Quantcast
compression in the digital domain - Gearspace.com
The No.1 Website for Pro Audio
compression in the digital domain
Old 4th May 2003
  #1
Gear Addict
 
Marshall Simmons's Avatar
 
🎧 15 years
compression in the digital domain

I've been thinking quite a bit about compression, mostly because with the current tools that i have, i have not been able to get a firm grasp on finer techniques of compression.

I'm in a DSP class using max/msp and A thought came to my head.


All compression is in its most basic form would be a condition statement:


If x(signal input) is greater then T(threshold) then divide x by R( compression ration)

In a situation like this, it would be an extremely simplistic compressor whithout the need for attack and release.

Sonically, it seems like it would avoid any attack and realease artifiacts, but perhaps i'm missing something.


Also, multiband compression would be relatively simple as well.
you could do it in the frequency domain instead of the time domain and have individual control over specific bins of the FFT, although this may pose a sonic problem during the resynthesis of the audio.


Any thoughts?
Marshall
Old 4th May 2003
  #2
Gear Maniac
 
Chae Ham's Avatar
 
🎧 15 years
The problem is that this could only exist in theory. Everything as we know it exists along a time continuum. In order for something to defy time (in your example, eliminating an attack or release time) it would have to be infinite. This is basically what Einstein's theory of relativity suggests: "That Energy in a mass equals a mass multiplied by the velocity of light squared."

If there isn't an attack time, at what point would it "kick in"? Never? Always(infinite)? Or at the speed of light(which would also suggest being infinite)? The same would apply to the release.

Sorry if I went off on a tangent there its the physics schooling kicking in...
Old 4th May 2003
  #3
Gear Addict
 
Marshall Simmons's Avatar
 
🎧 15 years
just like any other plugin, i would have to introduce a bit of latency to process.... but considering that this is simple multiplication it should only be in the order of 5-10 samples.

Marshall
Old 4th May 2003
  #4
Gear Maniac
 
Chae Ham's Avatar
 
🎧 15 years
Maybe I'm misunderstanding you, but are you saying that the attack time (and release) for the compressor to start working would be 5-10samples? If that is the case then, it would smooth out the transient because of such a fast attack, and possibly give a "pumping" effect because of the fast release followed by such a fast attack.

Although a compressor is a "dynamic processor," the nature of the way it works relies on some time variables. You can't eliminate the attack or release, because it has to "start" at some point and "end" at some point. You CAN set a constant unchangeble attack and release time, but that is just assigning the variables not eliminating them.
Old 4th May 2003
  #5
Gear Addict
 
Marshall Simmons's Avatar
 
🎧 15 years
ok... you are thinking in the standard paradigm of compressors...


think of it as a simple math problem instead.


think of looking at the amplitude of individual samples. if the sample X is above a specific number T (threshold) the divide the level by r (your ratio) and add T
so the equasion would be

x(n) > T
then (x(n)/R)+T


there would be no pumping because it is a product of a release control which isn't needed.... in the analogue world it it, but not digital

I hope that makes sense
marsh
Old 4th May 2003
  #6
Lives for gear
 
DanV's Avatar
 
🎧 15 years
marshall you are totally right, what you are talking about my friend is a limiter of sorts. compressors use dynamics like attack and decay to add a snappy or kicky quality. limiters skip that shet and basically are a no frills compressor with a more sophisticated gain management.
Old 4th May 2003
  #7
Gear Addict
 
Marshall Simmons's Avatar
 
🎧 15 years
Thanks!!!!

That helps a ton actually.


i understand that most people use real compressors for shaping and giving a feel to a track... but alot of times i just want to control the dynamics and nothing else

Marshall
Old 4th May 2003
  #8
Gear Maniac
 
Chae Ham's Avatar
 
🎧 15 years
That's what I was thinking, but I'm guessing that he still means a "compressor" In which case he is overlooking a few things.

...But with the right questions, comes great insight.

(At one point back in school I decided that I could design a broadband noise reduction algorithm based on the idea of frequency dependent compression... Long story short, I was quickly humbled as to why my original idea wouldn't work... but I'm stil working on it or other things of sorts...)

Good Luck Marshall!
Old 4th May 2003
  #9
Gear Addict
 
Marshall Simmons's Avatar
 
🎧 15 years
thanks!


i know i could write it in csound or max/msp

but i'd like to figure out how to do it in something that i could make a vst plugin.

marsh
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 164 views: 75505
Avatar for burns46824
burns46824 28th October 2015
replies: 28 views: 23118
Avatar for kilon
kilon 1st July 2010
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump