By | January 18, 2007
If you followed along with last weekâ€™s post, you should have a voice track along with a music bed track. You should also have some volume cues on the music track so that the volumes go up or down with the voice track.
Before we mix everything down, we just wanna do a little bit of processing on that voice track. Hopefully your audio levels are somewhere around -6 to -3dB. Thatâ€™s a pretty good place to be because when youâ€™re recording, you need to have a little bit of headroom. Headroom is the amount of â€˜cushionâ€™ that you have between your loudest sound and 0dB, which in the case of digital recording would mean distortion.
When your signal goes over 0dB, itâ€™s gonna sound terrible, so you wanna keep it at or below 0dB. To do that safely though, itâ€™s a good idea to shoot for something like -3dB and then if you need to, you can amplify it a bit, or better yet, you can apply some compression, which is what weâ€™re gonna do today.
Some of what Iâ€™m gonna tell you is kinda dependant on you specific situation, and a lot of people that give advice like this donâ€™t give specifics for that very reason, but I think I can give you some real numbers and just tell you where you can play with it.
Letâ€™s talk just a little bit of theory for a minute. The reason you want your signal in the -6 to -3dB range has to do with what is called the signal-to-noise ratio. When nobody is talking into your mic, and thereâ€™s nothing special going on, thatâ€™s your noise level. Try recording without anything going on and youâ€™ll see a level of noise.
You want your talking voice to be a lot louder than that noise level because any amplification or compression that we do, is going to affect that noise. Hereâ€™s what I mean. If your noise level is around -30db and when you talk it reads about -20db, that means your signal to noise ratio is 10dB, which isnâ€™t very good. If you amplify your signal, the noise is gonna get amplified as well.
On the other hand, if your noise level is around -50 and when you talk you register -6, thatâ€™s a difference of 44dB, much better. Now when you amplify your voice, itâ€™s mostly voice that gets amplified.
So youâ€™re gonna try to record your voice at between -6 and -3dB. If you go under a bit, or over a bit, no big deal, as long as you donâ€™t go over 0dB, which is called clipping. If you actually do that, youâ€™ll know why itâ€™s called clipping, your signal actually gets clipped.
Instead of doing a straight amplification, weâ€™re gonna use a compressor that comes with Audition. Before we do that though, letâ€™s get rid of as much noise as we can from the source signal.
In an earlier podcast we used the Noise Reduction effect and now weâ€™re gonna do the same thing. If youâ€™ve got your session open, double-click it so that youâ€™re in Edit View. Find a segment of silence and hit F9. Highlight the noise. Right-click it and capture the noise reduction profile. Then select the entire file and click Effects, Noise Reduction, and then Noise Reduction.
If you want you can play with the settings later, for now just take the defaults. This filter is gonna go through the whole file and filters out the sounds of the noise. Now it canâ€™t take out everything, but itâ€™ll usually give you something workable.
Go ahead and play your file now and see how it sounds. It should be a lot quieter. Also listen for digital artifacts which sometimes pop up if you applied the filter too strongly. Itâ€™ll make your voice sound very digitized.
Now with the entire file still selected, click Effects, Amplitude, and then Dynamics Processing. Without going into too much theory here, Iâ€™m just gonna give you the settings I typically use for a podcast voice track, and again, you can play with it if you want to. Iâ€™ll probably dedicate a whole post to how compression works later on, for now, make sure you have the Traditional tab selected, that way you can punch in some numbers. Thereâ€™s a picture on my post that shows you all the settings. Just make them match.
Click on the Attack/Release tab and take a look at the next picture. Basically, you want a quick attack time and a fairly quick release. Hereâ€™s why. When you start talking, you want the compressor to kick in right away. When you stop though, you might start again quickly, so you give yourself just a little bit of time, a split second. If you donâ€™t start talking within that 50 milliseconds, then the compressor is gonna shut the gate. Again, I donâ€™t wanna get too detailed about this just yet, so weâ€™ll leave it at that for now.
If you want, you can save your settings as a preset so you can pull them up later. Iâ€™m also gonna show you how you can apply this effect without actually changing your file, but that comes later.
So click OK and your audio is gonna get leveled out. This helps quiet passages sound louder, and louder passages sound a little quieter. It also helps your noise levels go way down when nobodyâ€™s talking.
Weâ€™ve got one more thing to do before your podcast is ready and thatâ€™s a little EQing. With the whole file still selected, click Effects, Filters, and Graphic Equalizer. Now this part, I can only give you a little advice. Everybodyâ€™s voice is different. If youâ€™re a woman youâ€™ll do this differently than if youâ€™re a man.
For a male voice, go ahead and give yourself a little bit of bass, not too much though. Donâ€™t make yourself sound fake. Also add some subtle highs and a little bit in the 1 to 2k range. Iâ€™ll post a picture of my EQ settings so you can get an idea of where to start. Use the preview button to hear what your voice sounds like with the EQ applied. Audition has a preset called 30-band Punch and Sparkle and thatâ€™s a good place to start. Also, if youâ€™re using a real Graphic Equalizer you can kinda make it match these settings and then you donâ€™t have to do it with software. If youâ€™re using a Sonic Maximizer, chances are you wonâ€™t need to do any of this EQ stuff at all.
So click OK and now we have a processed voice file. Listen to the voice all by itself and see what you think. If you use the Preview buttons then you can tweak the sounds before applying them. In another podcast Iâ€™m gonna show you how you can use the Effects bus to apply these effects without having to actually change your source file, which is an excellent feature.
Now go back to multitrack view and click File then Export and then Audio. Save your mixed down file as an MP3 and youâ€™ll want to look at the MP3 options and try one of the presets, like maybe 128k. I donâ€™t recommend that you go much lower than that because you wanna have a good quality broadcast and itâ€™ll start to get noisy below about 96k.
Give your newly created MP3 a listen. I always recommend that you listen to the whole thing once itâ€™s mixed down, just to make sure you didnâ€™t make a mistake somewhere. At this point you should have a good sounding MP3 and weâ€™re ready to talk about some hosting and platform options.