AMers often sound better on the air than in person... why?

[K1JJ]

Some thoughts about AM audio...

Back around 1976 my girlfriend wanted to listen to my headphones while I was testing on AM.  She immediately said that I didn't sound like that in person - I sounded way better in the headphones!

I've often thought about this and noticed the same thing with many of my ham friends.  At the flea markets I could easily recognize the speech patterns, but the big bass and clear highs were not there. Why is this so?  

I think the first thing is in person we are talking 4'-6'  away from someone's ear, thus losing some bass, and subtle sounds from the teeth, lips, mouth, etc.  In contrast, on the air we are talking into the mike only 1-3" away, thus faint sounds are preserved. On the receiving end, many listeners use either headphones to keep the sounds intact or use a big amplifier with large speakers for great sound transfer to the ears.

To enhance it even further, we often use audio limiting/ processing to make up for amplitude deficiencies in our voices as well as enhance or reduce audio frequencies using an equalizer. We can suck out the normal mid-bass mud freqs as well as enhance or roll off the extreme highs.  Bottom line is yes, we can and should sound "better" on the air than in person with the right effort.  Though much will depend on what kind of voice you have to work with in the first place.

Band conditions due to fading, noise or QRM can reduce intelligibility, of course.  Then again,  public settings have their own share of QRM and noise too.... :-)



Another subject:  Adaptive Predistortion:

Like it or not DSP adaptive predistortion is here to stay. It can make a ratty IMD amplifier chain sound pristine, often creating a -55 dB 3rd order wonder.  The bandwidth skirts can become sharp and narrow.  

There is another benefit I have not heard discussed very much... The ability to compress more density of audio into a given bandwidth without spilling over.  

For example, lets say we have +- 6KHz to work with.  And let's say we do NOT want our signal extending past 6.5KHz up the band unless it is way down, like  -40 dB after the 6.5  KHz point. If we use an old school DX-100, Apache, 4-400A modulated by 810s or any other analog AM rig, we cannot have dense audio at 6 KHz and still remain within the 6.5 KHz margin.   Rather, we must start throttling back at about 4.5 KHz to keep the audio energy from spilling past 6 KHz. We may see significant audio spilling into 7 to 8 KHz.   In contrast, with adaptive predistortion sharp skirts, we can go balls to the walls up to the 6 KHz point and have very little energy at 6.5 Khz. Thus we can have a denser overall audio  power band as it approaches the higher imposed bandwidth limits.

An analogy would be like running to the edge of a cliff.  With a standard old school AM analog rig, if we were running at full speed, we could not stop in time - and would go over the edge.  We must slow down first to stop.  But a predistortion rig could run to the edge at full speed and stop before going over.  It's a matter of audio frequency management.  

So in addition to a cleaner signal, we are able to get more high frequency density for a given bandwidth.   More density = more info and intelligibility for a given bandwidth.  And being a good neighbor at the same time.  


So what brought this on?  

I'm in the process of adding a new SCAF  (switched capacitor audio filter) to my old school analog AM rigs in an attempt to approach adaptive predistortion standards.  I'm working with Rick / W8KHK to adapt his Max 295 8-pole Butterworth filter into a stand-alone four chip cascaded chain. I will be posting a thread to continue an old one I was working on 3 years ago with great results.  After a lot of experiments I think the SCAF filter is the best way to go to manage analog rig bandwidth, variable from 4 KHz - 8 KHz on the fly.   I'd like to see these audio filters become available to the AMers once perfected.  More later.

Tom, K1JJ

[K1JJ]

BTW, here is the original thread about building and testing some sharp audio filters on the rig.  It tests the one I built as well as the SCAF filter.  I tested four SCAFs in cascade at one point. As seen from the spectrum pics, we can very well approach  predistortion standards assuming the analog rig IMD is clean to start.


This is a very interesting read:

http://amfone.net/Amforum/index.php?topic=45596.0

T

[W1RKW]

Tom,
Interesting subject and observation.

Something similar and related to some degree. When I was much younger a buddy and I played in a few bands.  Recently, the 2 of us have been swapping music over the internet for about a year especially stuff that is not digitized.  We've been digitizing it and sending it to each other.  We started going back to the day when we were in the bands and he had some raw recordings of us playing live including some VHS video footage.  I had no idea he had this stuff and I was intrigued by it then I was sort of embarrassed by it. Although we were tight, we sounded terrible. The sound was so bad I just felt the recordings were better served in a trash can.  But at the same time, we managed some studio time and managed to cut a piece of vinyl too.  The difference between the raw and processed audio was just like black and white. The studio recordings weren't really the way we sounded but we sounded killer.  What was behind all that, i don't know but I think it most likely relates to what you describe in your post even with processing being strictly analog based back then.  

