Topic: Compression question

[colargol]

Hi!

I've been reading up on old threads about compression here, then trying it out using good old CoolEditPro 2.1. I have been experimenting with minimal compression to raise the level a tiny bit on the entire recording. But when I run the compression on my recording, I am very sure it actually changes the parts of it that are well below my lowest threshold. Am I doing something wrong? Any settings I should change?

See the attachments for my settings.

Thanks for any help.

Colargol

[page]

I have been experimenting with minimal compression to raise the level a tiny bit on the entire recording.

The over/under thresholds are different than most plugins I've seen (most deal with an over and possibly a soft knee value).

In this instance, I think you'd be better served by using a limiter. If you're only looking for 1db or so, a limiter will do that in a more simplistic fashion.

[Gutbucket]

I have been experimenting with minimal compression to raise the level a tiny bit on the entire recording.

What you are doing, and what Page suggests, does what I think you are asking to do.  To clarify the conceptual basis behind what this is doing, and explain sort of the opposite way of doing something which may seem similar, but really isn't, I'll try to explain what's going on with both.  I hope that will, in part, help explain two different ways of using compression, but I also mention this because I have had others ask this question in a similar way, and what they really were trying to do was this second thing.

What you are doing here with the way your compression routine is set will automatically lower the level on just the loudest parts, which in turn provides more empty room 'up top' in the 'range of available loudnesses' and allows you to increase the level of everything without 'hitting the ceiling'.  The compression is making the very loudest parts less loud, and the 'make-up gain' you apply raises the level of everything to take advantage of that empty space up top.  This is basically the same as just turning up the volume knob a bit more and leaving it there all the time, except for the very loudest parts where the volume knob gets automatically turned down and back up again rapidly.  A limiter basically acts the same and just turns the volume knob faster when it kicks in.

If your goal was to make the loudness difference less extreme across the entire recording, which is often described as "turning up the volume of the entire recording" in similar language, then you would instead want to set a very low threshold and use a very low compression ratio like 1:1.2 or 1:1.5 or something, not more than 1:2.  With the appropriate amount of makup-gain applied, that would be more like constantly adjusting the volume knob automatically the entire time, turning the quiet parts up and the loud parts down so that in general, the entire recording sounds 'louder'.  That's very much the opposite of limiting.

Both techniques are reducing the overall dynamic range, which allows you to increase the average level with makeup-gain.  But the second is reducing the entire range more or less equally, the first is reducing just the peaks.

I've used regular language above to describe this to help make understanding it simpler, but actually I've used the common terms 'volume' and 'loudness' incorrectly in a technical sense.  Volume is technically a measure of 3-dimensional space, and has nothing to do with sound levels or signal level voltage.  Loudness is technically the human perception of sound level, or how loud sounds seem to be to the listener, despite their true levels.  It's closely related to how we perceive what compression is doing, but the compression routines themselves only deal with signal levels.  So you can actually substitute 'signal level' for 'loudness' and 'volume' in much of the explanation above.

I hope that's more helpful than confusing.

[capnhook]

Clear now....I learned something today, thanks Lee..

[colargol]

I have been experimenting with minimal compression to raise the level a tiny bit on the entire recording.

What you are doing, and what Page suggests, does what I think you are asking to do.  To clarify the conceptual basis behind what this is doing, and explain sort of the opposite way of doing something which may seem similar, but really isn't, I'll try to explain what's going on with both.  I hope that will, in part, help explain two different ways of using compression, but I also mention this because I have had others ask this question in a similar way, and what they really were trying to do was this second thing.

What you are doing here with the way your compression routine is set will automatically lower the level on just the loudest parts, which in turn provides more empty room 'up top' in the 'range of available loudnesses' and allows you to increase the level of everything without 'hitting the ceiling'.  The compression is making the very loudest parts less loud, and the 'make-up gain' you apply raises the level of everything to take advantage of that empty space up top.  This is basically the same as just turning up the volume knob a bit more and leaving it there all the time, except for the very loudest parts where the volume knob gets automatically turned down and back up again rapidly.  A limiter basically acts the same and just turns the volume knob faster when it kicks in.

