dB level for dialog .wav files?

I recorded dialog with boom and mic, to digital wav files (H4 Zoom recorder). Editing them in Sound Forge software. For normal dialog (talking, not shouting), what is a generally good dB level I should be seeing, or should adjust (using soundforge tools) the waveforms to? For example, should my peaks be around +/-3 db on the volume scale? +/- 6db? Other? Just looking for a ballpark figure. Otherwise it is tricky to know what volume I should be hearing since preceived volume is tricky due to my human ear sensitivity, volume knob on my speakers/headphone, volume setting for the soundcard using the software mixer, etc.
 
Thanks Mangler, much appreciated. That also fits with material I was reading this morning in The Filmmakers Handbook (incredible book, I read the section on recording sound levels and it is the first book to make it so understandable, I actually get it now, the whole dB thing, VU, 0dB, -6dB, etc.).

dialogue should be around -6db
 
I recorded dialog with boom and mic, to digital wav files (H4 Zoom recorder). Editing them in Sound Forge software. For normal dialog (talking, not shouting), what is a generally good dB level I should be seeing, or should adjust (using soundforge tools) the waveforms to? For example, should my peaks be around +/-3 db on the volume scale? +/- 6db? Other? Just looking for a ballpark figure. Otherwise it is tricky to know what volume I should be hearing since preceived volume is tricky due to my human ear sensitivity, volume knob on my speakers/headphone, volume setting for the soundcard using the software mixer, etc.

For your final mix levels it's going to depend on your delivery format - theatrical release, DVD, analog or digital broadcast all have different requirements.

For the mix you need to calibrate your environment so you can reproduce a standard setup. Get some setup tones - Blue Sky has some free for download on their website - consisting of mono 1kHz sinewave tone at a level of -20dBFS and pink noise at -20dBFS RMS (will read about -12 dBFS on a peak reading meter.) You also need a $50 Radio Shack sound pressue meter. Working one channel at a time, load the sine tone into your NLE and adjust the levels on the software meter to be outputting -20dBFS. Set the sound card output to maximum and leave it. Switch to the pink noise and wthout touching the software level controls, adjust the playback volume on your monitor until the sound reads 82dBSPL (85dB for theatrical mixes) on the sound meter held at the head position at the mixing desk. Mark the position of all controls so you can return them to the exact same spot.
 
Thanks Steve. It really is more complex than one (an amateur like me) would imagine-- one would think you just boost db volume, listen to playback, but what you say fits a bit with some material in a book i have been reading. So many links in the chain that determine "volume" I knew there must be some way to standardize the volumes somehow.

What is pink noise?


For your final mix levels it's going to depend on your delivery format - theatrical release, DVD, analog or digital broadcast all have different requirements.

For the mix you need to calibrate your environment so you can reproduce a standard setup. Get some setup tones - Blue Sky has some free for download on their website - consisting of mono 1kHz sinewave tone at a level of -20dBFS and pink noise at -20dBFS RMS (will read about -12 dBFS on a peak reading meter.) You also need a $50 Radio Shack sound pressue meter. Working one channel at a time, load the sine tone into your NLE and adjust the levels on the software meter to be outputting -20dBFS. Set the sound card output to maximum and leave it. Switch to the pink noise and wthout touching the software level controls, adjust the playback volume on your monitor until the sound reads 82dBSPL (85dB for theatrical mixes) on the sound meter held at the head position at the mixing desk. Mark the position of all controls so you can return them to the exact same spot.
 
What is pink noise?

Its a semi-random noise signal that has equal strength across all frequencies. Its rather like "white noise", only with more low-frequency energy. It sounds rather like a rushing waterfall. As Mr. House described, its an awfully useful diagnostic and benchmarking tool when your setting up a monitoring environment.
 
O/T Pros can modulate their voices so you get 'even' levels without strong peaks.

If they're non pro voices for a non broadcast DVD, to get them to modulate and sound better, I'll go through the wav files looking for big peaks and in Vegas or Soundforge, reduce them so I can get the average voice level up.

Non pros never do the same thing twice, they tend to drop away at the end of sentences too, so I'll maybe boost there slightly, maybe a couple of db.

Started doing this years ago in TV spots to boost the non pros to match the announcer doing the tag. IMO doing that with a good compressor or limiter doesn't sound the same.

Cheers.
 
Its a semi-random noise signal that has equal strength across all frequencies. Its rather like "white noise", only with more low-frequency energy. It sounds rather like a rushing waterfall. As Mr. House described, its an awfully useful diagnostic and benchmarking tool when your setting up a monitoring environment.

Actually white noise is equal energy for all feq's, pink noise is equal energy per octave. It is essentially white with a low pass filter on it. But you want to make sure your using calibrated pink noise. Blue Sky is good and Dolby pink is good. Some "pink" noise is filtered but isn't to spec and that will throw things off. Also it needs to be at the right level.

If your in a small home setup you will probably want to set your monitors a bit lower. 85 is the mix stage spec but that is a big space. A lot of folks on nearfields set the level around 79 - 80spl. If your in that ballpark (± a couple of dB) you should be fine but you need to pick a number and stick with it.
 
Back
Top