Actual 4-1000A plated modulated on-air AM recording for evaluation (short)

[K1JJ]

I'd like to get a hi-fi ADJUSTMENTS audio dialog started. This is an area many of us struggle with.  I find recordings are very valuable to get us in the ballpark - then opinions help us further.

Some general opinions of my recorded audio would be helpful.....   I'm in the last stages of making recordings and also getting on-air listener opinions.

This is my homebrew class C  4-1000A plate modulated by a pair using the  528e preamp >  DSP graphic EQ >   DSP 6.5 KHz  LP filter >   DSP  parametric EQ  > DSP limiter 5 DB > GFZ MOSFET audio driver board > 4X1 modulators... -12 DB audio negative feedback.   No compression, into a real antenna recorded off the air.  The second EQ is to help smooth and round out the first's lack of range.

Listen on both smaller computer speakers and large speakers to demonstrate the compromise of lows, depending on receiving systems. It makes a HUGE difference listening on a big amp and large speakers.  But I wanted it to sound decent on any smaller system too.  I can hear the lip and tongue ticking at times which shows clean highs.

On big speakers you will hear a faint hum which is in the receiver itself. Also, the blower noise is something I am working on and is reduced somewhat near the end of the recording using a noise gate... (downward expander)

Personally I am pleased with the sound. My goal is for a natural, high dynamic range, no processor pumping, no fake bass, full FM carrier quieting, crisp highs sound.  The transmit bandwidth is limited to +- 6.5 KHz using a DSP LP filter.  I am recording/receiving it in +- 5.5 KHz receiver bandwidth, just to make it realistic to an average on-air receiver.

Please add your opinion. Also, add some recording samples of your own rig(s) -  list your own objectives and results.  

*On the big speakers, I'm thinking the 150 Hz area needs to be cut slightly...

Maybe I need a reverb?  

Thanks.

Tom, K1JJ


Recording:  140% positive peaks at times, according to REA mod monitor...

[KK4YY]

Hi Tom,

Listening in my Sony MDR-7506 headphones, The esses are a bit too loud for me. So, maybe cut the very high end just a bit or de-ess it. Other than that, no problemo.

Don

[WD8KDG]

Hi Tom,

Listened to the recording, not bad on a computer. But have to question the definition of hi-fi. You have stated a plus or minus 6.5KHz. In my neck of the woods, you ain't hi-fi unless using somewhere near 20KHz. Then there is the positive peaks.

In another thread, it pointed out many of our communication receivers are not compatible with anything over 100% positive. This all leads up to my thoughts, I don't own a hi-fi AM communication receiver. It takes two to tango.

Yes I can open the IF bandwidth to 16KHz and I also get to listen to slop bucket that has a 3KHz bandwidth cutting through the AM bandwidth hogs. Not to mention anything over 100% positive peaks start to sound like crap.  

Craig,

[K1JJ]

Hi Tom,

Listening in my Sony MDR-7506 headphones, The esses are a bit too loud for me. So, maybe cut the very high end just a bit or de-ess it. Other than that, no problemo.

Don

Hi Don,

Thanks for the report.  I was wondering the same thing about the highs being too hot. But I don't know how much is lost on the air during rough condix  -  plus many of us are starting to lose our high frequency hearing.  Is this form of pre-emphasis needed to have some punch during nighttime condix?  I dunno and just have others' opinions to advise.

Right now the 6.5 KHz LP filter is rolling them off pretty well and the EQ is cut after 5.5 KHz too.  But being too hot before 6.5 KHz is the issue you bring up.

I'll get some more on-air reports about the highs now that you brought it up.

T

[K1JJ]

A friend just sent this custom photoshop to me... couldn't resist posting it - 

[KK4YY]

Tom,

When trying to limit bandwidth, I think our inclination is to get the most out of the high-end that we do allow, perhaps even over-emphasizing it. Though these frequencies need to be present for clarity, too much emphasis on them can add a harshness to the sound. That's the sense I'm getting when listening to your recording. If I turn the level up enough, I winch when those highs from the esses come through... ouch. I think if you had less high-end, it would bring out the lower frequencies by letting the overall level of received audio to be increased by the listener.

That said, a narrower receiver bandpass may roll those highs off and "correct" the sound. There's no way to control the receiver the other guy is using. So there we are.


Don

[MikeKE0ZUinkcmo]

Pluss one on the fotoshop pic.

