Propellerhead Software
  #1  
Old 2003-01-21, 03:37
gnorpf gnorpf is offline
 
Join Date: Aug 2002
Posts: 1
About compression

Hello everybody! I took a look at SuperSweets Reason FAQ today. Great page!

It looks like there are still a few open questions about compression, so I decided to make another experiment and post my findings here. What came out in the end is an extremely simple explanation. First of all, I converted the experiment into a .rps that you can download here:

http://dynamo.propellerheads.se/song...ng&songID=7129

Here's a primer how you can grasp compression very quickly: Play the rps and look at the LEDs on the first Mixer at Track 1 and you'll see that the sound gets 1 LED louder every second. Then turn the compressor from "Bypass" to "On" and set the following settings:
Ratio: 127 (16:1, the maximum)
Thresh: 34
Attack: 0
Release: 0
Before you play the song again, remember that the signal that goes into the compressor is still the same, but what comes out now is the processed signal. Play it. The fact that the outcoming signal is now louder is not important. The really special thing is that as soon as the signal reaches 3 yellow LEDs, it stays there. The input signal gets louder, but the output signal stays about the same. This is compression (D'oh!)

What do the knobs do?
THRESH defines how loud the incoming signal has to be that the compression kicks in. Here's a handy table:

Input1-20 Thresh
01 . . . . . 0
02 . . . . . 0
03 . . . . . 0
04 . . . . . 0
05 . . . . . 8
06 . . . . . 20
07 . . . . . 30
08 . . . . . 40
09 . . . . . 50
10 . . . . . 60
11 . . . . . 60
12 . . . . . 70
13 . . . . . 80
14 . . . . . 90
15 . . . . . 100
16 . . . . . 110
17 . . . . . 120
18 . . . . . 127
19 . . . . . 127
20 . . . . . 127

RATIO: You may have noticed in our experiment that the volume stays at 3 yellow LEDs for some time, but then still gets louder. This is because the Reason compressor doesn't limit the volume absolutely. What it does is rather: After the threshhold point, an additional increase in input volume result in a very small increase in output volume.

With the RATIO knob, you can decide how steeply this increase happens. The increase in our example is the flattest setting (16:1). The steepest setting on the other side is when an certain increase in input volume result in the same increase in output volume (1:1), which is the same thing as using no compression at all.


So what we experience as "compression" is the fact that louder parts of your music are turned down and closer to the soft parts, resulting in a more uniform loudness.

For all my experiments, I've set the parameters "ATTACK" and "RELEASE" to the minimum, so that the changes in amplification and compression happen immediately. This can be changed for certain effects. Let's say that any change in the input can only be of two forms:
1 - the input gets louder
2 - the input gets softer
ATTACK and RELEASE can control how the device reacts in these two cases. ATTACK controls the case of "input getting louder", and RELEASE controls the case of "input getting softer"

ATTACK: When you have a signal of average loudness and suddenly it gets louder, the device would normaly just cut back the signal according to your settings. But maybe you want to preserve a bit of the "punch" in the sudden increase of volume. You want that the signal to be louder for a short moment before it gets back to your specified settings. With the ATTACK parameter, you can decide how long the old compression of the softer part will still be applied until it gets to the new setting. The range of this knob is between 0 and about 0.3 seconds. Most common usage: preserve the "punch" of drums.

RELEASE: After the loud signal everything goes back to average. You want the listener to hear the decrease in loudness, so the old setting for the loud part should still apply for a short time, making the softer part even softer. The range for this know is between 0 and about 3 seconds.

I hope this is of use for some people
Cheers!
Gnorpf
Compressor experiment.rps
  #2  
Old 2003-01-21, 09:39
thral thral is offline
 
Join Date: Jan 2000
Posts: 0
a few corrections... okay a lot of corrections...

:Here's a primer how you can grasp compression very quickly: Play the rps and look at the LEDs on the first Mixer at Track 1 and you'll see that the sound gets 1 LED louder every second. Then turn the compressor from "Bypass" to "On" and set the following settings:
:Ratio: 127 (16:1, the maximum)
:Thresh: 34
:Attack: 0
:Release: 0
:Before you play the song again, remember that the signal that goes into the compressor is still the same, but what comes out now is the processed signal. Play it. The fact that the outcoming signal is now louder is not important. The really special thing is that as soon as the signal reaches 3 yellow LEDs, it stays there. The input signal gets louder, but the output signal stays about the same. This is compression (D'oh!)

