|
|||||||||
|
Thread Tools | Search this Thread |
April 26th, 2007, 01:05 PM | #1 |
Major Player
Join Date: May 2005
Location: Portland, OR
Posts: 277
|
Final Audio Track Processes? What do I do?
Dear Everyone.
After an unbelievable 2 years, my documentary is getting to be done and it's time to clean up the levels on the audio. Everything's in place, but some things are low, some are high, and none of it has that lovely "caterpillar" look of a finished waveform. You know, like when you look at CD audio. It's normalized, I think is what I mean. I'm a video guy, and so audio, despite its importance, has taken something of a back seat. What should I do to get it in shape? General or specific advice is great. I am using SOUND FORGE 8 DJ Kinney |
April 26th, 2007, 01:51 PM | #2 |
Inner Circle
Join Date: May 2006
Location: Camas, WA, USA
Posts: 5,513
|
Step 1: Upgrade to SoundForge 9.
Why? It includes noise reduction and mastering tools. Being a documentary, noise reduction will probably be very important. It has a mastering EQ that might help bring the dialog up front. The multi-band compressor and limiter are important for dynamic soundtracks - gotta get those drums to rock and the explosions to rumble. It might not be so important for a documentary, unless you are in a construction yard or a shooting range. With the multi-band compressor you can compress the bass frequencies while letting the mids and highs sing. And the limiter can tame the occasional peak with a soft, rather than hard, clip. You will want to mix the dialog so the meters max out around 12 dB under peak for comfortable dialog. That leaves room to let yells get louder and whispers quieter. Keep the dialog centered, pan any music or ambiance wide, if you can. For EQ, 300Hz or so is the voice fundamental. This area adds warmth and fullness, but can be muddy alone. 1200 Hz or so is where you find the consonants. Try boosting this region, if intelligibility is an issue. 2400 Hz is the range for differentiation. If two people sound similar, this range can help bring out the differences. (Same with musical instruments.) The 5k to 15k range is the "sizzle". Too much and you get noise. Just the right amount adds air. In general, you want to cut the unwanted frequencies that are outside of these ranges, and leave the EQ flat in the critical frequencies. If you boost, use a relatively wide boost, rather than a narrow one. Also, you can use some general compression to help keep the voice levels even. If you have the time, boosting and dropping the words syllable by syllable and line by line is best. What the mic records and what we expect to hear for consistent dialog are two different things. Watch especially closely for people who drop certain sounds, or drop their volume at the end of their lines. Best of luck!
__________________
Jon Fairhurst |
April 27th, 2007, 01:29 PM | #3 |
Major Player
Join Date: May 2005
Location: Portland, OR
Posts: 277
|
That's some incredible advice. Thank you very much. So Limit and normalize, huh? I've just never read much on how it all works, so I guess I will.
Thanks again, DJ Kinney |
April 27th, 2007, 05:03 PM | #4 |
Inner Circle
Join Date: May 2006
Location: Camas, WA, USA
Posts: 5,513
|
I'm glad to help.
So, Portland, huh? I think I've heard of it. ;)
__________________
Jon Fairhurst |
| ||||||
|
|