[N1BCG]

I’m ready

[K1JJ]

Tom,
When trying to limit bandwidth, I think our inclination is to get the most out of the high-end that we do allow, perhaps even over-emphasizing it. Though these frequencies need to be present for clarity, too much emphasis on them can add a harshness to the sound. That's the sense I'm getting when listening to your recording. If I turn the level up enough, I winch when those highs from the esses come through... ouch. I think if you had less high-end, it would bring out the lower frequencies by letting the overall level of received audio to be increased by the listener.
That said, a narrower receiver bandpass may roll those highs off and "correct" the sound. There's no way to control the receiver the other guy is using. So there we are.

Don

Don, you bring up some good points.  I remember in the past I would always have to boost the highs more than I expected in the monitor. To get thru a trip in the ionosphere and thru the filtering in receivers took its toll.  If I didn't boost the highs a bit hotter in the monitor, I would get a mid-bassy sound on the air.  Hot highs are often warm when they come outa the receiver.

The recording I posted came off my SDR, directly receiving the 4X1's RF, but still not taking a hit like a round trip would do.  What I need to do is have my buddy do a recording in NJ and email it back. He uses an R-390 off the detector, so will represent a typical ham receiver.

Posted below are the recordings I did today and an older tape from 3 years ago of the same 4X1 rig.    The older tape "calling W2NBC"  was recorded from CT to NJ.  I actually like the low end better in the older tape (EQing?)  but the highs in today's tape are cleaner. (though hotter as you said)   The difference is the old audio "hardware" chain  vs: today's DSP chain.

The blower noise was much lower 3 years ago... I think I was eating the mike, thus better S/N ratio.  I hate blower noise and will get it fixed somehow.

I'll leave the highs as they are until I get a recording off-air from 150 miles away.  It will be interesting to compare the two tapes; one in-house and one after the ionosphere trip and thru the other receiver..

T

Below - Now and back then:

[KK4YY]

Don, you bring up some good points.  I remember in the past I would always have to boost the highs more than I expected in the monitor. To get thru a trip in the ionosphere and thru the filtering in receivers took its toll.  If I didn't boost the highs a bit hotter in the monitor, I would get a mid-bassy sound on the air.  Hot highs are often warm when they come outa the receiver.

The recording I posted came off my SDR, directly receiving the 4X1's RF, but still not taking a hit like a round trip would do.  What I need to do is have my buddy do a recording in NJ and email it back. He uses an R-390 off the detector, so will represent a typical ham receiver.

Posted below are the recordings I did today and an older tape from 3 years ago of the same 4X1 rig.    The older tape "calling W2NBC"  was recorded from CT to NJ.  I actually like the low end better in the older tape (EQing?)  but the highs in today's tape are cleaner. (though hotter as you said)   The difference is the old audio "hardware" chain  vs: today's DSP chain.

The blower noise was much lower 3 years ago... I think I was eating the mike, thus better S/N ratio.  I hate blower noise and will get it fixed somehow.

I'll leave the highs as they are until I get a recording off-air from 150 miles away.  It will be interesting to compare the two tapes; one in-house and one after the ionosphere trip and thru the other receiver..

T

Below - Now and back then:

I do like the older recording better, for the reason you stated.

I'm curious how the ionosphere would effect the EQ'ing of the audio... if that's what you're saying.

One thing the two recording have in common is that you are vocally emphasizing the highs (sss, sss, sss, sixsss, sixsss, sixsss). I think a more conversational recording might be easier to judge than all that hissing. I think it may be triggering a primitive brain response in me. Snake! Snake! Snake! 

[K1JJ]

Don,

The ssssss, chhhhs,  ffffffs and thhhhhs  are to look for spitting and distortion that often rears its head when pushed hard. Each sound is centered on a certain highs frequency range.  IE, if you can hit it hard without those frequencies flat topping, going parasitic, spitting or whatever, they pass the acid test.  You can also hear when they are clean. Guess I'm used to it by now.

As for the EQing and ionosphere, atmospheric static and fading into the noise, I think a more aggressive highs curve will make the trip thru better.  I like running more balanced audio in the daytime and ramping up the S curve in the evening on 75M.  It's a mild take on DXing audio.

T

[K1JJ]

Based on comments, I made some changes:

1) Flat mic, no cut

2) less highs - still slightly hot, but need to check on-air reception -  could be reduced further if needed

3) One DSP parametric EQ now using different curve

4) Bigger low end with rumble when boom is tapped


5) Just 3 DB in DSP peak limiter

6) no compression


7)  + -  5.5 KHz transmit bandwidth

8]   much less blower noise

I have a desser available if the highs are still too strong, but I prefer not to use it.


