FUn with Audio

[KU8L]

GA Group:

I have been busy last few days making some comparisons of different audio chains/mics vs on-the-air-signals.

I have an R-4Bas my active receiver with a Softrock Lite as a first IF panadapter.  Running PowerSDR-IF on an P4 3GHz PC with audigy 2 ZS PCI soundcard.

I am particularly interested in corellating what I hear on-air vs what the signal looks like in the spectrum view.  As you can see, the spectrum views can observe the level of modulation as well as the bandwidth and the distribution of audio across the bandwidth.

I have been amazed at some of the strong carriers with very low modulation levels that are S-9 but not really readable. 

Here is a picture of really fine sounding Kw-1 at S-9 here and very smooth audio



Next picture is my own monitored signal from the Apache.  Audio is direct from Behringer Euro Board 3 band simple EQ into the Apache mod driver--bypassing the preamp, limiter, and BP filter.  Wideband mic yields a pretty nice sounding signal but I wont make any friends!

Just for grins, I decided to try an old Autek QF-1 active filter box.  Final picture is the BW limited signal from that combination--I think it is about as good as I will get without improving the outboard processing.  It would be nice to get a little compression, external limiting, and quieter BP filtering.  So more to do.

The pictures are set for peak hold so you get to see a representation of the maximums in every point but not necessarily all at the same time.

If you hear me on-air, please comment on how it sounds.

FWIW

Curt

[N2DTS]

There is much better software for looking at that sort of stuff.
I have an sdr-iq that runs 3 different programs, spectraview (crappy), hdsdr (nice), and sdr console (also very nice).
With the spectrum and waterfall displays running, you can see just where the audio power is located.
With the passband waterfall display in hdsdr, you get a big picture of the audio in its own sub display.

http://www.hdsdr.de/screenshots.html

Its interesting to see all the different eq on signals.

The oddest thing I ever noticed was a plate modulated rig with audio mostly in one sideband!

[W2VW]

The oddest thing I ever noticed was a plate modulated rig with audio mostly in one sideband!

BC-610?

[KU8L]

I have seen HPSDR.  Not sure I understand the note about seeing where the audio power is.  THe same spectragram, waterfall, and histograms are available in PSDR I just am more used to seeing the RF spectrum.

The amplitude at any frequency is representative of the audio power is it not?

Maybe I'm just missing something.  Anyway, it is fun looking.

When we characterize an audio space, we just feed noise source and look at FFT spectrum...same thing here I think. There are better visualizations available I guess...depends on what you want to see.

Without a wideband noise source test, we can't tell if it is the equipment or the input program that is responsible for the result--I guess thats why guys do checktapes.

I should load up HDSDRand give it another try--I like the simultaneous windows available.

Thanks

Curt

[N2DTS]

What you are doing is great with the peak hold, but with live signals, its easier to see on the waterfall I think, plus you can look back in time...

In the 3rd screen shot, everyone is going to think that is way too narrow, but the filtering looks great.

The audio (mostly) on one side of the carrier was a valiant with vfo problems.
Maybe that is what AM with some FM on it looks like.
With a crystal, it was normal (and less distorted).

There sure are some odd signals on the air, I looked at someone who's audio looked like it was EQ'd to boost every 100 or 200 Hz with the in between frequencies cut, sort of like |  |  |  |  |  on each side of the carrier...



 

[kb3ouk]

It's some kind of phase modulation going on that's causing one of the sidebands to be cancelled out. Look up the Kahn Powerside, it was an exciter for AM broadcast rigs that did the exact same thing, worked on the same principle as Kahn's AM stereo system.

[flintstone mop]

The oddest thing I ever noticed was a plate modulated rig with audio mostly in one sideband!

BC-610?

Dave,
I have seen that too. Not necessarily a BC 610, and it still sounded good.
BTW any of these PowerSDR or any others display the waveform 'scope pattern? There is a 'scope selection but it is the audio from the wave we see using a 'scope and RF pick-up.
NONE of the displays for Flex PowerSDR will show you if your mic audio is out of phase. Or can someone 'splain what I might be doing wrong?

As for the OP. It would depend on the voice characteristics. If the voice is a nice full audio and nice balance the entire display will dance evenly. The movement close to the carrier is the bass and the edges are the high audio freqs.
Fred

[WD5JKO]

Just for grins, I decided to try an old Autek QF-1 active filter box.  Final picture is the BW limited signal from that combination--I think it is about as good as I will get without improving the outboard processing.  It would be nice to get a little compression, external limiting, and quieter BP filtering.  So more to do.

  Hi Curt,

  I too use the QF1-A, but in my case for receiving audio. Its job for me is to take out that Hi-Fi bassy audio people strive for, by doing three things...notch out the low bass, accentuate the mids, and follow with about a 3 Khz low pass. The Autek can do these three tasks well simultaniously. The result is much easier copy. I get to turn it off when folks transmit the audio with these ideas in mind...K4KYV, WA1HLR, and others. Old George W2WLR was on to something way back.    http://amfone.net/Tech/EAM.pdf

  So you used the Autek on transmit? It would seem to present a challenge with the impedance and levels (around 8 ohms in and out?) when using it in the transmitter speech amp chain. You got me thinking to hook up my QF1-A to my audio spectrum analyzer and input it with white noise to look at the various filtering settings.

