PNW DMR  -  News   Forum   Repeaters   Talkgroups   Matrix; MMDVM OR VHF WA-W WA-E/ID   MCT   MMDVM   Links   Netwatch   History

Pacific Northwest DMR

Audio VU Meter and Correct Levels

DMR audio levels have always been problematic, even back in the early MotoTRBO days (more info).  This page will address ways to get better audio.  Trying to get everyone on the same level is difficult so we hope that some of the ideas discussed here will ease us all into better audio...both for the talker and the many listeners who are listening to that talker.

There are many ways to generate poor audio, both technical as well as user inattention or indifference:

  • Incorrect factory defaults in the client radio

    • No ability to set TX audio levels

  • Poor implementation of AGC or no AGC

  • Users are unaware of soft or hot audio

    •  or worse, simply do not care (they don't have to listen to their poor audio)

    • Holding the mic incorrectly in relation to the mouth

      • too close, speaking directly into the mic port (air stream into the mic opening)

      • breath puffs into the mic port (vocoders go ballistic)

      • harsh sibilant sounds being louder than the average speech (vocoder confused)

  • Over the air testing with other hams is problematic and subjective

  • We don't have the digital tools to adjust TX audio (as we do with an FM service monitor)

On the plus side, we have the following "tools" which are currently available to aid in our quest for better audio:


When testing your audio (with or without the LED VU meter resource), always use your normal voice, normal position, under normal conditions.  Sounds easy but it is not.  Think about how other people use cell phones, they talk louder when unaware of the distraction they are causing.  Most people will speak softer when they are aware of the noise they are creating or in this case when the OTA level may be offensive.  Most users will again, speak softer when advised of their hot audio but it will creep back up when they forget about it later.  So try to resist that temporary reaction. 

You could have your friends monitor your audio at random times and report back to you later as to where your audio falls when you are not thinking about your audio levels.  Most/many/some users tend to be offended when advised of poor audio, even if they don't indicate it.  So try to be gentle and diplomatic.  But a temporary solution by saying that you will talk softer or further from the mic IS NOT THE ANSWER. 

You must reprogram for your voice and style or you will fall back into that old method of speaking to your radio and continue to violate acceptable levels of audio as well as the clarity of your speech.  This human tendency cannot be emphasized enough!  The radio must be adjusted and likely the user will need to modify the way they position the radio.  This is has been amazingly difficult to get users to do.  They don't need to listen to their own audio but we do!


So how do we get good audios out of our radios that is a pleasure to listen to?  This is a method that I have used since the days of MotoTRBO/Hytera only DMR client radios.  The wide dynamic range in DMR audio makes it more difficult to get always perfect audio to the listeners.  It also makes it more difficult when trying to measure audio out of a DMR radio using common analog tools. 

I have used the SINADDER, Spectrum analyzer/Equalizers, audio compression and AGC gain and reduction when I was video streaming DMR onto a video streaming service.  It is hard to get perfect audio even when using this equipment with up to 6 dB compression and up to 9 dB gain or reduction.  But the key is to just get the levels close and position your radio properly when transmitting.  And this approach can be highly subjective when asking for an over the air, "How do I sound" report.  The person at the other end listening as an unknown listening level from his radio which at best provides a clarity report but not objective level reporting.

Most us of do not have a $42,000 AeroFlex 3920 DMR optioned Radio Test Set in our shacks, so some other method that provides a way for us to quantify our (average) levels at a cost that a typical ham is able to afford is needed.  This approach outlined below costs less than $25 and works well for the end goal of having effective and pleasant audio.  A spare radio dedicated to the project is a huge plus.

Parts needed:

  • Radio that provided earphone level audio output or speaker output massaged to reduce the levels so you have mid-range volume pot to be approximate level to drive into the yellow LED's.

    • Interface connector/cabling from radio to your LED module (some connectors are hard to find)

    • May need a 50-100 ohm resister to enable the external speaker/earphone output

      • for the GD-77, I'm using the speaker/mic earphone jack with 120 ohms across tip and ring to turn on the external speaker.