New Recording tapped off RF into SDR:

[KK4YY]

Perfect. Nicely balanced, to my ears. Sounds big in my headphones with the added bass. Highs are present, for clarity, but not overwhelming. Break off the knobs!

[N1BCG]

Anyone else hear distortion that's most prominent at :19 , :27, :50, and several other points? Apologies if it's at my end...

[K1JJ]

Anyone else hear distortion that's most prominent at :19 , :27, :50, and several other points? Apologies if it's at my end...

Hi Clark,

Yes, there is a momentary peak distortion, especially at :27  and :50.

I heard it right away but wanted to go for the EQ balance first.  This test was made with the FT-1000D. The 4X1 and FT-1000D sound identical in normal operation so I do my initial ballpark testing with the 1000D.  It was probably because I have little peak limiting or the linear amplifier is clipping... nothing to be concerned with right now.

I also hear some noise-gating to hide the blower noise. I'm thinking (and hoping) it will be masked when band noise is present.


Don, thanks for the EQ report. I can see that the "sound" is achievable now, so I just need to find a compromise on the air.

Thanks much, guys!

T


***  BTW, when someone forgets your name, just sing them this song...

https://youtu.be/noGjJyEDm5s?t=135

[K1JJ]

Here's the latest recording.   I optimized it again and you will hear it fresh off the SDR.

This afternoon and early evening on 3875 I put the FT-1000D on and then the 4-1000A plated modulated rig.  I ran into the Tron.  He said the new settings (from the audio file below) needed a 3 DB 150 Hz cut of mud frequencies and also needed about 3 DB more 2-4 KHz boost.

Listen to the file below and you may agree it sounds good as is.  So this kinda tells me what we hear in our own monitors and what comes out of a listener's speakers and into his ears may be different.

I will get a recording from NJ soon to hear the difference myself.

Throughout the afternoon I received more favorable on-air reports.

T

[KK4YY]

Sounds good to me.

By the way, here's the link that will make "You Know My Name" start at 2:15.
https://youtu.be/noGjJyEDm5s?t=135

[Opcom]

Listening description:

Listened several times to each file on heavily-equalized 4-way "hi-fi" speaker system on a 50-60W per channel solid state amp. It reproduces down to about 28-30Hz and is reasonably flat, actually only the long-excursion 10" woofer in the pair of sealed enclosures is EQ'd to make it do that low range. Of course it is not a loud system overall with all the power being eaten up by EQ-ing the speakers, but it is very clear sounding. EQ is by the DSP in the PC sound card audio. I do not believe I have 'over-emphasized' the bass but just flattened it without midbass humps as best possible to match the sealed boxes.

I listen to classical and rock on it, & think it's pretty good for a pair of PC speakers but an audiophile with good ears would find a few things to complain about, and one that was too nosy and had good eyesight would make a litany of complaints then puke. So that's it, that's my junk, now to my opinion as an uneducated listener:


========================

K1JJ Tron Approved 3-14-2020 Beaitius.mp3 (581.67 KB - downloaded 48 times.)
 - Sounded clear, with balanced bass in voice and good treble without peakiness or excessive emphasis in the 400Hz-1KHz range which is a common sin.
 - criticism - too 'essy' -sibilance. Bass lacks very lows (announcer lows), but they may not be present.

=========================

K1JJ Now vsThen-K1JJ.mp3 (2669.39 KB - downloaded 29 times.)
 - sounded very natural. Just a trifle boomy in the upper bass. Too much sibilance but maybe not as much as the above file.

==========================

K1JJ Test 3.mp3 (607.57 KB - downloaded 33 times.)
this sounded better. Heard the boom tones clearly. Much better de-essing, but still clear. If you could have just a little pre-emphasis without adding back the esses that might be a trifle clearer, but please don't bring back the esses.

Also, the 3 to 5 'pulses' that start off each test sounded 'nicer' than in previous tests but I guess that is an unfamiliar artifact so it means little to comment on it.

==========================

K1JJ Test5.mp3 (569.14 KB - downloaded 12 times.)

I like this very much. It seems best so far, the 30Hz cut is noticeable on the boom tones but don't seem to affect the voice. Hearing this makes me go back and notice a very slight or un-natural mid-bass boominess in the Test3, and the smoother bass in this Test5 result. Highs seem the same as Test3.

