Question on how compressor works

[neueliteratur]

Hi, here's something I've always wondered when using compressors,

as I understand it, a compressor reduces the volume of a soundwave that goes over a certain threshold. Intuitively, it makes me think that the compressed file should sound quieter than the initial file, but the result is the contrary. Can anyone explain why this is?

Appreciate it,

[BertKoor]

http://en.wikipedia.org/wiki/Dynamic_ra ... akeup_gain

[Burillo]

the result is "contrary" only if you add what's called a "makeup gain" after compressing. when you compress, the difference between loudest and quietest sounds decreases, which means your maximum peaks are lower and the overall sound is quieter. which means you can then turn up the overall volume without any negative effects. in other words, if your file peaked at -3dB and you compressed it so that it now peaks at -5dB, you can now bring up the overall volume by 2dB and get peaks at -3dB, but with the uncompressed parts being 2dB louder than they were.

[arseniy2]

Because you turn up the input signals or turn up the output of the compressed result.

[DJ Warmonger]

Compressor reduces the peak level of incoming track, allowing you to turn up teh volume without clipping.

[JCJR]

Some compressors, in attempt to be more user-friendly, and perhaps in case of some hardware devices, to reduce manufacturing expense by reducing the number of knobs, feature automatic make-up gain. Sometimes the feature can be disabled but sometimes it is always on.

In those, when you reduce the threshold, it automatically adds the amount of makeup gain it thinks you need. Depending on attack and release settings, it can sometimes be difficult for a device to automatically calculate the required amount of makeup gain, so the auto-makeup gain feature is a blessing on tracks for which it works, but a curse on tracks where it fails.

Though one might not see them much anymore in music/recording use, there are at least two approaches to the problem-- Downward compressors and upward compressors. Downward compressors don't modify quiet signals but they reduce loud signals (and then makeup gain can be applied).

Upward compressors normally have very high gain in absence of signal, and then as the input level increases the gain is reduced toward unity, and in some cases will reduce gain well below unity with a loud enough signal. Common examples include the compression stage of compander noise reductions such as old dbx noise reduction systems, and the ubiquitous auto level circuits found in old cheap cassette recorders, dictation machines (even digital dictation machines), and most likely used in assorted online chat softwares and perhaps some telecommunications systems today. The auto level feature optionally available in some television sets might more logically be accomplished with an upward compressor rather than a downward compressor, though perhaps different models have different solutions to the problem. The compression in some of the better fancier electronic hearing protection muffs and "hifi hearing aid" earplugs appear to be upward compression designs.

[robogone]

Compressors often have an input and output gain. If they are not set the same then you may notice a difference in volume - what you hear may be because the output gain is set higher than the input gain.

[neueliteratur]

Thanks.

Another question on how compression works: How do you compress a song that has quiet parts and loud parts (which is often the case)?

-When I try to compress the whole song according to the loud part: the quiet part doesn't get much compression.
-When I try to compress the whole song according to the quiet part: the loud part clips.
-If I compress the parts separately, the sound texture of each part becomes too different from each other.

[PurpleSunray]

Thanks.
Another question on how compression works: How do you compress a song that has quiet parts and loud parts (which is often the case)?
-When I try to compress the whole song according to the loud part: the quiet part doesn't get much compression.
-When I try to compress the whole song according to the quiet part: the loud part clips.
-If I compress the parts separately, the sound texture of each part becomes too different from each other.

Use a maximizer for that. It will try to push up everything to certain level - quite parts will get big push, loud parts less.
But think again about before you add a maximizer to your master channel.
The artist that made the track probably had some idea in mind when making loud and quite parts on a song. There are ppl out there that actually like to have some dynamic on a track, not all squished into a 3dB range near clipping. It will just sound horrible if you i.e. try to push the volume on a break (quite part) to the volume of the bass drop (loud part).

[neueliteratur]

Thanks.
Another question on how compression works: How do you compress a song that has quiet parts and loud parts (which is often the case)?
-When I try to compress the whole song according to the loud part: the quiet part doesn't get much compression.
-When I try to compress the whole song according to the quiet part: the loud part clips.
-If I compress the parts separately, the sound texture of each part becomes too different from each other.

Use a maximizer for that. It will try to push up everything to certain level - quite parts will get big push, loud parts less.
But think again about before you add a maximizer to your master channel.
The artist that made the track probably had some idea in mind when making loud and quite parts on a song. There are ppl out there that actually like to have some dynamic on a track, not all squished into a 3dB range near clipping. It will just sound horrible if you i.e. try to push the volume on a break (quite part) to the volume of the bass drop (loud part).

Any recommendations for free maximizers for Logic?

[PurpleSunray]

Voxengo Elephant has pretty good reputation, but it is too expensive for me since I rarely use maximer effects at all.. using LoudMax if I need a 'neutral' push, or TBT TLs maximizer if it may get some texture (both free)

[manducator]

I don't know if this works in Logic (no osc user), but I guess this is worth a look:

http://www.yohng.com/software/w1limit.html

[manducator]

Thanks.

Another question on how compression works: How do you compress a song that has quiet parts and loud parts (which is often the case)?

-When I try to compress the whole song according to the loud part: the quiet part doesn't get much compression.
-When I try to compress the whole song according to the quiet part: the loud part clips.
-If I compress the parts separately, the sound texture of each part becomes too different from each other.

Do you really want to compress the quiet parts? If yes, I guess you can't do it on the wav file of the whole song but you have to go back to the mix, make a buss of all quiet parts and put compression on that buss.

[neueliteratur]

Cool, I'll try that. Thanks.

[Hermetech Mastering]

-When I try to compress the whole song according to the loud part: the quiet part doesn't get much compression.

This is a problem?