-- If the input gets louder but the output stays the same, technically that's limiting. In fact, a compressor with a working ratio of greater than 10:1 is considered a limiter.

:What do the knobs do?
:THRESH defines how loud the incoming signal has to be that the compression kicks in. Here's a handy table:
:
:Input1-20 Thresh
:01 . . . . . 0
:02 . . . . . 0
:03 . . . . . 0
:04 . . . . . 0
:05 . . . . . 8
:06 . . . . . 20
:07 . . . . . 30
:08 . . . . . 40
:09 . . . . . 50
:10 . . . . . 60
:11 . . . . . 60
:12 . . . . . 70
:13 . . . . . 80
:14 . . . . . 90
:15 . . . . . 100
:16 . . . . . 110
:17 . . . . . 120
:18 . . . . . 127
:19 . . . . . 127
:20 . . . . . 127

-- If you could get that into dB it would be nice. :-)

:RATIO: You may have noticed in our experiment that the volume stays at 3 yellow LEDs for some time, but then still gets louder. This is because the Reason compressor doesn't limit the volume absolutely. What it does is rather: After the threshhold point, an additional increase in input volume result in a very small increase in output volume.
:
:With the RATIO knob, you can decide how steeply this increase happens. The increase in our example is the flattest setting (16:1). The steepest setting on the other side is when an certain increase in input volume result in the same increase in output volume (1:1), which is the same thing as using no compression at all.

-- Actually, attack would adjust how steeply the compression kicks in, not ratio... unless you are using a different use of the word steep (???).

:
:So what we experience as "compression" is the fact that louder parts of your music are turned down and closer to the soft parts, resulting in a more uniform loudness.

-- Also, quieter parts are turned up, since the reason compressor has auto-gain.

:For all my experiments, I've set the parameters "ATTACK" and "RELEASE" to the minimum, so that the changes in amplification and compression happen immediately.

-- I do not think it is exactly immediately... never seen a compressor working with an attack of 0ms... 10ms maybe. Maybe it does go to zero, but I don't think so.

This can be changed for certain effects. Let's say that any change in the input can only be of two forms:
:1 - the input gets louder
:2 - the input gets softer
:ATTACK and RELEASE can control how the device reacts in these two cases. ATTACK controls the case of "input getting louder", and RELEASE controls the case of "input getting softer"


:ATTACK: When you have a signal of average loudness and suddenly it gets louder, the device would normaly just cut back the signal according to your settings. But maybe you want to preserve a bit of the "punch" in the sudden increase of volume. You want that the signal to be louder for a short moment before it gets back to your specified settings. With the ATTACK parameter, you can decide how long the old compression of the softer part will still be applied until it gets to the new setting. The range of this knob is between 0 and about 0.3 seconds. Most common usage: preserve the "punch" of drums.

-- Also very commonly used on basslines.

:
:RELEASE: After the loud signal everything goes back to average. You want the listener to hear the decrease in loudness, so the old setting for the loud part should still apply for a short time, making the softer part even softer.

-- Actually, the idea with compression is that you do not want the listener to hear any major decreases or increases in volume. Setting the attack and release should help make the compression less evident, not more.

Eh... I donno.

- Springs
  #3  
Old 2003-01-21, 16:52
robinsarac robinsarac is offline
 
Join Date: Jan 2001
Posts: 16
Hmmm...

:-- Actually, attack would adjust how steeply the
:compression kicks in, not ratio... unless you are
:using a different use of the word steep (???).

I don't know that you're correct here. Attack determines *when* the signal is attenuated, and ratio determines by how much, no? A ratio of 1:1 is 1 dB of gain reduction for every 1 dB over the threshold, all the way up to 16:1 where the signal is decreased 16 dB for every 1 dB over the threshold. Attack tells the compressor when to kick in after the threshold is exceeded.

:-- Setting the attack and release should
:help make the compression less evident,
:not more.

That is entirely dependant on the producer concerned. Plenty of people use a compressor to purposely introduce a bit of pumping.
  #4  
Old 2003-01-22, 04:41
thral thral is offline
 
Join Date: Jan 2000
Posts: 0
Re: Hmmm...

::-- Actually, attack would adjust how steeply the
::compression kicks in, not ratio... unless you are
::using a different use of the word steep (???).
:
:I don't know that you're correct here. Attack determines *when* the signal is attenuated, and ratio determines by how much, no? A ratio of 1:1 is 1 dB of gain reduction for every 1 dB over the threshold, all the way up to 16:1 where the signal is decreased 16 dB for every 1 dB over the threshold. Attack tells the compressor when to kick in after the threshold is exceeded.