+++++++++++++++++++++++++++
opinion: test 5 is a keeper for this local listening kind of test.   It seems the increment of improvement is small between Test3 and Test5. I can't be more specific except tat this is the clearest and most naturally pleasant so far.
If it were mine and I wanted more low bass added to the voice without mid-bass boom then I would want to treat it like EQ-ing a subwoofer rather than a woofer.
If you do not naturally have those tones in the voice, then there is no way to add or create them or somehow bring them up, and have it sound more natural or better than it does at this point.
It's subjective here and if someone says I'm full of it they might be right.

[K1JJ]

Listening description:

+++++++++++++++++++++++++++
opinion: test 5 is a keeper for this local listening kind of test.   It seems the increment of improvement is small between Test3 and Test5. I can't be more specific except tat this is the clearest and most naturally pleasant so far.
If it were mine and I wanted more low bass added to the voice without mid-bass boom then I would want to treat it like EQ-ing a subwoofer rather than a woofer.
If you do not naturally have those tones in the voice, then there is no way to add or create them or somehow bring them up, and have it sound more natural or better than it does at this point.

Wow, that's an amazing analysis, Pat!   A good thread for the archives.

That's a lot of work and I appreciate it.   Yes, being subjective in opinion is what makes this art so difficult. Adjusting all these parameters to optimize a voice is not a hard science.

I agree that Test 5 is a good baseline.  It's up to me now to get real-whirl on-air recordings to refine things further.  Using the Test 5 DSP file, Tron told me to suck out -3 DB at 150 Hz and add 3 DB in the 2-4KHz area and I did.  I also notice I get the best reports when I run no compression and just a safety net of peak limiting. High dynamic range seems the smoothest on people's ears.

The rapid five pops on keyup is from the station sequencer timing that cannot be heard on the air, just on the computer soundcard - the SDR receiver is quiet in my headphones.  I am told there is a quiet on-air PTT keyup and release.

I know now to watch for too much ssss's, watch the 150 KHz mud frequencies and also carefully balance the 500 to 2K area.   At ~   1- 2KHz  it's too easy to have too little presence creating dullness and lost brilliance - or too much creating a hollowly sound.

I found that an interesting effect is to boost the 35 - 45Hz area and DSP cut it below 34 HZ. There is some voice resonance that is weak but can be gleaned at that low freq. As long as the curve drops after 70Hz into the 200 KHz mud abyss, it can work.  I forced myself to use an 8-band parametirc EQ for the first time. I've always liked the 31-band graphic EQ and thought I couldn't live without it, but the 8-band parametric using a DSP computer screen to change the curves with a mouse is pretty slick.

Yesterday as I talked with the guys on 3875. I really liked the look on the scope, the sound in my off-air monitor and the unsolicited reports on the 4-1000A rig. I simply liked the overall feel. It's all starting to come together.

There was another AM QSO going on 3885, 10 KHz away. I was able to match the sidebands of most of the stations in both QSOs who were running about +-6Khz. With my own DSP filter, I brought it down to +- 5.5 KHz on the fly using the mouse.  It was cool to watch the two QSOs with the sideband edges just touching each other at times. Except for a slight IMD bleed-over, the skirts held up pretty well, which is not that easy on a big rig.

Thanks again for the really thorough analysis, Pat!  (very cool audio system you have)

T

[N1BCG]

You know, Tom, if you really want to be pioneering in contouring your audio to everyone else's opinions, set up the first web controlled dehyellifier where ops can go to your site and adjust the various audio settings to suit their tastes. Mic gain, equalization, compression, bandwidth you name it, every parameter. Creating a free account lets users save their favorite settings so you would sound the way they like each time they log in during your QSO.

In fact, you could even use a button panel to select the various ops' settings so when you turn it over to someone, you can tap on their preferences first.

Some might dismiss this as you becoming the frequency "chew toy", "lab rat" or "dress-up doll" so it's important to rise above the critics and be the first U-Set-It hamateur radio operator!

[W2NBC]

Preliminary results of the K1JJ Audio Adjustments Connection system:

[KK4YY]

I'm suddenly concerned that photo-shopped images suggesting what a "K1JJ Dress-up Doll ®" might look like will begin to appear in this thread.

[N1BCG]

Preliminary results of the K1JJ Audio Adjustments Connection system:

Would everyone please stop fiddling with Tom? We're trying to have a serious QSO here!

[K1JJ]

You guys are a tough crowd.

Great app, Jeff!  I figure now we can get every ham in the country to eventually approve my audio.  

And the ones that don't, watch for sliming at the next flea market!

Hey, even The Tron said my audio has a good beat and is easy to dance to.

T