If your goal was to make the loudness difference less extreme across the entire recording, which is often described as "turning up the volume of the entire recording" in similar language, then you would instead want to set a very low threshold and use a very low compression ratio like 1:1.2 or 1:1.5 or something, not more than 1:2.  With the appropriate amount of makup-gain applied, that would be more like constantly adjusting the volume knob automatically the entire time, turning the quiet parts up and the loud parts down so that in general, the entire recording sounds 'louder'.  That's very much the opposite of limiting.

Both techniques are reducing the overall dynamic range, which allows you to increase the average level with makeup-gain.  But the second is reducing the entire range more or less equally, the first is reducing just the peaks.

I've used regular language above to describe this to help make understanding it simpler, but actually I've used the common terms 'volume' and 'loudness' incorrectly in a technical sense.  Volume is technically a measure of 3-dimensional space, and has nothing to do with sound levels or signal level voltage.  Loudness is technically the human perception of sound level, or how loud sounds seem to be to the listener, despite their true levels.  It's closely related to how we perceive what compression is doing, but the compression routines themselves only deal with signal levels.  So you can actually substitute 'signal level' for 'loudness' and 'volume' in much of the explanation above.

I hope that's more helpful than confusing.

Thank you, that is indeed helpful :-)
Will the limiter approach in any way compromise the sound quality in the loudest parts, or is this negligible if I only want to "cut off" a db or two?

-colargol 

[page]

Will the limiter approach in any way compromise the sound quality in the loudest parts, or is this negligible if I only want to "cut off" a db or two?

1) In practice, the better the plugin, the better the result. For a decibel or so, even Audacity's free limiter works if I use a wet/dry ratio of around 50% (I think it's called soft limiter in Audacity, would have to check to confirm). More then that and I'd reach for a third party one. I've been able to zonk around 6-8db after tuning Izotope's before getting a result I was unhappy with.

2) Limiters function as compressors with extremely fast attacks and releases. It's great for taming the odd transient here or there (super loud applause that is impeding your normalization of the rest of the set? Limit that), but you don't use it to control volume of a section so much as to buy headroom at the expense of the occasional transient. The differentiation is that compressors will affect much more of the signal than a limiter (if properly used) will.

[Gutbucket]

To expand on that-  improperly set, any compressor or limiter is going to compromise the sound.  Like Page mentions, the better quality ones pretty much get you two things: the ability to make more extensive changes in dynamics without them sounding bad; and easier and often somewhat more intuitive adjustments to get the dynamic change you want within the limits of what the routine is capable of.  The third, sort of subset of both those things is how good the plugin can sound when it is set properly within those limits.

The only way to tell when things are getting overly compromised is to listen.  Once you know what you want to adjust and what settings to target, hearing the change and deciding how much is too much is the trickiest part.

Here's a tip- often I find the best sounding approach overall is to use several different techniques, each making a relatively minor reduction, rather than trying to do too much with any one stage.  So you might manually fix those couple of really big spikes and that loud pop, run a small amount of some top down compression that helps better manage the loudest parts, some bottom up compression which brings up the quiet parts and makes the details more audible, and some limiting to reduce the transient peaks to allow you to get the average level up a bit more.  Similarly some mastering engineers run several passes of compression or limiting, each making only a minor reduction and with slightly different settings, so the overall change is less subconsciously predictable and audible. 

Try to listen two ways- intently and critically to the particular details of the changes you are making, then for longer periods without concentrating on the particular change or the specifics, just the overall feel and sense of the music and if that is right or not.  Both are important, and it can be quite difficult to switch mindsets to the second 'overall enjoyment' mode from the first 'critical analytic mode'.  If you have the time, I find it helps to go even one step farther- once you have settled on the best settings for everything and are happy with things in both senses, listen distractedly, doing something else without the music in mind, and see if your attention is drawn to problems in the music or if your overall impression is that everything is good.