Fun Stuff.

Jim
WD5JKO

[KU8L]

Hi Jim:

You mean you use the QF1 on receive to make the wide, full audio sound like communication quality, penetrating, intelligble SSB?  What an idea!  

Just kidding, but there is something to my madness....seems to be a lot of folks don't really realize what they sound like feeding wide-band audio to a 50W transmitter.   Unless you are working down the street, It takes a lot of poop to sound like an FM Sleepy Time DJ!

I am thinking that a lot of the Apache's rep for sounding "scratchy" is due to the use of the D-104 without a correct input load R.  I think they had it pretty much correct as far as the audio set-up internally.  Cut-off the really low stuff below a couple hundred Hz, clip so you can run hi levels without overmod--essentially compression, then filter at 3KHz.  That spectrum is pretty darned effective.

Using a different mic makes it sound much different--one on air report was that it sounded like a 32V collins rig----I thinked him!

I wasn't too concerned about the QF1 I/O impedance.  I use the effects send/return path from the Euro Mixer.  Lo Z going into Hi-Z usually works just fine and I think that is the situation on the QF output.  Going in, I don't imagine that there is a 8 OHM input Z on the QF.  It works because it gets a current driving speaker signal going into its moderate (hi) Z input.  Just talking out my butt on that because I have not actually looked at the schematic but it is so common in audio stuff that I took the chance.  It woks fine.  

You must have the A version...mine can only do one of the functions at a time.  It also has quite a lot of AC HUM that new caps did not fix.  But it was kind of a proof of concept for me.

Of course, all is out the window if you are trying to change your voice qualities vs simply reproducing your natural voice at the receiving end.

Regarding the scope display....yes the PSDR has a time domain (scope) display.  It is what is coming out of the SDR.   I find that simple RF sampler with the typical 1N34 detector for monitoring works but leaves a lot to be desired in the true reproduction of the input signal is concerned.  

My thinking is, there a multiple stages of the path that all have different effects on the output--we must understand where the big changes (distortions) take place and make changes we can control to create either a close facsimile of the input, or a copy that meets our ideas of how it should sound.  Because the receiving end is out of our control, and the ears(hardware) and subjective nature of the person(software) is different than our own, the improvement process can be confusing.

In the popular music business, it was common practice to mix masters that provided exaggerated EQ just so the broadcast to cheap AM radios would sound "punchy"---pretty much the antithesis of the ORTF/European one sweet mic--no eq--right to tape methods for classical recordings.

Anyway, its all fun....I'm off to map my actually voice characteristics on the input side.  I don't have anyone close enough and clear enough to do meaningful recording on air yet so will live with putting out what I think is good and wait for the reports to burn me!


BTW....'Splain to me what the mic phasing is all about.     I understand those effects in a multiple mic multiple source environment but not sure I grasp why we need to know in a single mic set-up.  If it is simply to insure that the larger voice peaks due to asymmetry are on the positive modulation part of the envelope, then I get that.  Is there another reason or do I have it?

Cheers

Curt
KU8L

[KU8L]

Well...I have both running and can use either.  It kind of depends on what I am looking to do.  I also switch back and forth between monitoring with the PC output or the R4B.

I find a couple of not so nice things about HDSDR tho.  It doesn't seem to save the frequency settings I have.  I may have not found that yet but it is a pain to have to reset the FREQ and LO settings every time.

It soes seem to use less resources than PSDR but in some cases I find the fexibility of set-up more in-depth with PSDR.

I also like the ability to change the frequency scale to read +/- offsets instead of the actual frequency.

I would like to get one of these working with a radio that had full cat control so it would track the frequencies properly--especially for CW---but all my current radios are up-converted so makes the softrocks a lot more difficult to interface without another mixer like the LP-PAN/FT950.

I have spent so much time in my career looking at FFT signatures, I kind of miss seeing both sides of the audio display in the sub-window on HDSDR....also, it really is a redundant display if you are looking at a zoomed image in the upper window, the lower window is showing same-same.  It is nicer when looking at the whole band segment above tho.

So I switch back and forth.

I really need to get a better soundcard or find the reason for the large central hump in my setup.  It is off to the side since The SR crystal is offset from the R4B IF.

Curt

[ke7trp]

You can set the LO Freq under options and Calibration if the SDR is looking at an IF.

For example, My IF output is 10550000hz.  You can then set offsets for each mode there as well. 

If you run Ham radio deluxe,  There is a built in DDE server.   So you can go to the DDE client setup on HDSDR and click HRD.  Bingo, The Frequency shows up and tracks the radio.  You have several modes also.  I use the LO mode so the freq bar scrolls across. 

You dont have to use the HRD DDE. You can use any that you would like.  The clients are listed in the selection screen.

If you dont want to use any other logging or control program, you can have HDSDR use the CAT feature right to your rig.  I like to run HRD to hunt and log DX so it makes sense for me to use the DDR so my scope tracks.

C