breakout of compression/downward expansion

  • 1
  • Idea
  • Updated 4 years ago
  • Planned
It's time to lose the one-size-fits all "compander" and introduce a real compressor/limiter and downward expansion/gating. I would love to sell a rack of audio gear if only Flex can do it right in the digital domain where it should be :-) and all in the right codec order. Make the balanced line-in capable of at least +4dbu or AES/EBU standards for the near future. I wished in one hand and wanted in the other for a digital audio Toslink/S/PDIF input when the 6XXX was announced. Alas.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes

Posted 4 years ago

  • 1
Photo of Scott AC8DE

Scott AC8DE

  • 34 Posts
  • 16 Reply Likes
It is my understanding that there is a plan to break out the expander and compressor, but not until later this year as other more important things are taking priority.

As far as making the XLR take +4 levels, I've been told this won't happen because of the limitations of the chip's input.  AES/EBU standards don't apply, as that is the pro version of S/PDIF, which won't work into an analog input.

In the future, you should be able to take DAX input for your mic audio if you route it through your PC.  Frankly, it should work now, but I've not given it a go.  But you will be introducing some digital latency when you do so.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
That latency is a killer.
Photo of Paul Christensen, W9AC

Paul Christensen, W9AC, Elmer

  • 313 Posts
  • 132 Reply Likes
Before DAX, I was dissapointed that a digital audio interface was not included in the F6K series. But with DAX, it's actually better. We can still take advantage of devices with S/PDIF or any other common digital connectivity standard. Yes, that requires a PC, but we need one anyway :-) Likewise, there are many good software packages that include compression/expansion tools and that too can be imported/exported through DAX. I think Flex did the right thing by implementing DAX, then leave audio processing and interfacing to others since a continuous digital stream is maintained.

Paul, W9AC
Photo of Michael - N5TGL

Michael - N5TGL

  • 307 Posts
  • 101 Reply Likes
According to the roadmap from Oct 2013, the DE is scheduled for Oct 2014.  This is after waterfall, simple mode flex control, SAM (synchronus AM), TX profiles, panafall, quick rec/play, tracking notch filter and CWX.  I agree with their prioitization, as the compressor works pretty well now and I wouldn't want to give any of that up to get a better compressor.

As for that rack-o-gear, I run none.  I use a nice mic (RE320) run straight in to the flex.  TX EQ is not even on.  Virtually every QSO results in unsolicited compliments on my audio, with further questions as to what the heck am I running.  :)

I think the flex as it is now is capable of very good audio quality.  I have a background in audio engineering, so I'm not completely ignorant of what it takes to make things work...but then I also have a "less is more" philosophy.  What works for me may not work for others.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9032 Posts
  • 3431 Reply Likes
Michael - I too have a little experience in audio (engineering), though not as extensive as Scott (a shout out to my peep).  I have gone the minimalist route by removing the outboard audio processing rack I used with the FLEX-5000 and just use a good mic (PR40, SM58, Marshall XML) into a small mic tube preamp, feeding the 6700 using the balanced line in.  Good audio quality is all about generating sufficient gain at the beginning of the audio chain for the best signal to noise radio.  Using a differential signaling input (balanced mic/line in) is a must for RF environments.  I have also been getting unsolicited reports of "great audio" with the 6700.

There is some room for improvement in the entire AF signal path and the DSP engine in the 6000 Signature series radios has some really cool audio processing feature that we plan to expose in the software once the road map is complete.  The 6000s have a lot of potential in this area and we have only seen the tip of the iceberg.
Photo of Charles - K5UA

Charles - K5UA

  • 310 Posts
  • 86 Reply Likes
Tim, after all the 1.1.2 turmoil is over and you have time to breathe again, I wonder if there is a simple way to identify an audio input over-drive condition by an on-screen flashing over-drive "lamp"? I know we "should" be able to determine sufficient drive levels from the mic level indicator, but if there was a flashing red lamp on-screen we would instinctively know we are driving the audio too hard. The level at which the lamp would flash could be set just a few dB before the level at which the clipping occurs. I know that the better professonal mic pre-amps allow gain to be set in several of the stages to insure these stages have enough signal to perform their functions. Just a thought to make setting audio gain a little easier.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
That's why I like the hard limiter in the DYN menu of the DEQ2496 considering the wimpy -10dbv "max" input into the balanced input of the 6700 ( a retrograde from the max in from the 5000, by the way). Kinda sad we have audio dynamic range and TX/RX bandwidth to die for, world class IMD TX performance and get choked feeding audio into a consumer grade input :-(  
It's all good though and Flex will follow thru with the timeline. 
Photo of Paul Christensen, W9AC

