Compressor settings
 

Hi,
Just wonder if and how you put a compressor filter on speech. Since speech is so dynamic I want to compress it, but the noisy pauses between words easily goes up to high. In this case it's a laveliere interview in a silent room (except distant traffic). If you use a compressor, what values do you apply? I work in Soundtrack.

I don't use Soundtrack, but here's how I approach compression for an interview.

First, ratio to apx. 2.5 (or, on some tools, 2.5:1).
Start with threshold at 0.
Attack at 5ms (milliseconds).
Release at 500ms (that's about half a second).
Lower threshold until gain reduction meter shows -3 to -6db.

This will get you started, then listen carefully on reference monitors and perhaps increase release time, reduce or increase ratio (probably the range you'll be working in is a ratio of about 2 to 4, the upper end is pretty extreme).

Typically, the goals for compression of an interview will be to:
* Make the words more understandable (sometimes this is to make up for poor mic technique).
* Help the voice stand out against a backing track of sfx and/or music.

Compression might also be used across the master in a final mix to limit the dynamic range of the entire program, for example, playback in a house supports a wider dynamic range than playback in a car.

If you can't hear the difference between this filter and straight audio then you might need better monitoring, or someone to sit with you and point out the difference in sound as you adjust these settings. Sadly, not everyone will be able to distinguish the effects of these different settings.

Ear training in general helps, musicians tend to be able to hear better. It's important to remember that hearing takes place in the brain, not in the ear, and therefore "ear training" actually can be very helpful, we can learn to hear better!

I like combining a compressed track with an uncompressed track. The uncompressed track keeps the peaks vibrant, while the compressed track ensures that soft sounds are not dropped.

For the "big announcer sound" try -25 dB and 20:1. It's very aggressive, but can be mixed with the original to taste.

You might still need to manually reduce stray peaks, or apply a limiter if the dynamic range is still too high.

Thanks, I'll try this. Do you normalize? Or use multiband compressor? I'm quite new to audio editing (have until now mostly done it quite simply in the video app), so I appreciate getting deeper into the workflow. Good Soundtrack Pro tutorials are appreciated!

I don't use the program you have, but are there possibly some presets you can try?

You don't necessarily "normalize" when you are using a compressor. However, since you will be bringing down the highs, it is often necessary to increase the overall output level, to bring it back up. This in turn can bring up the noise floor so a gate might be added to the chain, etc. etc. It's all pretty straight forward, but it seems complicated, and it can be with all the little settings. Thus, I suggest looking for some presets to get things pointed in the right direction.

If your program accepts plugins, here is a great one for voice:
http://www.voxengo.com/product/voxformer/

Here are some links to give you an idea of what you are dealing with:
http://en.wikipedia.org/wiki/Audio_level_compression
http://www.soundonsound.com/sos/1997...mpressors.html
http://www.homestudiozone.com/compressors.php
http://www.pcmus.com/compressors.htm
http://www.musiconpc.com/id_19/

And here's an article on recording the voiceover:
http://mixonline.com/mag/audio_recording_voiceover/

Thanks Jack, I'll check this out.

I use a multiband compressor and limiting on the final master, but it depends on the target. If you're targeting the web, squish it. If it goes to TV or a theater, tread lightly and leave some dynamics. The broadcaster will do their own compression, and a theater doesn't need any.

I over-compressed & limited this year's 48-hour entry, and it was WAY louder than everybody else's uncompressed mix. What was I thinking??? (Answer: after 48-hours of little sleep and too much caffeine, not much!!!)

For voice recordings for TV I almost always use something like this:

ratio: 6:1 - 10:1
threshold: depends on the source, mostly between -10dB and -25dB
attack: 0ms
release: 110ms

These settings work pretty well for me, I use them as a basis in either LogicPro, a Yamaha LS9-32 live mixer or in Final Cut Pro, for voice-overs, studio talks or eng interviews. However, I am not a sound professional, I have only basic knowledge in sound but I have to deal with sound quite often. That's why I like the most simple solution that works for me ;)

Heiko: And if doing it just for dvd, do you still compress voices?

Jon: 'squish'?

"Squish" means to squeeze or squash... or in other words, compress a whole lot.

The more difficult the listening environment, the more compression that is in order.

Radio, for example is compressed almost flat, with no dynamic range. Listen to a talk show. The host will go from talking softly to screaming and the volume will be about the same over the radio. If you record a radio talk show, then bring it into an audio program on the computer, it is obvious from the wave form how much compression there is. Also, when listening to the file with head phones, you can hear the noise gates that are used to cut out low level noise as they open and close.

Thus, for the internet, with bad speakers, compression is in order.

When I transferred a lot of education videos from VHS to DVD, I compressed the audio a bit and raised the level as much as possible. In noisy classrooms, DVD players are often not able to play loud enough at normal levels.

However, if something is going to be played on good equipment in a good environment, you want more dynamic range and overall compression would be much less.

One way to see how compression is used is to rip a track off a pop album, rip a track off a classical album, and compare the wave forms. The pop track is maximum loudness with very little dynamid range. The classical track has much lower average amplitude with a huge dynamic range.

There are standards (as outlined above) that apply in various situations. However, in general, for a DVD for example, the main issues to consider are:
1. type of material in recording
2. Equipment used for playback
3. Environment where playback will be made.

Thanks, Jack, I will chew this and squish the most out of your recommendations.

Excellent explanation, Jack!