work better with the "raw" signal. I didn't spend enough time switching it on and off to tell whether one setting was better than the other.

[K6SDW]

Great questions for the digital wizards!  I believe I've read that NR off, RF gain  backed down and AGC set to fast or off, then let the Joe Taylor software do the heavy lifting...

What sez U FT8/4 guys?

GL/73

[G4AON]

The less processing, of any sort, that takes place in the transceiver the less it will foul up data modes.

Transceivers that have a dedicated data mode will usually bypass any EQ (tone control) and may bypass some of the overall DSP processing in SDR based transceivers.

You can often improve reception by reducing bandwidth, something perhaps not usually considered.

The “Achilles Heel” of SDR transceivers is latency, which is the delay through the processing side of the radio and is measured from the line in to RF out, and vice versa on RX. The more DSP processing the radio does, the longer time it takes for signals to pass through. While it probably makes no difference to slow change over modes, such as FT8 or RTTY, ARQ modes (Pactor, Amtor and to a much less extent Vara) can cease to work with long latency radios, or at least limit the max range.

Fortunately, the humble and popular IC-7300 has incredibly short latency on data settings (around 3.5 to 4.5 milliseconds on TX and RX), while some PC based SDR transceivers are in excess of 100 mS. The IC-7300 has good cooling too.

73 Dave

[K4WH]

No.  Turn it off.   Stop falling into this trap that all these latest features are actually needed.  Most are not.  Learn to use your RF gain.  You will be amazed how little RF gain you need for digital modes.

[WA3SKN]

The answer here is that any distortion to the audio is not good and to turn the NR off.

-Mike.

[KE6SLS]

No.  Turn it off.   Learn to use your RF gain.  You will be amazed how little RF gain you need for digital modes.

I couldn't agree more!  I am very active in digital modes and most often on 80m.  The NR & DSB & NB should all be off.

I run my RF gain on my TS590 at 6, sometimes, especially during heavy lightning crashes, I turn it back to 5.  AGC fast also helps.

The cool gizmos on our radios are made for our ears, not for digital software.

73
Jaye

[VE3PP]

I have been operating digital modes since 2007. Olivia, Contestia, JT 65, FT8.

I never have the DNR or NB turned on when operating those modes.

No problems.

[AA4HA]

It even goes to the differences between using MIC-IN and LINE-IN for audio and where you take the audio from the radio.

Some radios have pre- and de- emphasis circuits; You do not want any of that to be mangling up the audio coming out of (or going in to) your radio.

[G8FXC]

One look at the spectrum display in WSJT-X is enough to answer this question. With all signal processing off, you'll see distinct bands of colour representing the various signals - each pretty well defined. Turn on the DNR and the edges of each band become diffuse, tending to merge together. Strong signals will still decode, but weaker signals close to strong ones may get swamped. I've owned both an ICOM IC-7300 and Yaesu FTdx101D and both behave similarly despite the significant differences in DNR implementation.

Martin (G8FXC)

[K6OK]

One of the authors of WSJT-X, G4WJS (SK), wrote in 2020: "NR on modern rigs and SDRs seems to be harmless for WSJT-X modes so long as the level is set reasonably."  That's been my experience, I leave my NR on and I've not seen any harm to FT modes.  If there is an objective study based on actual measurements showing NR causes a reduction in FT reception then I would be open to changing my mind, but until then, I'll go with what Mr. Somerville wrote. 

[K0UA]

One of the authors of WSJT-X, G4WJS (SK), wrote in 2020: "NR on modern rigs and SDRs seems to be harmless for WSJT-X modes so long as the level is set reasonably."  That's been my experience, I leave my NR on and I've not seen any harm to FT modes.  If there is an objective study based on actual measurements showing NR causes a reduction in FT reception then I would be open to changing my mind, but until then, I'll go with what Mr. Somerville wrote.

My extensive A/B switching with it on and off yields about the same results. It doesn't seem to matter at least with lower settings on my 7300 or 7610, Other than the "smearing" of the waterfall traces as noted above. I usually run with it OFF. But if I accidently leave it on, it doesn't seem to be a big deal or much effect on decodes, just the waterfall "look" and the imprecise actual traces. So six of one and about half a dozen on the other hand. That is the best I have to offer.

[K6OK]

@K0UA -- Here's evidence the NR in the 7300 might improve FT8 decoding: in AB4OJ's extensive lab testing, he fed a -122 dBm signal (6 dB SINAD) to the 7300 under test.  With NR set at a moderate 4 level, he measured SINAD at 10 dB versus 6 dB with NR off.  This equates to a 4 dB improvement in SNR with NR on.  That's intriguing. Additionally, I suspect the 7300's NR would not harm FT8 because the DSP algorithm looks to remove random noise components.  FT modes are a form of MFSK, a series of steady tones with 6.25 Hz shifts.  In theory the NR should not touch continuous signals.

A possible way to test this is to receive WSPR (also a form of MFSK) across multiple days and bands with NR on and off, building a database of perhaps 1,000 decodes, then running a statistical analysis to see if there is a significant improvement in decodes and recorded SNR. With WSPR there's a good chance of hearing the same stations over long periods, some of them are basically beacons. 

[VA3VF]

Here's evidence the NR in the 7300 might improve FT8 decoding: in AB4OJ's extensive lab testing, he fed a -122 dBm signal (6 dB SINAD) to the 7300 under test.  With NR set at a moderate 4 level, he measured SINAD at 10 dB versus 6 dB with NR off.  This equates to a 4 dB improvement in SNR with NR on.  That's intriguing.

Intriguing indeed. I'll give it a try.

Visually speaking, I'm with G8FXC, the change on the waterfall is clear for all to see.

As a side note, I'm pretty sure K1JT used to say in the early days that AGC should be off. More recently I'm seeing people say that it should be set to fast. Has anything changed? What am I missing?

[WA3SKN]

It's all about the "signal-to-noise ratio".
With RTTY you are shifting the carrier.  With most all the newer modes you are using audio tones in the audio channel... about 300Hz-3000Hz in frequency will pass, but any distortions can cause problems.
Just like using a modem over a telephone line, there are certain parameters that need to be met for it to be successful.

-Mike.