Sign up ×
Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It's 100% free, no registration required.

Total newbie here, not sure if this is the right place to post this. I'm trying to record some station IDs for an online radio station thingamajig I'm doing. I'm using a Samson C03U. My voiceovers sound okay when isolated.

However, I've noticed that my voice sounds very quiet when played over music, even when I lower the volume of the music and raise that of my voice. I've made sure to record my voice at the highest possible volume without generating peaks and to stay close to the microphone. The waveform doesn't look odd to my untrained eye.

I've found these related threads on the matter, but I'm wondering specifically if there are any audio FX plugins for Sonar I can run my voice through to help me achieve better results.

As I said, I'm a total newbie so I don't know anything about compression and all that jazz. I find it hard to wrap my head around abstract sound theory without hearing how certain effects affect input. Until now I've just been trying out stuff and seeing what sticks.

My microphone has a bunch of recording modes but I don't know if they're relevant to what I want to do. The theoretical explanation in the manual isn't doing it for me.

share|improve this question

migrated from Jan 24 '14 at 12:01

This question came from our site for engineers, producers, editors, and enthusiasts spanning the fields of video, and media creation.

If you need more help with this, just tell me in the comments! – Friend of Kim Feb 13 '12 at 20:26

5 Answers 5

up vote 7 down vote accepted

Just add a compressor on your voice-over. Set the ratio to 2:1. Set a middle long release and a bit faster attack. (Or auto if you have it.) Then you lower the threshold until it gets better. Just play with the ratio and threshold until you get it right.

What the compressor does is that if the audio gets over the threshold in volume it will wait the number of milliseconds you set as the attack. Then it will increase 1 dB per 2 dB the audio really increases. This is the ratio. 2:1 is this: If the audio increases 4 dB, the output will be 2 dB increasment.

When the level dips below the threshold it will turn off the compressing (the 2:1) after the selected milliseconds in release. This will turn down your volume, thus allowing you to turn up everything. Therefore everything will be louder, instead of like it is before the compressor: Not loud, with loud peaks. The compressor reduces the peaks, so you can increase the volume to get the overall volume up.

Good luck!

EDIT: Since the compressor turns everything down you can increase the output gain to turn the volume back up.

The soft knee turns the compressor on gradually, the more knee the more soft in the beginning. It's like this: If you have ratio to 2:1 and threshold of -10.

Without knee (in consept):
-10: -10
-8: -9
-5: -7.5
-2: -6
-1: -5.5

With knee (in consept):
-10: -10
-8: -9.5
-5: -8
-2: -6
-1: -5.5

The gate threshold is the minimum volume the audio got to have to go through the compressor. A gate is simply a expander (see below) with attack = 0 ms, release = 0 ms and ratio = 999999999:1.

Oh, I see that you have set up the expander! The compressor ratio turns the volume down if it gets over the threshold, the expander ratio turns the volume down if it gets BELOW the threshold. Thus making the sound twice as low if it is below the threshold, as you have set it up now. Set it to 0, because you don't want to reduce what's below the threshold, because you want to reduce what't over the threshold to turn UP what's below it, not DOWN.

The graph tells you how the audio will be affected by the compressor. The x-axis (the horizontal) tells you the input, and the y-axis (the vertical) tells you the output level.

share|improve this answer
2:1 is rather unusually gentle for a voice-over, so that might well not be enough. But it's certainly a setting one can start with. – leftaroundabout Feb 13 '12 at 20:44
Yes, indeed. Yes, go higher if it isn't enough. But, if it starts sounding flat and/or dull, then you have too much ratio, or too low threshold. – Friend of Kim Feb 13 '12 at 20:45
I think I understand the concept of compression now. I'm gonna need some help with this, though. I recognize the Compressor Ratio, but I've got all kinds of knobs and settings I don't understand. What do I do? – Pieter Feb 13 '12 at 20:50
Look at the updated answer! :=) – Friend of Kim Feb 13 '12 at 20:59
Updated and added the expander. – Friend of Kim Feb 13 '12 at 21:03

A lot's been said about compression but nothing's been said about intelligibility. If you're producing VOs, trails, bumpers, idents or jingles with vocals over a bed, EQ is also your friend here. If you critically listen to professional productions, they often lowpass AND highpass music under speech in addition to heavily compressing and EQing the vocals. Here's how I'd produce a jingle with music:

on the music:

  1. compress at a high ratio (start at 5:1 and consider anything up to or over 10:1)
  2. lowpass the bed at ~15 kHz and highpass at ~200 Hz
  3. take some of the high end (10 khz+) out of the music and also notch out a little at 1, 2 and 5 kHz

  4. ...and on the vocals:

  5. add an exciter to the vocals and adjust to taste (basically adds some 'shimmer' in the high frequencies, the smallest amount of short-tail reverb and likely also some gentle mid-side compression to subtly widen the stereo image
    • If you want to do it the manual way, sculpt the EQ - fill in the gaps you made in the music, so 1, 2 and 5 kHz (area of greatest intelligibility in human speech to begin with) and also crank the frequencies above 10 kHz and below 150 Hz.
  6. compress the vocals at a high ration (>=5:1,
  7. set the master buss compressor so it 'pumps' nicely; the audio rides at a consistent level until the vocal, at a much higher level, makes the compressor duck the level of the mixed signals whilst keeping the main output level the same.

