The AM Forum
March 28, 2024, 05:20:09 AM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: The Phase-O-Matic ... Changing audio polarity multi-band  (Read 5209 times)
0 Members and 1 Guest are viewing this topic.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« on: February 28, 2020, 02:43:32 PM »

Or, in the 1950's maybe it would be called a Phase-o-Tron?   Grin

Background:

This idea has been around for years...  splitting the audio into six bands, DC-100 Hz, 100 - 500 Hz, 500- 2000 Hz, etc.  Then using a 180 degree polarity switch on the output of each band, to individually flip phase. Then combine the six phase-adjusted band channels into one to feed the transmitter / audio chain.

This would allow an individual operator to optimize his audio polarity for his own voice and stack the peaks in a more desirable manner.  This would not be a good idea for broadcast stations where an all-pass filter makes the audio symmetrical for many different voices.


Implementation:

How many times have I heard someone say that their highs are out of phase with their lows... and then just compress or limit it away. I have that same problem - but I do love the lows polarity arrangement that now exists because of the one mic polarity switch. But I want more control.

In addition, instead of a 180 degree toggled phase shift, what if we could TUNE the phase shift for each band with a pot from 0 to 180 degrees, hitting 22, 45, 67, 90 degrees, along the way? After all, the phase change required will not always be a simple 0 or 180 degrees. It will likely be a more random mix. What kind of waveform could we build that might sound better than what we have with a simple single mic polarity switch?

I'd like to try something like this. I need some ideas though. How about using an old 6 band EQ and put an emitter follower on each band output that we can tap either 0 or 180 degrees from... and pick the right phase with a switch... then combine the six paths.  It's somewhat akin to changing the sliders on an EQ and hearing a different sound - there is some phase shifting going on too.

How would someone make a variable audio phase control for each band?  

If we homebrewed the six band splitter would this be a simple bandpass filter for each audio band?

If anyone has modeled or given the idea some thought, please chime in and I will take the best ideas and build it. WTF, it's worth a try.

Tom, K1JJ

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
KK4YY
Member

Offline Offline

Posts: 565


Your best isn't as good as you can be.


« Reply #1 on: February 28, 2020, 03:01:18 PM »

This concept had occurred to me, but I wondered how it could be implemented. Upon recombination, where adjacent frequency bands overlap, there would be cancellation if the phases where 180 degrees apart. This would produce drop-outs of those overlap frequencies. I don't know of a way around that in an analog environment. You'd need some pretty darn sharp filters!


Don
Logged

All your worries won't add a day to your life, or make the ones you have any happier.
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8308



WWW
« Reply #2 on: February 28, 2020, 09:44:15 PM »


A narrow notch here and there may not be audible or detrimental on the other end, but knowing about it would bother me until I knew it was harmless to intelligibility.

There are lots of old 5-6 channel car audio EQs. Almost free once the user has blown out the "200 Watt Per Channel" 8-watt audio IC, which is most of the time.

Make a recording of your voice how you speak on the air, play it into the setup and adjust it for minimum peak-stacking with your voice (trying to keep the frequency response flat except for notches or additions), then compare it to the original. Ask others to judge them in an a/b test?.
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
KK4YY
Member

Offline Offline

Posts: 565


Your best isn't as good as you can be.


« Reply #3 on: February 29, 2020, 08:45:22 AM »

Maybe a step in this direction might be to split the lows from the highs, as with a crossover, and apply an all-pass filter only to the highs letting the lows be asymmetric and properly phased. This could probably be done in an analog environment, whereas multi-band phase correction might better be implemented with DSP (and its attendant latency).

It's a challenge.
Logged

All your worries won't add a day to your life, or make the ones you have any happier.
K8DI
Contributing
Member
*
Offline Offline

Posts: 423


« Reply #4 on: February 29, 2020, 09:26:11 AM »

You guys are killing me here...

Complicated and precise audio manipulation in the analog domain is hard. The precision and odd-valued resistors and capacitors alone are a challenge. Circuit changes mean PCB changes, time and money and more parts...

Modern open architecture audio DSPs are easily configured to do this stuff. Since these units get put in corporate environments (as well as stadiums, churches, auditoriums, theatres, etc.) they show up on eBay every time a boardroom gets remodeled for the next wall street CEO.

I've attached a screen shot of an enhanced variation of your audio ideas I created over a cup of coffee. The DSP unit is a BSS London; their basic BLU-50 unit is about $1400 list price, new -- much cheaper used on eBay. They're designed to run 24/7, they last forever if you replace the fan every decade or so.

The best things about doing it this way are that completely reworking your processing is easy and fast, you can tweak any parameter in real time while passing signal, and when your design is done, you can store multiple presets and recall them with a contact closure switch (or a computer, or a serial command from a control system), you can add a pot/knob to control any variable anything in the box with a standard knob, and it just works.

