Multiband Compression - Attack & Release times
-
- KVRist
- Topic Starter
- 106 posts since 22 Mar, 2011
Hi guys,
After many years of ignoring it, I've finally got curious about about multiband compression and decided to see what I might be missing out on. I've just downloaded the demo of FabFilter Pro-MB.
I really haven't put stacks of time into this yet and I know this will be very dependent on the material in question, but I was keen to see if my initial guess that (in general terms), the attack and release times that work best will gradually shorten as you go up the frequency bands?
The goal being to even out a mix that little extra bit when mastering. I'm doing my own mastering at the moment, mainly because it's given me the focus to really up my game with mixing and mastering skills, so this is the next area I want to get tooled up on
After many years of ignoring it, I've finally got curious about about multiband compression and decided to see what I might be missing out on. I've just downloaded the demo of FabFilter Pro-MB.
I really haven't put stacks of time into this yet and I know this will be very dependent on the material in question, but I was keen to see if my initial guess that (in general terms), the attack and release times that work best will gradually shorten as you go up the frequency bands?
The goal being to even out a mix that little extra bit when mastering. I'm doing my own mastering at the moment, mainly because it's given me the focus to really up my game with mixing and mastering skills, so this is the next area I want to get tooled up on
-
- KVRAF
- 1888 posts since 13 Aug, 2011 from Berlin
Generally yes. Shortening the times you will find the character of that band changing. I suggest to start with the bass band to get a feeling for the times that are giving you the desired results. Distortion is easily heard there when times are too short.
-
- KVRist
- Topic Starter
- 106 posts since 22 Mar, 2011
Yeah sounds like a good plan.
The whole reason I started looking into multiband was because I find my kicks get way too easily distorted when it comes to the final limiting. This way I can get that loud and punchy sound without having to destroy my hard crafted kick (house music).
It's also made me be a little more gentle with that kick in the mix to begin with.
Please do pass over you multiband compression tips for all types of sounds and frequency ranges!
Cheers
The whole reason I started looking into multiband was because I find my kicks get way too easily distorted when it comes to the final limiting. This way I can get that loud and punchy sound without having to destroy my hard crafted kick (house music).
It's also made me be a little more gentle with that kick in the mix to begin with.
Please do pass over you multiband compression tips for all types of sounds and frequency ranges!
Cheers
Last edited by stevemac on Fri Oct 31, 2014 9:00 am, edited 1 time in total.
-
- KVRian
- 750 posts since 30 Aug, 2011 from somewhere in universe
For mixing it's generally better to keep attack/release times close to each other. Otherwise you might mess up your transients. It's more important for attack and a bit less so for release. Just make sure when tweaking specific bands you pay close attention to transients of the sounds with a body below that band.
Wonder whether my advice worth a penny? Check my music at Soundcloud and decide for yourself.
re:vibe and Loki Fuego @ Soundcloud
re:vibe and Loki Fuego @ Soundcloud
- KVRAF
- 1821 posts since 26 Nov, 2005 from Where silence and chaos meet.
a general compression tip: the way the brain works means that we need to hear a few cycles in a transient to get a feel of it. therefore, the higher the frequency, the higher the amount of cycles you would hear per transcient if you set your compressor at specific attack.
somehow, to have a more homogenic feel in the transients using a multiband, leave a reasonable amount of attack in the bass area, shorter it in the mid, and shorter it even more in the tops.
if you are used to compress individual drums ''properly'' (i understand this is subjective). it comes back to the kick having like more or less 30 to 50 ms, the snare 25ms or so, and 5 to 10 ms for the hats. enough to mould a descent attack.
somehow, to have a more homogenic feel in the transients using a multiband, leave a reasonable amount of attack in the bass area, shorter it in the mid, and shorter it even more in the tops.
if you are used to compress individual drums ''properly'' (i understand this is subjective). it comes back to the kick having like more or less 30 to 50 ms, the snare 25ms or so, and 5 to 10 ms for the hats. enough to mould a descent attack.
It's not what you use, it's how you use it...
-
- KVRist
- Topic Starter
- 106 posts since 22 Mar, 2011
Good post!Chopper wrote:a general compression tip: the way the brain works means that we need to hear a few cycles in a transient to get a feel of it. therefore, the higher the frequency, the higher the amount of cycles you would hear per transcient if you set your compressor at specific attack.
somehow, to have a more homogenic feel in the transients using a multiband, leave a reasonable amount of attack in the bass area, shorter it in the mid, and shorter it even more in the tops.
if you are used to compress individual drums ''properly'' (i understand this is subjective). it comes back to the kick having like more or less 30 to 50 ms, the snare 25ms or so, and 5 to 10 ms for the hats. enough to mould a descent attack.
-
- KVRAF
- 6741 posts since 25 Mar, 2002 from sheffield, england
The numbers above are useless. There is no concensus as to how attack and release times should be defined and measured, so unless you are referring to a specific compressor attack and release time in ms are meaningless.
Pro-MB also has extremely program dependent attack and release times, and also takes account of the frequency band being processed, hence the use of percentage values not times in ms.
So forget everything posted so far and use your ears. If you want a punchy kick be careful not to set too fast an attack time: 30 or 40% should be a good starting point. And use the release to control the tail of the kick, and glueing effects with bass parts (assuming you're processing the full mix).
