Question about audio chain order sequence

[K1JJ]

I decided to put my CRL PMC-300A peak limiter back online.  It needs some adjustments and I'm going over the manual now. But initial tests are impressive and it definitely limits a sine wave's peaks and gives a denser audio on the air. My audio level is all over the place and needs a limiter badly.

Question:  I'm trying to decide where to put the new SCAF filter.  I know a limiter should go at the end of the chain, but I am thinking that the low pass filter at say, 6.5 KHz could help to filter out any harmonics generated by the limiter. When limiting hard, I see the test sinewave's peaks get a crushed rounded look which indicates harmonics to me. Why not put the LP filter at the end of the chain after the PMC -300A?


The present order is now:  Shure broadcast mic >  528E  processor >  31 band EQ side-chained into 528E  >  528E audio compression, 3-6DB  >   PMC-300A peak limiter >  SCAF  low pass audio filter > GFZ MOSFET audio driver > modulators.


T

[N1BCG]

Peak limiters or clippers always, without exception, are the last stage before the modulator. Your CRL is designed with a low overshoot LPF to address clipping artifacts so all frequency shaping should precede that.

The first version of the MAX processor was laid out just as you suggested. Tests revealed that while the 295 Butterworth filters have a smooth response and decent rolloff characteristics, they also wiped out the positive and negative peak control of the clipper stage, requiring a reduction in modulation to compensate.

To address this, we switched to the 296 Bessel filters which exhibit low overshoot. The problem then was the loss of the tight bandwidth filtering due to the more gradual curves of the Bessel design.

What we settled on was to place the 295 filters after the compressor/limiter and before the clipper. This way any overshoot would be addressed and maximum modulation preserved. This also improved S/N and gave the filter stage a very controlled audio level.

[K1JJ]

Thanks, Clark.  

Then I'll place the PMC-300A at the end of the chain, just before the GFZ MOSFET audio driver board. This will give me a similar chain order as you described was best for the Max experiments.

Yes, the overshoot is definately there in the SCAF filter and is the one thing I have not been able to reduce, due to the limited access to the chip internal parts. Though, all analog filters (non-DSP look-ahead)  probably have some overshoot or other compromise.

The PMC-300A is not an easy box to get adjusted and optimized right. I'm working thru it now.

** I see that the PMC-300A comes stock with 10 KHz LP filtering parts installed.  Should I leave it as 10 KHz or is it worth the effort and future limitations if I bring it down to the next level of 8KHz to help the filtering further?


T

[K1JJ]

Clark,

You were correct about placing the PMC-300A last in the chain.  I moved the MAXIM 295 SCAF filter position to just after the EQ and 528E compressor.... before the PMC-300A peak clipper.  

The spill-over fuzz on the waterfall past the skirts is gone. This is about the cleanest config so far.  I'm really pleased with the results.  

Also, I was having trouble with the PMC-300A interacting with the filter, causing all highs and no lows. Strange.  By putting the filter into the 528E sidechain, it cured the problem.

This is the first time I have used peak limiting in a long time and the small amount of extra blower noise is worth being able to scream or whisper into the mic without staring at the modulation level all the time.  The audio density has increased while the positive peaks are peaking at 160%. I'm close to reaching my goals for this overall project.

T

[N1BCG]

Great to hear, Tom! It can be frustrating working toward the ideal sound, but once there, it’s all worth it.

Excessive background noise (blowers) can be reduced with a slight reduction in drive to your 528 and a lower “Process” setting on your CRL.

Most b’cast processors employ a gating function where the AGC gain freezes when the input level drops below a set level. Gain riding then only occurs when there’s enough S/N ratio. Rick and I added this feature, the “Gain Gate”, to the latest MAX processor to address the noisy shack issue. Aggressive dynamic range processing when there’s enough audio to work with and unity gain below.

Looking forward to hearing more FB audio!

[K1JJ]

Great to hear, Tom! It can be frustrating working toward the ideal sound, but once there, it’s all worth it.

Excessive background noise (blowers) can be reduced with a slight reduction in drive to your 528 and a lower “Process” setting on your CRL.

Most b’cast processors employ a gating function where the AGC gain freezes when the input level drops below a set level. Gain riding then only occurs when there’s enough S/N ratio. Rick and I added this feature, the “Gain Gate”, to the latest MAX processor to address the noisy shack issue. Aggressive dynamic range processing when there’s enough audio to work with and unity gain below.

Looking forward to hearing more FB audio!

About noise gates:  I'm actually using an Aphex noise gate in the sidechain.  My observations:  It is great on ssb when there is blower noise, but not TOO MUCH blower noise. If the blower noise is too loud, then the gate is ineffective and the white noise just swirls around the speech syllables in the background.  When the blower noise is low, the noise gate is very valuable to drop it 10-15 DB which can appear quiet on the air. The release should be slow and smooth, not abrupt.


My own minimum S/N blower noise measurement standard is -30 DB.  You can tune in your own ssb signal on a receiver and see the signal at S9 +30 over. The blower noise should drop to S9. A better goal is -40DB on ssb.  At -40 DB blower noise, it is quite acceptable. Same for AM.