Paul Christensen, W9AC, Elmer

  • 313 Posts
  • 132 Reply Likes


However, the Flex 5K did not include a 20 dB preamp at the XLR input and it's easier to optimize for one level rather than serve as both a mic and line level input as in the case of the Flex 6K. 

If changing output level on pro gear is inconvenient, one solution is to use a 4:1 transformer (e.g., Jensen JT-10KB-D) ahead of the XLR input when being driven from a +4 dBu source.   A secondary benefit of a high-quality 4:1 transformer is that it will offer superb CMRR across a broad RF spectrum range -- something lacking in even the best 3-stage op-amp based instrumentation input.  This can be important in difficult RFI abatement cases.

Paul, W9AC 

Photo of Michael - N5TGL

Michael - N5TGL

  • 307 Posts
  • 101 Reply Likes
I agree about having sufficient gain at the beginning.  With the RE320 I need to run mic gain at 95, but it still sounds pretty darn quiet in the headphones...and by the time it makes it to the other side of the world, there's enough HF noise in there that it doesn't really matter.

I admit I have given some thought to a mic preamp.  I prefer not to have to run things wide open if I can avoid it.  Mebbe I need to build one as a project.

Definitely looking forward to other audio processing in the future!
Photo of Scott AC8DE

Scott AC8DE

  • 34 Posts
  • 16 Reply Likes
Thanks for the shout out Tim.  I'm blushing.  I used to be a FOH engineer for a touring company, then decided after 10 years that I wanted to do something that made real money.  Long story for another time.

As far as the compressor/expander working well as is... I'll just say that it does enough to even the peaks out, but its got a long way to go before it acts like a good compressor or expander.  The boys at Flex got it working well enough to be able to have something functional, and they did a great job at that particular juncture for the time they had.  But if you really know how a compressor and expander is supposed to function and feel, you start adjusting it and wonder "What's up?".  Joe obviously picked up on this.

There is absolutely no doubt you can get great audio from a nice mic directly into the 6000 series with no outboard gear as is.  But remember, the quality of the audio is judged by the guys on the other end, who usually are limited to a 100-2700 Hz receive bandwidth with crappy speakers or a cheap set of headphones, so any good mic (RE320 is a great one - Chinese version of the RE20) will get you 95% of the way there at that bandwidth.  For the best audio with a mic straight in, you want to use a mic that has a slightly higher impedance than the typical 100-200 ohm output.  Something above 400 ohm will give you a lot more output allowing a better S/N ratio as a general rule, but it is not always the case.  It is mic dependent.  But you DO NOT want a true high impedance mic.

I practice ESSB a lot and often run at 7.5 KHz wide.  When other guys are listening that wide that have good audio monitoring equipment, they will hear issues without outboard processing.  But that is a very specialized segment of the hobby, so it's not really fair to even go into that here.  But they all agree that the front end of the Flex-6000 series sounds flat from 20-10KHz, which is a great thing.

As Tim mentioned, there are a few things that outboard processing does for you on the input of the Flex-6000 series... or any radio.  A good argument can be made to continue to use a minimal amount of outboard equipment.  I have to start with the fact that the 6000 series input is limited to -10 dbV consumer levels (a bit higher than -10 dbV actually).  So it can comfortably take mic level to a modest level of preamplification.  While a good mic works great into the input of the 6000 directly, a minimalist outboard equipment approach is better hands down; on ANY rig I might add.  The first thing is to amplify the mic; pushes S/N up a lot.  Then turn on the outboard compressor (careful on adjustment and don't overdo it) and you squash the input audio modulation envelope in such a way that you can stop loud talking peaks from clipping the input of the radio and still sound loud at the other, end even when talking soft. This evens out output RF power like the onboard compressor, BUT keeps you from clipping the input side of the rig and gives a much better S/N ratio, which gets noticeable when the bands have little noise.

