Best way to set sound levals with various projects

flicktease wrote on 5/11/2003, 9:19 PM
Whenever I make a longer project I usualy complete a section & render it as an AVI file. I then put a collection of AVI files on the timeline & output as one large file. What I noticed is that the waveform on the timeline has a wide variation of heights from one clip to the next. Obviously I didn't do a good job of keeping each section around the same volume. I have tried the normalize command on the quieter sections but this produces only the smallest improvement. I have tried putting a volume envelope on the audio track & increasing the volume of the quieter sections. This makes these sections louder but moving the line doesn't redraw the waveforms as a guide for getting the sections all about the same loudness. What is the best way to correct this?

Comments

GaryKleiner wrote on 5/11/2003, 9:57 PM
>moving the line doesn't redraw the waveforms as a guide for getting the sections all about the same loudness.<

You should be setting your levels using the master bus of the mixer as a reference.

Gary K.
filmy wrote on 5/11/2003, 10:07 PM
Audio is percieved a lot of times. Meaning just because the sound wave itself isn't "loud" does not mean the actual sound will not be percieved a being loud. I don't know what your project is but remember dynamics are your friend.

So that being said - If you don't have it already get Sound Forge and make use of the "wave hammer" plug-in. I used to do this by hand on the Amiga with Studio 16 so trust me this is much much easier. No matter what you are doing please be aware of dynamics from scene to scene. If you are trying to match audio from one scene in the same setting to another scene in the same setting you might want to use the Acoustic Mirror plug-in.

Also what is your final output going to be? Do you need wide dynamics or narrow? Do you need the ultra low bass? There are tons of DX plug-ins out there that will help you get your audio to where to want it but the most important thing to start off with are your ears, and, maybe even more important, your clients ears. Even if you don't like what you hear your client might love it or vice versa. ("I want the gushots louder! Make them BIGGER sounding") But make sure the client is going to watch and listen in the same enviroment that they will after they leave your place. Don't play it back on PC speakers if they are going to play it back in a large boomy hall because it will *not* sound the same. Likewise don't play it back on a THX Certified system if all they are going to do is plop it into a mono vcr for use in staff meetings.

A common mistake that people make when looking a graphs is they think what they see, if it isn't all the way at the top, isn't correct - or isn't loud enough. For example the music can be loud if it is all that is going on and it is an action scene or a loud club - but if it is a sad scene the music could almost be 'off' if there is nothing else in the scene. Same for dialog or narration. Yelling can be loud - but a yelling whisper doesn't have to be as loud on a graph. Having said that - if that yelling whisper takes place in a loud setting and the plain 'ole yell is in total silence the graph could show you something totally different. It does not mean you need to make one as loud as the other.

And finally to repeat what I first said - dynamics. Learn about them and love them. If you do an action scene/film and the producer wants to hear those punches and swishes cut through and you need to have all the other effects/music/dialog cut through as well you need dynamics. Simply making the punches louder won't make the mix better. (Much the same way simply adding grain to the video doesn't make it 'film look') It doesn't even mean the punches or swishes would be the "loudest" thing on the graph, it just means that how you add the elements hopefully will work in how the audience percieves the loudness of them.

Hope that was of some help. I tried to keep it basic and hopefully easy in it's concept.