[w8khk]

BTW, here is the original thread about building and testing some sharp audio filters on the rig.  It tests the one I built as well as the SCAF filter.  I tested four SCAFs in cascade at one point. As seen from the spectrum pics, we can very well approach  predistortion standards assuming the analog rig IMD is clean to start.


This is a very interesting read:

http://amfone.net/Amforum/index.php?topic=45596.0

T

AM'ers often sound better on the air than in person... why?

Tom, coincidentally, I noticed that recently myself.  Possible reasons might be:

1) Many ops have  taken the ARRL Professional Speaking Self-Study Course
2) The op is cognizant of diction and enunciation, and wishes to make a positive impression on his peers
3) AM ops are constantly focused on syllables and words, and avoid modulating their transmitters with the random sounds of other bodily functions
4) The op knows that other AM ops are aware of the efforts made toward sounding great, efforts which go unnoticed if the carrier and one sideband are sadly missing
5) AM ops have the utmost respect for the listener, and make every effort to avoid causing any "ear bleed", which does not seem to be an issue if only sharing a signal report and grid square

Other, possibly inconsequential issues that come to mind are the wide frequency response and low distortion of the entire transmitter, or maybe judicious use of audio processing? I dunno....

[WD8BIL]

AM'ers often sound better on the air than in person... why?

Might be the beer!

[K1JJ]

heheh..... FB, OM!

And for another angle to enhance this point, let's change the title to:  "SSBers often sound  WORSE on the air than in person.    Why?"

Isn't this true?  If  some unfortunate soul walked around in person talking with a typical ssb voice, IE chopped off from 300Hz to  2.7 KHz,  we might do a double take and then pity the guy for having a speech impediment.      


Thank God we are able to communicate on a mode with decent fidelity.

I go back and forth from ssb to AM over time.  After a stint on ssb, when switching back to AM I am shocked.   And, on AM for a while and switching back to ssb, I am stunned and convinced something has gone wrong with my ssb rig.  It can't really sound THAT bad!    

The exception is the ESSB guys -  some who sound absolutely marvelous at 4 KHz.  There are a handful that have "the voice" and optimized gear working soooo FB.

T

[KD1SH]

Quotes from Rick, K8KHK:

"1) Many ops have  taken the ARRL Professional Speaking Self-Study Course"
Okay, I'll admit, I actually Googled that.

"2) The op is cognizant of his diction and enunciation, and wishes to make a positive impression on his peers"
Yes! I actually started - perhaps subconsciously - practicing that way back in the 70's, in my CB days. It's not just striving for clear enunciation; it's almost a sort of "organic compression" wherein you try to speak in a way that keeps the meter or scope or whatever showing "density".

"3) He is constantly focused on syllables and words, and avoids modulating his transmitter with the sounds of other bodily functions"
Well, there are other schools of thought...

And speaking of the old CB days, I often found, upon meeting fellow CB ops, that they probably looked better, and perhaps smelled better as well, on the air.

[K8DI]

Couple points to toss in here --

Most of the microphones we use for radio are directional, by design.  A side effect of directional mics is something called proximity effect, which boosts low end.  Most of the mics we use are NOT flat at all, they are designed with various emphasis at the upper mid/high range to increase "presence" or enhance the fricatives to improve intelligibility. So, you sound different through a microphone, "better", no matter how flat the rest of the system is.

Second, on the filtering:  one word: digital.  If we can stop being scared of it we can use a proper digital audio chain. You want a 52dB/octave low-pass? No problem.  Cheap-ass DSP boards are available. There's configurable audio DSP's from the professional AV install world on eBay, although they've become less common/more money over the last couple years as people have bought used when no chips were available for new.  Check out what you can do with this: Dayton Audio/Parts Express cheap DSP boardhttps://www.parts-express.com/Dayton-Audio-DSPB-K-DSP-Kernel-Board-and-DSPB-KE-Kernel-DSP-Expansion-2-In-3-Out-325-1300?quantity=1  Necessary programming interface https://www.parts-express.com/Dayton-Audio-DSPB-ICP1-In-Circuit-Programmer-USB-325-132?quantity=1  Configurable Pro DSP eBay example  https://www.ebay.com/itm/255957507695 or (my preferred brand) https://www.ebay.com/itm/266116665016

Ed

[Todd, KA1KAQ]

AM'ers often sound better on the air than in person... why?