The only caveat I give on external audio processing is RF ingression.  If you do not have a good handle on your RF in the shack (or even if you do) adding those extra wires can bite you in the rear and cause you to add isolation and additional grounding.

So bottom line:
- Joe, have patience my friend.  We'll get there with the onboard processing.
- Running just one little cheap preamp/compressor box can make a big difference in your input AF S/N ratio on ANY rig, not just the Flex-6000 series.

73,
Scott AC8DE
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
I learned years back with the 5000A to only use my 528E and DEQ2496 and forego the internal EQ and processing in SSDR and PSDR!
Photo of Charles - K5UA

Charles - K5UA

  • 310 Posts
  • 86 Reply Likes
Hi Scott,

Do you know anyone using an Orban unit driving a 6000 series radio? Have never heard anything as rich and powerful as the HF Orban used by shortwave broadcasters. Would love to hear one on a 6500 or 6700.
Photo of Scott AC8DE

Scott AC8DE

  • 34 Posts
  • 16 Reply Likes
No, I can't say that I do.  Some of the broadcast stuff gets into some pretty big bucks.  And unless you are doing AM or ESSB, it won't do much for you.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
An Aphex project channel box would be peachy keen if only the 6000 could take advantage of the digi outputs as well as a Vorbis M1 box. Straightjacket audio guys need not apply.
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 3872 Posts
  • 1177 Reply Likes

I also am constantly getting "fantastic" audio reports with a Heil PR22 into the BAL-TRS, after tweaking the EQ.  ( I can't understand why anyone would hook up a new mike and not spend 15 minutes with another receiver, or at least looking at their own TX Spectrum on the panadapter while making a few adjustments.)

But just for fun I ran my mike through a USB mixer, through "Stereo Tool," and back into the rig via DAX.  It didn't get me a whole lot extra, but was fun to play with.  I haven't really fine tuned it all that well, yet. The exception was that the ESSB guys on the low end of 80 said that I was getting some great richness using Stereo Tool when stretched out to 6K wide.  (I did go out to about 8K once but felt too guilty!)  The biggest problem I had was about 1/4 second latency on transmit.  I probably could have achieved the same thing with judicious adjustment of the inboard TX EQ on the 6500.  But I didn't want to mess up my other settings.  (TX profiles will make this more easy to play with.)

As far as "normal" bandwidth operating, I am looking forward to definable TX profiles that store all of the bandwidth, EQ, gain & compression parameters.  I tend to "gadfly" about between rag chews and DX hunting and am always changing my TX Bandwidth.  I would change EQ settings and compression along with it but it takes too long and too many mouse/keystrokes.  Profiles, and the addition of some more aggressive compression/processing for added DX/weak signal punch (just a bit more than we can get now) things will be great. 

But I agree with the others...It is really good now, and I am content to wait until the other features are delivered before adding extra processing (Unless it is really easy and won't sidetrack the developers!)

For the future?  Here are some dreams...

TX Profiles (already on the roadmap)
Definable Parametric EQ?  
Or the ability to draw your own EQ pattern as Stereo Tool does? 
Multiband compression with definable compression level and post-compression output level for each band?  This would be killer!
Noise Gate/Downward Expansion.
Ability to define the PTT/Footswitch to DAX audio input channel if desired.

Just a little dreaming... And I still don't know enough of the capabilities Tim is talking about to even know what I want!

Keep up the good work!

Ken - NM9P

Photo of Scott AC8DE

Scott AC8DE

  • 34 Posts
  • 16 Reply Likes
Ken,

I think I remember when you visited us at the bottom of 80M.  Fun stuff, huh?  Sorry you felt guilty.

Nice wish list.  I'll comment on a couple.

I for one would LOVE to see a parametric instead of a graphic EQ with ability to drag points and change filter Q, as it is much more usable and WAY more effective.  (I do that outboard currently.) However, it would be a stretch for a lot of hams, as it is way more complex to adjust.  But why not?  Very useful and anyone can learn to use it in time.