The screen shot shows some of the control windows open. The design  shown allows for multiple mics, multiple transmitters without changing connections, just recall the combo you want with a preset. There's a gate and a leveler before the band-splitting, and a look-ahead limiter after the recombining. By adding a delay to the limiter's sidechain input, you can have the limiter decrease gain before the peak gets to it -- it is totally overshoot-proof, and brick wall. Also note the filters -- Neville-Thiel Method, 52dB/octave. Also available, 6/12/18/24/36/48 dB/octave, Bessel, Butterworth, Linkwitz/Riley. The design here uses 18% of the available horsepower in their smallest unit. There's also an entire logic section available with gates, inputs, outputs, preset recalls, delays, etc. you can use for sequencing. I use an Agilent signal generator as my station VFO; I use the logic section, with my foot switch PTT to turn the RF output on and off over an RS232 connection, because there's enough bleed that it gets into the receiver. Crazy stuff like that is easy to implement.

What's more, you don't have to use the limiter. Or anything other than the split and phase and recombine. Every block has a bypass button. As an experiment with audio processing tool, you can't beat it.....

Ed



* multiband example.jpg (1802.3 KB, 2719x1386 - viewed 351 times.)
Logged

Ed, K8DI, warming the air with RF, and working on lighting the shack with thoriated tungsten and mercury vapor...
KK4YY
Member

Offline Offline

Posts: 565


Your best isn't as good as you can be.


« Reply #5 on: February 29, 2020, 12:16:58 PM »

Ed,

I have a Symetrix 322 DSP audio processor wired to a TS-440S that works well. It uses an RS232 interface and has four preset configurations. A USB to RS232 cable lets me program it with a PC. The software has a similar GUI to your BSS screenshot. Very easy to use. It doesn't seem to have a way to manipulate phase, though.

The BSS stuff looks to be much more advanced than my old Symetrix 322. Even Symetrix has more advanced hardware than the 322. And yes, I took advantage of the used market to get it. I spent less than $100 for the unit and power supply on eBay.

I also have a Behringer 8024 DSP that has menu driven programming which is a real PITA to use. Right now all it does is sit here looking pretty in the rack.


Don


Edit:
One thing to keep in mind is that the software for some of these high-end processors may require licensing. Finding that out after you bought the hardware can ruin your sunny day.
Logged

All your worries won't add a day to your life, or make the ones you have any happier.
N1BCG
Member

Offline Offline

Posts: 836


« Reply #6 on: February 29, 2020, 12:45:31 PM »

All this effort for asymmetric peaks? How much modulation energy is contained in half-cycle peaks?
Logged
KK4YY
Member

Offline Offline

Posts: 565


Your best isn't as good as you can be.


« Reply #7 on: February 29, 2020, 01:20:15 PM »

All this effort for asymmetric peaks? How much modulation energy is contained in half-cycle peaks?
Gotta love a good rhetorical question.

The issue is not about having big positive peaks. The problem is having big negative peaks — you have to lower the gain so as not to over-modulate. It you flip a negative peak to positive, you can raise the overall gain without over-modulation. Making the audio symmetrical with an APF does solve that problem, but flipping the negative peak to positive solves it too.

So you need to ask the right question. Grin
Logged

All your worries won't add a day to your life, or make the ones you have any happier.
W2NBC
Contributing
Member
*
Offline Offline

Posts: 327



« Reply #8 on: February 29, 2020, 02:05:34 PM »

“Phase-O-MATIC”..  attempted via an interesting processor, the BL-40 Modulimiter. Unfortunately, the polarity switch (flip) was audible (pops), and most engineers just turned that function off.

 If you sharply bandpass sections of the audio spectrum of course you could flip polarity individually.. but there are problems! The human voice is not a fundamental tone producing a predictable sine wave.. It’s a biological signal generator that’s a waveform hot mess.. The HIGH “eeee” into a mic that may be out of “phase” still contains harmonic content from the LOW “OOOO’s” you make for that JJ big announcer man.. It’s unfortunately a complex waveform, all interlaced to produce a “voice”..
The real problem for an analog approach (as Don suggested earlier) are the “multiband filter roll-offs which would create holes (valleys) in the waveform. The crossovers will need to overlap to avoid this, and flipping phase on one band would effect the tail end of the other..

Seems to me that to truly be phase -o-“Matic” a Neg/Pos look-ahead flip via DSP would be just about right..

*The easiest way to level the polar playing field is sitting on that post and “rotate”. All-pass tickets to ride that BIG voice ..

PS.
I can see you flipping and rotating dozens of switches just to accommodate different times of the day :
“This is my morning voice” .. “damn, where is my cheat sheet?”
Logged

Vintage Radio Pages- http://www.dealamerica.com
K6JEK
Contributing
Member
*
Offline Offline

Posts: 1189


RF in the shack


« Reply #9 on: February 29, 2020, 05:06:44 PM »

...
Modern open architecture audio DSPs are easily configured to do this stuff. Since these units get put in corporate environments (as well as stadiums, churches, auditoriums, theatres, etc.) they show up on eBay every time a boardroom gets remodeled for the next wall street CEO.

