Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[hit_ny]

Compiling a wedding video for a friend and find the MC at the reception (open-air) is sounding a bit muffled. The lower tones of his voice seem to overpower the mid-range ones. It's not like the volume isn't high enough, but could be better if one did not have to strain to hear what he was saying. The footage done indoors is fine but i guess recording outdoor sessions is an art in itself. (nope, i wasn't the cameraman )

I tried playing in Wavlab and lowering, raising certain frequency bands but not too convinced with the result. Best i end up is making it sound nasal and telphone like. There isn't much music accompanying the MC just speech.

It could be that nothing can be done, but maybe not.

Any ideas?

[glynor]

Recording outdoors is certainly a much more challenging "art" than indoors.  That's why they pay all that money for fancy shotgun microphones and boom mics on TV news crews.  With good software tools (WaveLab is perfectly good too), you might be able to clean it up some, but... In my experience with these things (and it is sort-of my job) the general rule of thumb is:

You can't make chicken salad out of chicken crap (using "crap" because the forum will edit the real phrase).

Adjust your expectations accordingly and good luck.

[hit_ny]

Right, i certainly did not see any boom microphones during this gig

The voice becomes more intelligibile if you boost the volume up on the tv, so maybe something could be done. But then the sound levels are already recorded at 0dB peak, can't increase gain any more here. So instead of cutting the bass, push the mids up a little.

Voice is what, somewhere between 1khz & 4khz, maybe a 3dB jump across that band. I can't hear much of a diff, maybe 6dB, then the question becomes if the difference is clearly audible, it might be worth it even with a re-encode at the end.

I was hoping there would be some kind of plugins that could do this sort of thing, but AFAICT they are only available for vinyl recordings, where there is a constant frequency which can be identifed and removed easily.

[Tangoman]

Hi hit_ny
One would really need to hear a sample to advise good. This is my work and hobby. I have been in the sound business since the mid 80's everything from pirate radio to Oscar winning features. My fun at the moment is restoring Tango music recorded in the early 1900's from 78 rpm's  if you fancy to send me an mp3 a minute or 2 or 3 I maybe can advise.
email to gerrysan     at   gmail    dot    com
Cheers Tangodude

type Gerard Roche into IMBD to see an online resume

[hit_ny]

Thx for the offer Tango, and apologies for getting back so much later but i've put the sound restoration on the backburner for the moment and instead working on the slideshow with the large amount of pictures.  The slideshow would be included on the DVD with the video segments as well. This has been quite a long exercise, the result is quite appealing, pictures, that pan, zoom etc  more lively than a static slideshow.

Which brings me to the next quesiton, soundtrack, its been challeniging putting music to the photos, i can't compose the music myself so i have had to use various tracks that fit as well as i can make the pictures dance to them. MC has been quite helpful here as sometimes i would try to think of words that describe what i was watching and then use them as search items for track names. Found a few tracks i might not have normally selected based just on rating but fit the mood better.

Now the tracks all have different sound levels, i would like the sound to be fairly uniform along the show. So you can watch it without having to hit the volume button every now & then.

does one
- just normalise the big wav at the end to 0dB ?
This would just knock some of the louder sounding tracks a bit down. I was playing with Pinnacle's Clean and it has a meta-normalise function, sounds like album gain to me.

- Analyse each track in wavlab and apply a gain accordingly so that the average levels are similar for each track ?
This would make the softer segments stand out a bit, there is variation in sound levels within some tracks, but not too much. This option is more time consuming, as you would have to work-out how much gain is required to get to the desired level..for each track.

- I might be able to use the values MC calculates here for replay gain and just add/or take away that much gain for the individual wav. So in MC's settings set it to make tracks the same volume and use the resulting +/-dB calculated.

Any thing else i might have missed ?

[Tangoman]

Hi NY,
If it was my call I would go down the wavelab route. Make the video import it then lay your music adjust as necessary burn when ready. I reckon this would be the quickest way. Simple volume graphing is the way to control your sound levels. Hope this helps.
Tangodude

[hit_ny]

Could you explain what you mean by volume graphing ?

setting the y-axis to dBs, scrolling through the waveform, and lowering gain in sections where its 0dB ?

..there is no clipping anywhere in the soundtrack.

If so, what avg level do you aim for ?

When i made the mix for the slideshows, i nomalised all the tracks to 0dB ( had to use goldwave for this, wavelab will normalise to anything lower but 0dB), then analysed them in wavlab to see what their RMS, peaks & avg were, i found for some tracks the RMS peak was higher than others, i lowered the gain (-3dB) on those tracks till i got a range of (-7dB to -10dB) for the peaks, at this point i could tell the mix sounded ok.

Doing the soundtrack in the video where ppl speak is much longer by this route.