An Introduction to Equalization - A Free Guide from Audio Masterclass

Equipping Your Home Recording Studio - A Free Guide from Audio Masterclass

An Introduction to Compression: Basic Compression - A Free Guide from Audio Masterclass

Where in the signal chain you should compress, and where you should not

A post by David Mellor
Monday August 14, 2006
In the entire signal chain from microphone through recording to loudspeaker, there are certain places you should compress, and certain places you should not. Where and where?
Where in the signal chain you should compress, and where you should not

In a recent article on managing the noise level in vinyl mastering, a Audio Masterclass visitor commented that compression makes noise more audible, so compression can not be used as a means of improving signal-to-noise ratio.

This is an interesting point. Compression always reduces signal-to-noise ratio. You can see this easily - the compressor works by reducing the peak level of the signal. Overall therefore, level has been lost. This is restored using the make-up gain function of the compressor.

But doing this raises the noise floor. So it is inherent in compression that the signal-to-noise ratio is reduced and the signal is noisier than it was before.

But this contradicts the article on managing the noise level in vinyl mastering where it was advocated that compression should be used to improve the signal-to-noise ratio.

Surely there's a conflict of opinion here?

Not really. It's a question of where in the signal chain compression is applied.

Let's suppose that you have made a digital recording (without compression) that you want to have mastered to vinyl. Before mastering, the majority of the noise will come from the microphone and microphone pre-amplifier (and indeed the mic will pick up acoustic noise too).

But if the recording is well-made, the dynamic range will be wide and the noise level low.

If this is transferred to vinyl, the highest signal level will be limited by the maximum acceleration that the stylus can be expected to track on playback of the finished product.

Since there is a wide dynamic range in the original recording, the quiet sections will be a lot lower in level.

Now, since vinyl is inherently a noisy medium, much more so than even a 16-bit digital recording, during the low-level sections, the noise from the record will be clearly audible.

To prevent this, the recording should be compressed while it still has a good signal-to-noise ratio.

You can compress a 16-bit digital recording by as much as 20 dB and it will still have a signal-to-noise ratio of 70 dB or so.

This is significantly better than the noise level of vinyl, so any noise in the digital recording that has been brought up in level by compression will be lost among the vinyl noise.

So although the signal-to-noise ratio of the digital recording is worse, this will not be audible on vinyl.

And the benefit is that now all the low-level sections of the recording are higher in level than they were, and some distance above the noise level of the record.

The net result is an increase in signal-to-noise ratio, in the end product.

In conclusion, although compression always increases noise, if you do it in the right place in the signal chain, the net result can be a better signal-to-noise ratio than before.

A post by David Mellor
Monday August 14, 2006 ARCHIVE
David Mellor has been creating music and recording in professional and home studios for more than 30 years. This website is all about learning how to improve and have more fun with music and recording. If you enjoy creating music and recording it, then you're definitely in the right place :-)