You probably don't have to worry about fast release times distorting: Pro-MB is too clever for that. But fast release times might spoil the "shape" of the low end and make it seem congested. Experiment with settings between about 30% (fast but not too fast) or 70% (bouncing with the beat).
Pro-MB also has extremely program dependent attack and release times, and also takes account of the frequency band being processed, hence the use of percentage values not times in ms.
So forget everything posted so far and use your ears. If you want a punchy kick be careful not to set too fast an attack time: 30 or 40% should be a good starting point. And use the release to control the tail of the kick, and glueing effects with bass parts (assuming you're processing the full mix).
You probably don't have to worry about fast release times distorting: Pro-MB is too clever for that. But fast release times might spoil the "shape" of the low end and make it seem congested. Experiment with settings between about 30% (fast but not too fast) or 70% (bouncing with the beat).
-
- KVRAF
- 6741 posts since 25 Mar, 2002 from sheffield, england
Pro-MB kind of does that for you. The attack and release times are made to be as easy and intuitive as possible. Whatever frequency band you are processing, use values below about 50% to shape faster transients, and values above 50% for slower dynamic changes. And always use your ears!stevemac wrote:I was keen to see if my initial guess that (in general terms), the attack and release times that work best will gradually shorten as you go up the frequency bands?
-
- KVRist
- 201 posts since 14 May, 2008
I love Pro-MB as a subtle mixbuss compressor or to glue complex harmonies.
However, just like you, I have a hard time distinguishing what the attack and release times are exactly doing to the sound...but it does a lot of stuff to the signal, kind of automatically.
It makes everything sound modern and crisp. That's why every time I use it I end up using as a parallel compressor, with the mix knob around 90/95%...I little goes a long way with that mix knob.
Make sure to take advantage of its mid/side component. It's great for making sure low frequencies are mono and to do a little stereo expansion in the higher frequencies.
However, just like you, I have a hard time distinguishing what the attack and release times are exactly doing to the sound...but it does a lot of stuff to the signal, kind of automatically.
It makes everything sound modern and crisp. That's why every time I use it I end up using as a parallel compressor, with the mix knob around 90/95%...I little goes a long way with that mix knob.
Make sure to take advantage of its mid/side component. It's great for making sure low frequencies are mono and to do a little stereo expansion in the higher frequencies.
-
- KVRist
- Topic Starter
- 106 posts since 22 Mar, 2011
Good to know, I must admit that my first use of it last night was on the basis that the controls were 'normal' and not programme dependent. The attack and release controls being % rather than time values should have given that way really Just reading the manual now...IIRs wrote:Pro-MB kind of does that for you. The attack and release times are made to be as easy and intuitive as possible. Whatever frequency band you are processing, use values below about 50% to shape faster transients, and values above 50% for slower dynamic changes. And always use your ears!stevemac wrote:I was keen to see if my initial guess that (in general terms), the attack and release times that work best will gradually shorten as you go up the frequency bands?
I have Saturn (which, amongst other tasks, is a permanent resident of my mix buss - on 15% warm tape as my mix glue), and I have to say that FabFilter really do have their shit together.
-
- KVRAF
- 6741 posts since 25 Mar, 2002 from sheffield, england
FabFilter Pro MB - Downward compression: http://youtu.be/O1Qz_libwBk
-
- KVRAF
- 7383 posts since 17 Feb, 2005
This is the only thing that's been putting me off of Pro-MB. The use of percentages raises more questions than it answers. Are you saying it also changes the timing if you shift the band?IIRs wrote: Pro-MB also has extremely program dependent attack and release times, and also takes account of the frequency band being processed, hence the use of percentage values not times in ms.
-
- KVRAF
- 6741 posts since 25 Mar, 2002 from sheffield, england
Potentially yes, but in practise it's not really noticeable. After all, a significant change to the band frequencies will change the dynamics of the input signal and hence the compression will change anyway. (Unless your Sidechain filters are set to free and don't change. Probably not a common scenario.)camsr wrote: Are you saying it also changes the timing if you shift the band?
As far as I can tell they just use the info to prevent the compression from ever reacting to single cycles of the wave and hence distorting. That's just speculation though.
You should certainly not be put off by the lack of ms values. These are meaningless in all other compressors anyway! The only way to accurately measure a compressor's gain modulation rate is in dB/s or dB/ms, and even that only works with linear designs (ie: almost none of them!)
The attack and release parameters in Pro-MB are well judged and work very intuitively: that's all you need to worry about
-
- KVRAF
- 7383 posts since 17 Feb, 2005
Maybe, but here is the confusing part...
Let's use the example I set a band to downward compress above 10khz. I set the release all the way to 100% (let's say the actual time doesn't matter). Now I may like how this works rhythmically, but I think it's better to move the band lower to 5khz. Would the release be twice as long? That would be intuitive to me. Maybe there is some kind of envelope shaping to retain the complete release time, but minimize distortion? If there is, is it intuitive?
Let's use the example I set a band to downward compress above 10khz. I set the release all the way to 100% (let's say the actual time doesn't matter). Now I may like how this works rhythmically, but I think it's better to move the band lower to 5khz. Would the release be twice as long? That would be intuitive to me. Maybe there is some kind of envelope shaping to retain the complete release time, but minimize distortion? If there is, is it intuitive?
-
- KVRAF
- 6741 posts since 25 Mar, 2002 from sheffield, england
No it's more complex and subtle than that. I suggest you try the demo...