Something as esoteric as multiband compression would be a lot less useful in the ham radio world for a number of reasons, at least on normal SSB.  First, most people really don't really have the subtleties down on proper adjustment of a single band compressor or understand how that affects transmitted frequencies.  Now add in a multiband version, and I could see heads exploding and some really screwed up audio.  The only people that would get any real benefit would be AM and ESSB guys (especially AM) because without being at least 5KHz wide on TX, it's affect would be almost negligible.

Downward expansion (just expander for short, which is what most know it as) is already there coupled to the compressor, although its not really usable yet as it has a really low threshold which is obviously not yet adjustable.

Scott AC8DE
Photo of Charles - K5UA

Charles - K5UA

  • 310 Posts
  • 86 Reply Likes
Scott, took your suggestion of using a bit of compression from my 528E voice processor. You can literally see the amp output start to decease at a compression ratio of about 1.7 to 1. I wish the 528E also had a limiter, but the mild compression appears to even out the input signal enough to prevent any drastic overdrive peaks. Thanks for reminding us. Somewhere (not here) a good tutorial on how to set up an outboard compressor-limiter would be beneficial to a lot of us.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
The current compression ratio is fixed 5:1 in SSDR. Too aggressive IMHO, especially if one tries to use vox with an amp's noisy fan. I use the 528E's downward expander wired in the side chain EQ with a -15db cut at .3 octave centered on the ambient room/amp blower noise frequency (~145hz) to maximize the expanders usefulness.
Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
Ken, my vote is for EQ before the compression happens in SSDR...easy to do with a 528E or in combo with a 2496. How Flex arranges this is up to us to voice our opinions.
Photo of Scott AC8DE

Scott AC8DE

  • 34 Posts
  • 16 Reply Likes
Ken,

EQ/Compressor VS Compressor/EQ is an interesting topic.   I understand the argument both ways, but it's not a simple thing to decide.  Or is it?

For guys who TRULY understand compressor/EQ relationship and have years of professional audio knob tweaking under their belt, whether you EQ before or after the compressor isn't that important, as long as you know the signal path order.  EQ before a compressor has some major advantages when you are talking about signals that are heavy at certain frequencies, like a kick drum.  Such signal are very consistent, so using EQ before compressor makes sense and solves many leveling issues; EQ out the heavy beater pop of the kick that's pegging the meters and compress to level requirements.  Easy Peasy.  But with the AVERAGE human voice, EQ before compressor is a lot trickier and I'll guess more than the average ham wants to grasp.  When an EQ is before a compressor, every time you boost or cut the EQ more than a db or two at a specific frequency, it will change the overall level of the signal relative to the compressor threshold setting you've already chosen.  Therefore it screws with the adjusted compression sound, unless you keep going back and forth adjusting ratio controls and resetting the compressor threshold!  If you are an old salt to this stuff, it's no big deal but does take a bit more care.  But if someone doesn't understand the dynamics going on (which is the average ham), they are going to get frustrated really fast and think something is wrong with the EQ or compressor.  For that reason alone, I would not suggest EQ before Compressor order for most hams due to that back and forth adjustment problem.  There are plenty of arguments the other way, I know, but the reality is we have to set this up for the average ham to be able to handle easily, as they are not sound engineers.

That said, there really wouldn't be any reason that a check box couldn't be put somewhere to change the order of the compressor/EQ chain if someone so desired.  It's software after all.

Scott AC8DE




Photo of K4EAR

K4EAR

  • 138 Posts
  • 17 Reply Likes
I'll throw my hat in for an option... click pre/post.
NU9N espouses the following:

If the EQ were placed AFTER the compressor, high frequency energy from the EQ would be able to distort the audio since the compressor would not see it, therefore not being able to react to it.

With the EQ BEFORE the compressor, we can cut midrange frequencies from influencing the compression action of the high frequencies before it gets to the compressor. And, if the compressor has a built in low frequency filter preventing the lows from influencing the compressor, we have an elegant solution where we can adjust the compressor so that ONLY high frequencies will be limited aggressively, while the low and midrange frequencies will be compressed conservatively.
Different strokes for different folks Take a look at the "generic" EQ graph in SSDR  and come to your own decisions.

I used a TC-electronic Triple-C multi-band compressor before it went south on me on the 1000/5000 in the early days. Not voodoo stuff either.