Ed, Who uses these things and why?
Logged
KD6VXI
Contributing
Member
*
Offline Offline

Posts: 2648


Making AM GREAT Again!


« Reply #10 on: February 29, 2020, 09:31:03 PM »

Best method I've found is this.

All pass filter, 4 to 8 poles.

(multiband) compress

Limit

Run through audio processor I built and designed that allows for any positive peak you want.  Turn to 0 for normal peaks.

Modulate, set audio level for about 95 pct negative.

Turn pot to run to 125 pct positive.

Done.

--Shane
KD6VXI
Logged
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #11 on: February 29, 2020, 11:58:01 PM »

The issue is not about having big positive peaks. The problem is having big negative peaks — you have to lower the gain so as not to over-modulate. It you flip a negative peak to positive, you can raise the overall gain without over-modulation. Making the audio symmetrical with an APF does solve that problem, but flipping the negative peak to positive solves it too.

A good, simple explanation.

Well, all in all I'm convinced that a DSP solution is the way to go here. The software solution is probably available now. Maybe someone will come up with an inexpensive executable program that will allow polarity band tuning for us to experiemnt with.

Thanks for the info and directing the proper path, everyone.

T
 

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
K8DI
Contributing
Member
*
Offline Offline

Posts: 423


« Reply #12 on: March 01, 2020, 10:48:25 AM »

Ed, Who uses these things and why?

Because they are fully configurable, and controllable, they’re very versatile.

Think of this: hotel with three or four ballroom/meeting rooms and connecting lobby.  Moveable walls (“airwalls”) between them.  So room combining, mic in one room heard in one or more rooms.  Background music per room.  Ambient noise level compensation. Music with ducking during paging.  Phone paging. Audio from the video/PowerPoint per room or combined.  All the audio preamps and level controls too. As well as eq and delay to line up audio and video. Add in auto feedback eliminators. All in one box, all automated by a control system via a touch panel on the wall.  Add to that remote administration, timer controls. That remote admin is network based— I fix/change stuff for customers in New York City from my office in Michigan.

That’s just a hotel.

Some examples:

Chicago’s Navy Pier district.  ALL audio is controlled through a single BSS processor. Not my install, but it’s kinda cool..

The Columbus Zoo and Aquarium, including the Safari Golf course, the Zoombezi Bay water park, and the Africa Event Center:  ALL audio, including zoned phone and desk mic paging, 32 separate audio tracks, and over 100 zones of audio, all controlled and routed through two BSS processors. The audio is sent over the 580 acre campus using Dante, over the existing enterprise network.  Dante digital audio is directly supported by some BSS processors — the audio tracks come in and leave as digital, all the way to the Dante digital input amplifiers.   The BSS system we installed replaced six large separate systems and tied dozens of smaller single zone systems together, administered from a single point.

Corporate boardroom, identity confidential: ceiling mounted, hidden microphones.  VoIP phone included in the DSP. Teleconferencing, local voice lift, acoustic echo cancellation, etc. All invisible. There’s a touchpad in a pocket in the middle of the huge granite table. Dial a number, and talk. Plug a laptop in, and the video system activates, and the audio from the laptop is heard.  No controls. Simple.  This one used a Biamp DSP, they’re better with VoIP.  Again, I can change/fix/maintain over the internet. 

Ed

Logged

Ed, K8DI, warming the air with RF, and working on lighting the shack with thoriated tungsten and mercury vapor...
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8308



WWW
« Reply #13 on: March 07, 2020, 12:01:39 AM »

Good to have choices to get to the same thing. That DSP looks good for those large applications though I am amused at so much complexity and cost just to do this one simple ham radio thing.
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
K8DI
Contributing
Member
*
Offline Offline

Posts: 423


« Reply #14 on: March 07, 2020, 09:01:13 AM »

Good to have choices to get to the same thing. That DSP looks good for those large applications though I am amused at so much complexity and cost just to do this one simple ham radio thing.
When you work with them all the time, so you have to have half a dozen various brands of them in a rack for program testing and development, that happens to be five feet from your radio bench, it becomes more reasonable... ;-)

I’ve just started playing around with some home stereo grade maker-styled dsp products from Parts Express, though.  Much less intuitive user interface, fewer features but workable, and the dsp, analog interface, and usb programming adapter, shipped, was under $70. I’ll be sharing my progress on that in a couple weeks. That price point is a game changer. 

Ed
Logged

Ed, K8DI, warming the air with RF, and working on lighting the shack with thoriated tungsten and mercury vapor...
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8308



WWW
« Reply #15 on: March 07, 2020, 08:15:36 PM »

Good point, if you are already used to them. $70 is a much more reasonable cost for home use. Hopefully it can be more or less set and forget unless other changes are made.
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.066 seconds with 18 queries.