On AM, I hate noise gates - because the carrier quiets the background and as Tron says, "a noise gate is unnerving to listen to." The background ambience coming and going is distracting. Thus a gate that has the ability to leave some background noise in to anchor the brain to is better. A little gate goes a long way on AM.

All in all, I never use the gate on AM, but do all the time on ssb.  It will be a nice option to use in Max when needed.


Yes, I already have the minimum 528E input level to run things and the CRL "Process" control is in the lowest setting. I got by with no clipping for years, so I can live with just a touch now. I have this thing about getting rid of blower noise as much as possible. It is almost impossible without putting the big rig in a different room.  So many really FB sounding rigs are ruined by bearing whine or blower air white noise.  

One of the best thangs I've done for shack noise was three years ago.  I took all my rig blowers to a motor repair shop. (five) They replaced all the bearings. Huge difference.  Cost only about $100 total. Those used blowers we all buy at flea markets are often at least 5 DB acoustically louder than they should be.


T

[N1BCG]

That’s a “noise gate” which is a common feature on music store processors. Some studio mic processors use elaborate noise gates to effectively turn off microphones that are not being spoken in to. This is handy for round table interviews to reduce reflected sound from being picked up by idle mics.

I’m referring to a “gain gate” function that holds the gain at unity below a threshold. Above this point is the full processing, but below is a range of unity 1:1 pass through. No compression, no gating, no expansion, and no increase in background noise.

[K1JJ]

Very interesting variation.  Gain gate vs: noise gate...

I can visualize doing manually what you are doing automatically.  Riding the knob to decrease gain between words.  I'll have to see that in action one day.

If it can reduce shack noise transparently by even 5-8 DB, it would be worthwhile to me.

T

[Opcom]

On AM, I hate noise gates - because the carrier quiets the background and as Tron says, "a noise gate is unnerving to listen to." The background ambience coming and going is distracting. Thus a gate that has the ability to leave some background noise in to anchor the brain to is better. A little gate goes a long way on AM.
T

I had no idea. Apparently my noise gate is subtly torturing the listeners.
A case of "looks great here on the scope, annoying there in the speaker".

That said, the gate could probably be modded to allow something through. It's a simple one and is used for transmitter blower and HVAC noise.

But how much to allow? That is, by how many dB should the background noise be attenuated below the 100% modulation value when the gate is shut, to give the anchoring effect? 10dB?
Keeping in mind the compressor pushes the noise way down when any speech is present.

[N1BCG]

If you have a noisy shack then that’s where the problem should be addressed first. Some ops are able to reduce background noise by moving their mic and noise source apart or use cardioid pattern mics to reduce pickup from other directions.

An easy solution is to reduce compression. If audio processing brings background noise to full modulation between words then it’s set too high for that environment. Another option is to eliminate any AGC action and utilize just limiting for peak modulation control. This gives an ultra-clean sound with minimal noise pickup.

Talking closer to the mic is yet another way to increase the S/N ratio of what’s being transmitted. Combine several of these and you’re sure to receive compliments.

[K1JJ]

On AM, I hate noise gates - because the carrier quiets the background and as Tron says, "a noise gate is unnerving to listen to." The background ambience coming and going is distracting. Thus a gate that has the ability to leave some background noise in to anchor the brain to is better. A little gate goes a long way on AM.
T

But how much to allow? That is, by how many dB should the background noise be attenuated below the 100% modulation value when the gate is shut, to give the anchoring effect? 10dB?

Hi Pat,

Without a noise gate, my personal standard is to see blower noise on an ssb signal drop at LEAST -30 DB between voice word pauses.  (using PTT, not VOX - just watch your S-meter)  If you can get a  -40DB drop or more, then you've arrived and no one will really notice the noise.

The same applies to AM, though the S/N is harder to measure with the carrier on, requiring a DB calibrated audio meter. (or an o'scope measurement converted to DB)

Sometimes when we optimize an EQ for our voice and preferences, we get stuck amplifying a noise area where the blower whine/noise exists. The only thing to do is reduce the noise or drop the EQ in that area.  

I spent a lot of time thinking about audio absorbing material in the shack, moving rigs behind a glass enclosure, etc.  But I always come back to the rig standing 5' away and fighting the -40DB minimum standard as best I can.

As far as anchoring, IE, leaving some noise in there, just listen in the monitor and make a judgement call. It's just like camouflage in the woods. It's an art. You will find if you go too low in residual noise, the gating effect takes over and is distracting again. But leave a touch of noise and most of us will ignore the gating effect because it is not abrupt.  It's just like listening to your refrigerator. It can make a small racket for hours, but when it stops abruptly, we notice.

BTW, I use an Aphex Model 612 Expander/Gate. It has variable release times and the ability to leave residual noise at any level. Some gates do not have this feature. You can get them for about $100 used these days.

https://www.worldmusicsupply.com/aphex-expander-gate-model-612-stereo-expander-noise-gate

And, the new Max processor by Clark and Rick will have a "gain gate" so there is a lot to look forward to and consider.


T