levels looked OK but recording sounded distorted

[Richard Gooderick]

If anyone has an idea of what was going on in the sample below I would be grateful to hear your views.

The mic used is a DPA 4061 lavaliere feeding directly into the camera via an xlr cable.

I had not used the camera before (Sony F3) but the levels looked fine on the onscreen display.

However it did not sound good to me so I changed the mic. It sounds 'gritty', like it is distorting.

I had the mic checked by my local audio specialist and they can find nothing wrong.

password is:
sound

This is a password protected video on Vimeo

Am I imagining this! Any thoughts, anyone?

[Steve House]

Nope, I don't think you're imagining it. I too can hear a number of crackles in there. How were you connecting/powering the mic? You said via XLR cable but was the mic setup with a phantom power supply or what?

edit.....................

I concur with Andrew that the mic seems to be too high in the throat, it looks to be touching the Adam's Apple. If so, vibration from the throat is going to be transmitted directly into the mic capsule at a very hgh level.

[Andrew Smith]

The good news is that the distortion (or over resonating) us only happening at certain frequencies. :-P

I think there are a couple of things that are happening here ...

1. The position of the lav mic is too close to the mouth of the speaker. Hence the distortion when he emphasises his pronunciation of a word. Try one button further down the shirt.

2. You are possibly getting extra reflections from the room on certain frequencies that in turn are adding to the levels (of those frequencies) entering in to the lav mic. Hence the distortion on those types of sounds.

Please bear in mind that I'm not a dedicated audio expert. I may even be 'full of it'. But thoses are my immediate guesses.

BTW, a longer sample would be handy. I had to play it a few times in order to get a bit of a grip on what was happening.

Andrew

[Richard Gooderick]

Thank you Steve and Andrew

The xlr cable came out of the camera into an xlr socket on the waistband of the subject. It is then plugged into a DPA xlr downconverter and becomes a thin wire that runs to the mic. The mic is hard wired at one end of the thin wire. The wire screws into the xlr downconverter at the other end.

Phantom power was being provided by the camera.

I am enclosing two photos of the location. It is all hard surfaces.

You can also see the Sanken CS3e shotgun mic that I used instead. Which sounded fine.

I am uploading a longer clip which will replace the first one. It should be ready in about 4 hours.

[Steve House]

I don't read anything into your desription that would account for a problem. How hot were you recording? I keep coming back to Andrew's idea of mic position and picking up vibration from the speaker's throat causing very brief clipping.

[Richard Gooderick]

Thanks Steve

I wasn't recording hot at all, I would say. The levels looked very conservative on the OSD. And I used the same levels for the Sanken, which sounded great.

So it appears that the proximity of the mic to the speaker's mouth and larynx and/or sound reflections from the hard surfaces are the most likely explanation?

[Andrew Smith]

I really do think it will be the lav mic being placed too close to the speakers mouth ... which is an easy fix.

Andrew

[Richard Gooderick]

Good. That's an easy lesson learned.

I knew that placing the mic too close to the mouth would change the tone but I never knew it would have this effect.

[Warren Kawamoto]

I think your microphone input level into the camera is too hot. I'm not talking about the input level wheel, I'm talking about trim. Does the F3 have an audio trim in it's menu? Try lowering the trim. I disagree with previous posts that putting the mic up too close to the throat would distort it. I've had ministers put it there and shout as loud as they can without any distortion. On an EX-1, if your trim is correctly set to pad the input, it will not distort. Changing microphones requires some tweaking with the trim because some are a little hotter than others. Try experimenting with it and see for yourself.

[Rick Reineke]

Either the mic itself is distorting, (which I doubt, since I have used DPA lavs before in much louder circumstances and never had any issues) or the camera's input stage (pre-amp) is being over-driven. ( and yes, adjusting the the record volume would not make any difference) I'm not familiar with the F3 so I don't know if it has the software gain/trim settings like the E1/E3. Too late now if any case.

[Richard Gooderick]

Thank you Warren and Rick
The F3 does have trim controls within the menu. I just checked the manual. I enclose the relevant parts below in case anyone reading this finds it useful.
I was renting the camera and was not very familiar with it. It was a last minute decision to use the F3. The rental shop did say that they would set it up for me and that I wouldn't have to worry but I should have taken that with a pinch of salt.
Can I ask for some advice, just to get this clear for the next time I use this camera?
Would you set the input level wheels at their halfway setting and the use the trim controls in the menu to get the appropriate level? Is that how it is done?
I don't quite understand that the manual seems to be saying to view the meter levels (on the on-screen display) to monitor the audio when using the trim controls to fine tune to suit a microphone when you first set the camera up. But the meter levels were looking fine on the camera when I was recording. I could see that the levels were fine but could hear the audio distorting. That really confused me.

[Matt Portwood]

Essentially, what I think is happening in this instance is that the signal at the input stage of the camera going into the pre-amp was too hot and distorted the signal before it really hit the pre-amp itself. So when you adjusted the levels on the pre-amp you were just adjusting up or down an already distorted signal. When you change the trim levels you are essentially fine tuning the level before it hits the pre-amp stage so you can make sure it is not overloading the pre.

For comparison consider a standard microphone pre-amp. Most pre's have a selector switch on the back of the unit that says -10 dB or +4 dB. That switch tells the pre-amp what level of signal you are sending to the pre-amp. If you set it to -10 dB the pre-amp knows that you are sending a weaker signal and if you set it to +4 dB you are telling the pre-amp that you are sending it a hotter signal. To put it simply different types of equipment of different quality have different ratings on the level of signal they output. If you send a +4 dB signal into a pre-amp that is set to -10 dB the signal is boosted at the input stage because the pre-amp thinks it is receiving a lower level signal than it actually is. Thus it distorts before it hits the pre. If the amp would have been set to +4 dB it would have adjusted the signal as needed at the input stage so as not to overload the pre-amp.

In a nutshell that is what I believe happened in your case. You were probably sending a hotter signal into the camera than it was set to receive. Adjust the trim and you are good!

Wow that was way more than I planned on typing...

[Rick Reineke]

"Would you set the input level wheels at their halfway setting and the use the trim controls in the menu to get the appropriate level? Is that how it is done?"
Yeah, that would work, set the Rec Volume from 50 to 75% , then adj. the internal trim to get the levels in the 'ball park'. I used to do the same thing on audio mixers that had no pre-fader metering option. I set the channel fader to 'Unity', then adj. the (pre-amp) gain/trim pot for adequate levels and anticipated headroom.
Cheers

[Richard Gooderick]

Thank you everyone for your replies which were all very helpful.

I guess I had a bit to learn about trim controls.

[Tom Morrow]

I went ahead and opened up the audio from this video in Audition. I didn't visually see any signs of clipping; usually clipping will cause a "flat top" look to the waveform where all the peaks are the same exact size.

But I suppose that the re-encoding involved in putting this video on the web could make such visual analysis difficult, as compression or other artifacts feather the peaks away from flat.

The waveform did make me suspect that compression had been done somewhere along the way, as the peaks for the loud parts weren't that much bigger than the peaks of average-level material.

I'd be curious to hear from the more seasoned pros whether they've seen clipping that doesn't show up visually as a "flat top" on the waveform. I attached the waveform of a part that sounds harsh to me especially at the beginning.

Another possibility is speaking louder than the excursion limits of one's vocal cords.