Might be the beer!

I'm with Buddly. It's the same reason ol' Hortense down at the end of the bar looks better toward closing time.

The longer you listen, the better they sound.......?

[WA2SQQ]

The logical extension of this topic is why our own recorded voices don’t sound the same as the way we hear them. I understand why, but after all the EQ’ing and processing, are you hearing the way I’m hearing it? Now add it the element of hearing loss … Oy!

[W1DAN]

proximity effect

[WO4K]

It's the 20 minutes of voice exercises AMers do right before calling CQ, such as:

Should saucy sharks seek shelter soon? Or
A proper cup of coffee from a proper copper coffee pot.

Wait. What? YOU don't do vocal exercises?

[W7TFO]

A lot of us look better on the air than in person, too...

 

73DG

[K1JJ]

Just an update on the filter project:


"So what brought this on?  

I'm in the process of adding a new SCAF  (switched capacitor audio filter) to my old school analog AM rigs in an attempt to approach adaptive predistortion standards.  I'm working with Rick / W8KHK to adapt his Max 295 8-pole Butterworth filter into a stand-alone four chip cascaded chain. I will be posting a thread to continue an old one I was working on 3 years ago with great results.  After a lot of experiments I think the SCAF filter is the best way to go to manage analog rig bandwidth, variable from 4 KHz - 8 KHz on the fly.   I'd like to see these audio filters become available to the AMers once perfected.  More later.

http://amfone.net/Amforum/index.php?topic=45596.0

Tom, K1JJ"



UPDATE  2/11/2023:

We are trying to decide which audio filter approach to take.  The Maxin  Max295 chip is an 8th order Butterworth that has excellent roll off charactoristics, but the ripple is not great when four are in cascade.   In contrast, the Max 296 is a Bessel which has nice ripple specs but less sharp roll off.   We will be testing whether four of each, or two of each or whatever combination is the best when four are cascaded.  (an incredible 32nd order filter)    I will not be satisfied until I see a waterfall drop-off that is reasonably clean....   We will also incorporate a toggle switch that allows two or four chips in cascade just in case four is overdoing it.   The chips are about $14 each, so it's a little over $112 + shipping for the experiment.  Stay away from eBay and the fake chips. Buy from DigiKey or Mouser.


BTW, yes, mike proximity has a lot to do with the enhanced sound on the air.   Our voices change throughout the day. I notice that I can talk across the mike in the morning, but need to talk more directly into it later in the day as my voice loses its lower frequencies.... just to maintain the same low end.  


T

[Opcom]

...
The Maxin  Max295 chip is an 8th order Butterworth that has excellent roll off charactoristics, but the ripple is not great when four are in cascade.  
...
T

Does it make any sense to slightly 'stagger' the frequencies of the four Max295 chips to smooth the ripple, or would that ruin the steepness of the slope?

[w8khk]

...
The Maxin  Max295 chip is an 8th order Butterworth that has excellent roll off charactoristics, but the ripple is not great when four are in cascade.  
...
T

Does it make any sense to slightly 'stagger' the frequencies of the four Max295 chips to smooth the ripple, or would that ruin the steepness of the slope?

That is a good question, Patrick.  I never considered that option.  Of course we would then need two or more clock generators, running at 50x the desired filter bandwidth.  We had to add a butterworth LPF after the SCAFs to expunge the clock residue from the audio.  I think a concern of this approach might be heterodynes from the multiple non-sinusoidal clock signals mixing within the signal path. 

One thing I would like to test would be a mix of 295 and 296 SCAFS; possibly improved ripple, not at the expense of slope steepness.  A four-SCAF prototype filter test is in the short-term plan.

[K1JJ]

Maybe even a Butterworth/Bessel  295 > 296  > 295  >  296  alternating sequence would help to smooth out the ripple and maintain the steep slope.

Then again, a full Bessel  296>296>296>296 sequence may make a great ripple spec, but using FOUR in cascade may be what's needed to create that sharp roll off  at the same time.

After using my analog 5.5 KHz filter, I see it can use some more skirt roll off.  To get that SDR predistortion skirt look will certainly take some doing.

As for a low IMD transmitter, I am impressed by my recent tests using a single class C 813 -  class A series modulated by a pair of 813s, driven by the GFZ MOSFET audio driver..  It will easily modulate over 190% positive with a natural soft negative peak limiter effect...  and at 150% is super clean driving a robust linear amp in the "class C AM mode."   I'm looking forward to teaming it up with the new SCAF filter to approach an SDR skirt look.


T