Audio Help

[[email protected]]

Please help - I recorded a 1 hour speech and upon playback realized that somehow the +48V on the camera was flipped on (obviously accidentally. and sometime after the camera was set and tested). The actual audio is balanced but it's just "hot" sounding.

How do I fix it? Is there an Edius solution or are we talking something else like Soundbooth (which I have but would require additional help in choosing correct fix setting)?

I know there IS a fix somehow! right?
TY!

[Khoi Pham]

What do you meant by hot? too loud or distorted? you could try eq the midrange down a little and audio restored in Edius and see if it help.

[[email protected]]

I would have to say it's over modulated. the speaker sound almost ok when she speaks softly, but normal speech sounds over modulated - distorted. It's classic 48+V sound mess. I can't believe this happened. I'm really upset.
I'll try that - any other suggestions???

[zvit]

Can you provide a link to download some of the audio to check? (or if it's a short speech - the whole thing)

[dpalomaki]

Hot, as in too loud and clipped?
What was the sound source and how was it conencted to the camcorder?

What do you mean by classic +48 volt sound mess? Feeding phantom power to an unbalanced mic that cannot cope with it causing a positive bias on the audio waveform? Some quality dynamic mics and similar sorurces will ignore it.

Too loud can be fixed, and modest amounts of clipping can sometimes be improved using audio editing tools such as Audition (and maybe Soundbooth, but I've not used it). An approach is to export the audio to a file, fix it in your sound editing program, than add the repaired audio back as a new audio track and mute the original.

[[email protected]]

Yes - here is the link to the audio - NOTE!! the opening titles says ch L = Mic, it's actually Ch R that is the mic. But the rest of the titles are correct.
CLICK HERE

I had the 48+v switched ON when patching into a mixer (provided by the venue)- Connected to camera via XLR into channel R - Ch L was room mic. The audio was not hot from the mixer - it was balanced. On the clips above they are NOT peaking at all. The audio was not peaking on the VU meters on the camera at the event. The switch was somehow bumped or hit before recording. I could play the blame game but ultimately it's my mess.

However, it was exactly as dpalomaki said; Feeding phantom power to an unbalanced mic that cannot cope with it causing a positive bias on the audio waveform. (THANK YOU, that IS exactly it).

I am not an audio editor per se. I have outside audio programs and messing around with them is always an option. I have done that before for short clips. I was hoping to get a better handle on the real fix, and was hoping it was in Edius. THIS production is about 1hr 15 min long.

THANK YOU!!!

[zvit]

Youtube is one of the worst compressors there are. Can you zip and upload the original to the forum? Just the audio - no video.

[digitalgecko]

Was your audio input set to MIC or Line level? If your coming in from the board it should have been Line. If it were set to MIC level, then you would have gotten distortion as well.

[John Hooper]

Yes - here is the link to the audio - NOTE!! the opening titles says ch L = Mic, it's actually Ch R that is the mic. But the rest of the titles are correct.
CLICK HERE

I had the 48+v switched ON when patching into a mixer (provided by the venue)- Connected to camera via XLR into channel R - Ch L was room mic. The audio was not hot from the mixer - it was balanced. On the clips above they are NOT peaking at all. The audio was not peaking on the VU meters on the camera at the event. The switch was somehow bumped or hit before recording. I could play the blame game but ultimately it's my mess.

However, it was exactly as dpalomaki said; Feeding phantom power to an unbalanced mic that cannot cope with it causing a positive bias on the audio waveform. (THANK YOU, that IS exactly it).

I am not an audio editor per se. I have outside audio programs and messing around with them is always an option. I have done that before for short clips. I was hoping to get a better handle on the real fix, and was hoping it was in Edius. THIS production is about 1hr 15 min long.

THANK YOU!!!

Hi Sherri,

I don't think that the source from the mixer is correctable. Was the mixer giving you a mic level output? It sounds like the mixer was set to line level and connected to your mic input.

Anyway, what is done, is done. One solution would be to use the room mic track, increase the level and apply some EQ. It would still sound like a mic set some distance from the source, but at least would be less painfull to listen to.

[[email protected]]

ZVIT - Gladly!

John & Digitalgecko- YUP - it was set MIC 48+V. NOT LINE. It was switched somehow. HOWEVER, I was monitoring the audio out on the VU meters and it looked good. I know, I should have had the headset on too. MY bad there - but the venue was extremely difficult, and in the stress forgot.

I kinda thought it wasn't fixable but I know there are miracles that can be done in Hollywood - and I know some of you are geniuses... so....

????

[iibw]

Sherri, what the guys are saying is that the MIXER OUTPUT was probably line level, not the camera's input. If this is true, then even if you turned the gain down on the camera's mic input, you'd still get distortion, even if there was no peak clipping. The only way you could have avoided this is if you monitored with headphones, you would have heard the distortion right away. The fix here was to have put your camera's input to line.

There are some programs such as Soundforge or Audacity that can reduce clipping, but not entirely eliminate it.

[John Hooper]

There are some programs such as Soundforge or Audacity that can reduce clipping, but not entirely eliminate it.

I did try this for Sherri with her sample clip, using Soundforge, but the distortion is too far gone to be salvaged, in my opinion.

[[email protected]]

THANK you everyone - Esp John trying to fix it - I appreciate all the advice and ideas.

Just for the record, the camera WAS set to line (the switch was set all the way to the left on the camera for line). It was somehow pushed all the way to the right which was 48+Mic. So when we did our sound check (LINE in) it was fine, the switch happened, and when we went to record, it was not caught. MY stupid was not using headsets. Fully take that one on. But the jury is out on what happened to the switch. :(

my ONE SAVING GRACE, (just to finish the story), is that the speaker was supposed to stay in spot A and she roamed around, making the video look more profile shot than full or 3/4 shot. She isn't sure she even likes that - so the audio issue COULD be a non-issue anyways.

IN ANY CASE... thanks everyone!!!

[dpalomaki]

The posted waveform does look like an over driven analog input rather thatn digital clipping, and it is reasonably symmetric. It also is fairly low level (-24 dB or so below max digital level.

Phantom power at +48VDC is generally a fairly high impedance source, and I would expect most most balanced outputs on decent mixer to swamp it.

What camcorder (or audio recorder), mixer, and any other devices in the audio chain were being used?

[zvit]

Worked on it (Audition and Edius) for some time but like others said... probably can't get it to perfect.

The 2 options that I would do if it were my client would be a little high pass with high-pass filter or EQ (raise right side and lower left side) and then export and reimport and then using EQ, add some EQ bass (raising some left side) and then lowering the over-all volume of the track and raising the on-cam mic volume. So the main track would be the on-cam mic and the problematic track would give a little fill-in.

I would also have some explaining to do with my client. I wouldn't lie (like saying it's the mic's fault or Uncle Bob's fault) because... well, lying never helped anyone and the truth will set us free.