      • For the CS-580, I used the same 120 ohms via the speaker/mic

  • 30 segment 3 color LED VU module (30 segment green/yello/red for Amazon, ~$40)

  • 20 segment 3 color LED VU meter (32 segment green/red sample from:  Ebay, ~$25) [NOTE:I am having issues with this Ebay 32 segment, FYI at this time only]

  • Once the audio level on the HT's volume control is set, mark it and then leave it alone.

    • You can leave the 580 in the charger 7/24 or use a battery eliminator pack.

    • Put the radio in digital monitor mode, single or dual slot so that you can "see" all traffic.

a

The LED VU bar meter shown on the left provides an objective view of the range of talkgroup audio levels.  The target range is the 6 yellow bars  but 1 red square is acceptable on peaks only.  Here is an image of a 30 segment module

Click on the Image to see a numerical description edition to match a verbal report.

The meter is adjusted so that reasonable audio (no piercing audio, no distortion and breath puffs minimized) is likely when your average audio falls within the yellow range.  This set-up provides a more complete "picture" of all audio conditions.  The actual dB values is less important in this context.  The peak audio level is held for 1 second.

0dB = 3 reds (HOT HOT) // -3dB = 1 red (HOT) // -8 dB = 3 yellows (Perfect) // -12dB = 0-1 Yellow // -15db = 3 greens

The responsive range of this meter is 4-20 though the audio level input is can be lower than 4 and higher than 20. 

The VU meter is calibrated to a baseline of all yellow and green bars plus one red just flashing on the edge of illuminating while pressing Touchtone number 5 on a MotoTRBO XPR-6500 HT keypad as it sends out roughly 1,000hz (dual) tone.  Calibration provides the dB levels vs LED segments as measured by my Sinadder 3 (More Info).

To test with this resource, always use your normal voice, normal position, under normal conditions.  Sounds easy but it is not.  Think about how other people use cell phones, they talk louder when unaware of the distraction they are causing.  Most people will speak softer when they are aware of the noise they are creating or in this case when the OTA level may be offensive.  Most users will again, speak softer when advised of their hot audio but it will creep back up when they forget about it later.  So try to resist that temporary reaction.  You could have your friends monitor your audio at random times and report back to you later as to where your audio falls when you are not thinking about your audio levels.  Most users tend to be offended, even if they don't indicate it.  So try to be gentle and diplomatic.  But a temporary solution by saying that you will talk softer or further from the mic IS NOT THE ANSWER.  You must reprogram for your voice and style or you will fall back into that old method of speaking to your radio and continue to violate acceptable levels of audio as well as the clarity of your speech.  This human tendency cannot be emphasized enough!  The radio must be adjusted and likely the user will need to modify the way they position the radio.  This is has been amazingly difficult to get users to do.  They don't need to listen to their own audio but we do!

This image on the left was what was seen when Livestreamed back in 2013.  The target was to hit the yellow bars without going into the red, paying attention to the average range of the bars.  1 red bar occasionally is OK but no higher.  Anything else should merit your attention or correction.  This is different from a normal VU level in other applications but was done to keep it visually simple, just hit the 4 yellow segments on average but never into the reds.  Greens only are too low but better than hitting the reds.

 The sample at left shows 4 of 5 yellows which is essentially perfect.

We suggest saying something like, "This is K6HOT, VU Audio Test 1, 2, 3, 4, 5".  To "exercise" your audio, you might say "4" several times; long and sustained as it provides a better view of the bars and tends to be the hottest audio in the number string.

Turn your local audio up and listen closely for syllables that confuse the vocoder as well as breath puffs across the mic.  These are points that cause the most grief for the listeners when your overall audio is too hot.  Even when audio is in the yellows, the vocoder may still have trouble processing the sounds so listening closely should help you determine if you should adjust your position relative to the mic on the radio.


We hope that you will become "Audio Active" in that you will try to generate good audio over the PNW DMR network as well as help others to so.  If you hear poor audio, let the user know and over to assist them with your VU meter.  You don't need to be the "Audio Police" but you can make a difference if you make the effort on behalf of all members of our network.

-- Mike, NO7RF, 3-29-2018

 

 

 Revised: 09/17/2018 22:56

 

Contact PNW

Webmaster