Choppy music audio during Webinar

[marktx]

Hello All !

I need to play soft background music while speaking through a microphone.  During LIVE stream webinar tests, microphone audio is clear and stable.  However, low volume background instrumental music is choppy, regarless of whether mic is being used or not.

I have tried numerous options for feeding music into GoToWebinar.  Mixer to computer, Mixer to SparkoCam, Cell phone playing into microphone, etc.  I have tried during business hours and in the evenings.  Same issue:  choppy, unstable music.

I have a live webinar in 2 days and am really at wits end trying to find a solution.  Will be greatly appreciative for any guidance.  Even if it's not possible, that would be good to know.  Thanks in advance.

[Chris Droessler]

My guess as an audio professional is that there is a noise gate and compressor working on the audio somewhere.  Not sure if it is in the hardware (computer) or in the GoTo software, or both.

The noise gate will "turn off" or mute an audio source when the sound level (volume) gets below a certain level.  Noise gates are often used on drums where you want them to "turn on" automatically when there is a loud sound and then turn off when the drum sound stops to eliminate background noise when that drum is not playing.  This noise gate is a good thing if someone is talking and you want the noise gate to automatically mute the mic during the pauses between words and anytime you are not talking. 

This can have an adverse effect on soft music, because the noise gate will mute the music anytime the volume is below a certain level.  That makes it choppy.

To test this theory, try playing the music loud and see if there is a difference between playing soft and loud.

You could try to defeat the noise gate by adding a low frequency carrier wave to the music, like 40 Hz.  Something that is low enough frequency that is will still be heard by the noise gate but not transmitted by GoTo.  You mentioned an audio mixer.  You could play your music through one channel and add a tone generator to another channel. Adjust the level of the tone and frequency until you get rid of the chop, but still have a pleasant music background.

[marktx]

Hi Chris,

Thank you very much for this reply.  I experimented yesterday with our mixer.  Each channel has High, Mid and Low knobs, as well as a Low Cut button.  I ran 12 tests, each time turning one knob, two knobs and three knobs max or minimum, then repeated with the Low Cut button depressed.  For the test, the volume was turned up above the voice level to rule out source volume being too low.  Unfortuntely, this did not yield any improvement.    

To rule out the laptop as an issue, I repeated the test on a desktop with huge RAM and processing capacity, but the result was the same:  choppy, inconsistent instrumental music.

To rule out network traffic as an issue, I repeated the test early in the morning, late at night and during the day.  I am connected to the internet with an ethernet cable. 

I wasn't clear on how to add a tone generator to another channel.  Could you give some guidance on this?

thanks, Mark

[Chris Droessler]

So was the audio still choppy at the louder level?

[marktx]

Yes it was still choppy at the louder level.

[Chris Droessler]

I just did a GoToWebinar test with music myself, and I got the choppy music as well. It is a different chop based on my assumptions of your comments, but I think I know what is happening.

WARNING: This will be a Physics class in audio processing. All of this may be on the final exam!

It sounds like something is taking all sound below a certain volume (loudness) and reducing the volume on that, and leaving the audio above that level alone.

It sounds like GoTo uses an audio processor we call the "downward expander." My earlier guess of a noise gate is close. A noise gate would eliminate all sounds below a certain volume, which is not what is happening here. A noise gate is a downward expander taken to the extreme. In this case the sound under a certain volume is still there, just at a reduced volume.

My guess is that GoTo added this downward expander audio processor to reduce the volume of the background sounds, which would make a normal presenter voice clearer. That is a good intention.

I've always suspected that GoTo added a compressor to process the audio, but now I think they must use a downward expander first, then add a compressor next. Probably a de-esser after that.

I always add a compressor to the audio track in my GoTo recordings before I post them. You know how TV commercials are always louder than the TV shows. The commercials have been compressed to the point the softest sounds are just as loud as the loudest sounds. In a TV program you have the whole range of loud to soft sounds. The commercials are not really louder, they just bring all the soft and loud sounds up to the same level as the loud sounds that are scattered throughout the TV show. A little extra compression on the audio track can help to reduce the difference in volume levels of loud and softer speakers.

For normal people talking in GoTo, the default settings are fine, but for soft-spoken presenters, and for playing music, I wish there was a way to defeat the default settings. 99 percent of the GoTo users would not have a clue as to setting compressors and expanders, but might find solace in having a switch that turns it on and off. The switch could be labeled "music," which means if you want to play music, enable this switch. Otherwise leave it off to get better sounding voice with reduced background sounds.

This is the same "pulsing" sound I get from a soft-spoken presenter. Sometimes you just can't get your presenter loud enough, even with the microphone inside their mouth. This new "music" playback switch could help keep the soft-spoken person from fading in and out. Then you just hope they are in a quiet room.

I know GlennD at GoTo has some background in sound, maybe he can shed some light on this and translate to the programmers.

If you want further reading on these audio terms, may I suggest:

https://www.izotope.com/en/learn/audio-dynamics-101-compressors-limiters-expanders-and-gates.html

[SteveSgt]

Chris,

I agree they are doing some significant processing of the audio (which they expect to be coming from a modest microphone in a noisy environment). I'm sure there is filtration, expansion with some speech-favoring sidechain or even spectral-envelope recogniton, additional compression, etc. At the end, anything that isn't human speech would be irrepairably damaged by the processing intended to make speech more intelligible.

What I would l like to see would be for the app to allow the organizer of a webinar to have a second unprocessed (or lightly compressed but broadband) audio input that is mixed with the speech input after it is processed. Some other platforms do provide such a feature (and no, I won't name names in GoToWebinar's forum).

The lack of this feature has caused my organization to more carefully pre-produce more of our events as videos, but that's not an ideal solution for many reasons.

[Mel-SoTSA]

A mixer with a pre-fader aux out to a  dynamtric TMP 646 phone patch to real  landline gives the best perfomance. 

[SteveSgt]

That is sad, that the best music would be if sent over a dial-up line, which is 8 bit/8 kHz, and limted to 300-3kHz audio bandwidth.

With most people joining on smart phones or computers, they should be able to get spotify/youtube quality.

[Chris Droessler]

In these days of smart phones who has access to a land line anymore?