You could even slap a multiband compressor on the output instead if you want to further sculpt the response in frequency bands.

Overengineered? Probably. However, that's the standard process I'd go through for "pro" radio jingles. You can get all fancy-pants on them but think of the following whilst editing:

  • what kind medium am I producing for? (radio which is always heavily processed to begin with to deal with small speakers and their limited frequency response)
  • ASSUME MONO! Clock radios, portables, in-store PAs and even TVs don't necessarily have two speakers so if you apply loads of mid/side processing or 'stereo widening' effects your music might disappear when summed to mono!
  • how obnoxiously in peoples' faces do you want to be? The station will likely use processing to 'calm' any severely loud signals, but there are now established EBU standards for "loudness" which you would do well to read up on. Nobody likes having to reach for the radio to turn it down during the ad breaks. I vastly prefer a jingle / advert / ident where it's CLEAR but not simply LOUD.

Final afterthoughts: when setting up a mic, set it to cardioid or super-/hypercardioid (to eliminate as much room noise as possible) and experiment with 'working' the mic; the closer you are to the capsule, the more 'presence' you will hear in the resulting sound.

The same voice recorded six inches and thirty-six inches from the same microphone will sound remarkably different, and you usually want to go for the close-miced option for the desirable boost in the midrange.

Happy producing :)

share|improve this answer
Too bad I can't accept multiple answers -- that's helpful stuff. I'll try this out today. – Pieter Feb 15 '12 at 15:24
I haven't used the exciter before. I'm seeing this and I'm not sure what the knobs do. Any settings you can recommend for a male voice recording? – Pieter Feb 15 '12 at 16:43
Sorry I never saw your question -- never got notified about it. The forte icon with 6000 looks like the centre frequency; drive is likely the amount of saturation or compression applied around that frequency and spread is probably adjusting the mid/side balance. Mix is simply the wet/dry balance coming out of the plugin, if you were using it as a Send effect you'd set it to 100% wet. Most of the time it's on a single channel strip as an Insert so you'd adjust that to taste around its mid point. – Christopher Woods Apr 15 at 18:48

In addition to the fine suggestion from 50ndr33 to compress your voice channel, you might also set up your compressor to use a side channel: use your voice channel as a trigger to compress everything including the music. This will cause the music to "duck" when you speak.

share|improve this answer
I don't know if I can do this with Sonar LE, but I'll investigate. – Pieter Feb 15 '12 at 16:22
This is very possible, especially if you were to send all the music to one bus and put a compressor on that bus. Sidechain the compressor on that bus to your voice and boom - the music will duck every time you speak. – Sean Michael Dorian Feb 20 '12 at 18:07
The compressor has a side-chain option. I could put a compressor effect on the master bus. But how does it know which tracks need to duck and which tracks should be dominant? – Pieter Feb 21 '12 at 19:36
I guess you'll have to read the manual. From your screenshot it looks like maybe you click on the side channel and it offers a choice of what you'll route to it (the voice)? But maybe you don't want to duck everything, you want to duck one bus containing all the music. – Liudvikas Bukys Feb 21 '12 at 20:22
This is what I'd suggest you try. This is my go-to before EQ and compression on the voice-over track. – Sean Michael Dorian Mar 12 '12 at 1:30

Are you using an effect called "ducking"? Ducking is used all over the place. Basically what it does is automatically reduce the volume of backing tracks (music in this case) when it detects input from an audio source such as a microphone.

This effect is used in DJing all the time and it's also used in many video games and movies.

share|improve this answer

"Trying out stuff" is still an option for even the seasoned recording engineer who is faced with some thing they have not experienced yet. There is no end to the over all learning curve as technology changes the target moves.

Before you try to compress your voice over track, review your signal path from mic to recording software to be sure all is optimized. Try another mic to see if you see any improvements, also with the original mic try a different recording package. Are you getting the same result when changing these variables?

Where are you recording the voice? Are you in an isolation booth like environment that blocks all the ambient background sounds?

Consider that there are natural born talented voices and some of these folks have developed this talent with a lot of hard work and now make a living to do VO work. Their voices are already pleasing and interesting and carry volume without compression from the get go. Although audio engineers like to boost things anyways. :>

Treat your voice over like you are going on stage at the MET. Warm up your voice. Sing some scales, hold long notes, take deep breaths, stand up right, learn to manage your diaphragm, and rehearse the content. Get a friend to listen to you while you perform for positive feedback.


Lastly, you might not be cut out for this kind of work, so don't be too proud to allow a substitute to see if this is better or not.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.