No.. attack controlls the steepness of the attack curve after it is triggered. And it's not decreased 16dB for every one. If it was, if it went 6dB over the threshold, it would silence the signal... and it doesn't.

:
::-- Setting the attack and release should
::help make the compression less evident,
::not more.
:
:That is entirely dependant on the producer concerned. Plenty of people use a compressor to purposely introduce a bit of pumping.
:
  #5  
Old 2003-01-21, 16:41
EnochLight's Avatar
EnochLight EnochLight is offline
 
Join Date: May 2001
Posts: 5,916
Thanks gnorpf, nice work (n/t)

:Hello everybody! I took a look at SuperSweets Reason FAQ today. Great page!
:
:It looks like there are still a few open questions about compression, so I decided to make another experiment and post my findings here. What came out in the end is an extremely simple explanation. First of all, I converted the experiment into a .rps that you can download here:
:
:http://dynamo.propellerheads.se/song...ng&songID=7129
:
:Here's a primer how you can grasp compression very quickly: Play the rps and look at the LEDs on the first Mixer at Track 1 and you'll see that the sound gets 1 LED louder every second. Then turn the compressor from "Bypass" to "On" and set the following settings:
:Ratio: 127 (16:1, the maximum)
:Thresh: 34
:Attack: 0
:Release: 0
:Before you play the song again, remember that the signal that goes into the compressor is still the same, but what comes out now is the processed signal. Play it. The fact that the outcoming signal is now louder is not important. The really special thing is that as soon as the signal reaches 3 yellow LEDs, it stays there. The input signal gets louder, but the output signal stays about the same. This is compression (D'oh!)
:
:What do the knobs do?
:THRESH defines how loud the incoming signal has to be that the compression kicks in. Here's a handy table:
:
:Input1-20 Thresh
:01 . . . . . 0
:02 . . . . . 0
:03 . . . . . 0
:04 . . . . . 0
:05 . . . . . 8
:06 . . . . . 20
:07 . . . . . 30
:08 . . . . . 40
:09 . . . . . 50
:10 . . . . . 60
:11 . . . . . 60
:12 . . . . . 70
:13 . . . . . 80
:14 . . . . . 90
:15 . . . . . 100
:16 . . . . . 110
:17 . . . . . 120
:18 . . . . . 127
:19 . . . . . 127
:20 . . . . . 127
:
:RATIO: You may have noticed in our experiment that the volume stays at 3 yellow LEDs for some time, but then still gets louder. This is because the Reason compressor doesn't limit the volume absolutely. What it does is rather: After the threshhold point, an additional increase in input volume result in a very small increase in output volume.
:
:With the RATIO knob, you can decide how steeply this increase happens. The increase in our example is the flattest setting (16:1). The steepest setting on the other side is when an certain increase in input volume result in the same increase in output volume (1:1), which is the same thing as using no compression at all.
:
:
:So what we experience as "compression" is the fact that louder parts of your music are turned down and closer to the soft parts, resulting in a more uniform loudness.
:
:For all my experiments, I've set the parameters "ATTACK" and "RELEASE" to the minimum, so that the changes in amplification and compression happen immediately. This can be changed for certain effects. Let's say that any change in the input can only be of two forms:
:1 - the input gets louder
:2 - the input gets softer
:ATTACK and RELEASE can control how the device reacts in these two cases. ATTACK controls the case of "input getting louder", and RELEASE controls the case of "input getting softer"
:
:ATTACK: When you have a signal of average loudness and suddenly it gets louder, the device would normaly just cut back the signal according to your settings. But maybe you want to preserve a bit of the "punch" in the sudden increase of volume. You want that the signal to be louder for a short moment before it gets back to your specified settings. With the ATTACK parameter, you can decide how long the old compression of the softer part will still be applied until it gets to the new setting. The range of this knob is between 0 and about 0.3 seconds. Most common usage: preserve the "punch" of drums.
:
:RELEASE: After the loud signal everything goes back to average. You want the listener to hear the decrease in loudness, so the old setting for the loud part should still apply for a short time, making the softer part even softer. The range for this know is between 0 and about 3 seconds.
:
:I hope this is of use for some people
:Cheers!
:Gnorpf
  #6  
Old 2003-01-22, 04:57
alientom alientom is offline
 
Join Date: Oct 2002
Posts: 0
What would be really killer is a limiter n/t

 

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT +2. The time now is 06:08.