Dealing With Volume Spikes In The Recorded Waveform

  • Thread starter Thread starter Dr. Varney
  • Start date Start date
philbagg said:
Varney, just curious, did you really give my long post a good read-through?
Click to expand...

Not at first, but I read technical stuff in passes and I never expect to thoroughly understand something in one pass. I've gone back to it a few times and will definitely keep referring to it. There's a lot there.

Some of the comments which came after your question though have also been helpful and make me want to go back and read again, with them in mind, the parts about attack & release.

Dr. V
 
Dr. Varney said:
Not at first, but I read technical stuff in passes and I never expect to thoroughly understand something in one pass. I've gone back to it a few times and will definitely keep referring to it. There's a lot there.

Some of the comments which came after your question though have also been helpful and make me want to go back and read again, with them in mind, the parts about attack & release.

Dr. V
Click to expand...

Coolio :) I know it's a long post, but it's a complicated topic.

Let us know how you get on dude.
 
Question: Assuming (and taking into account) different models of compressor may differ... does leaving the attack and release dials down (at zero?) make for 'short' or does it mean they are 'off', making for no effect?

Dr. V
 
Dr. Varney said:
Question: Assuming (and taking into account) different models of compressor may differ... does leaving the attack and release dials down (at zero?) make for 'short' or does it mean they are 'off', making for no effect?

Dr. V
Click to expand...

You can never turn a compressors attack or release "off", otherwise it just wouldn't work. They are essential to how the unit works. Some compressors have shorter/longer attack and release times than others, some of the knobs even go the opposite way.

On most compressors, turning the knob anti-clockwise means a shorter attack/release time, whereas the likes of the Urei 1176 goes clockwise for shorter a/r times.

Also, there's no such thing as zero for attack and release times. The knob will simply go from the shortest attack/release times, to the longest - I've given examples of shortest/longest attack and release times in the long post :)
 
Dr. Varney said:
Question: Assuming (and taking into account) different models of compressor may differ... does leaving the attack and release dials down (at zero?) make for 'short' or does it mean they are 'off', making for no effect?

Dr. V
Click to expand...

Lower attack and release times mean faster action, not less action.
 
Joza said:
Actually very very long attack and release mean no compression at all.
Click to expand...

Depends on the material.

Besides, most compressors only go as far as 50ms attack. There's very little that wouldn't be affected by that.
 
Well.. depends. I guess you are more right.

But if attack is say 1000000s then it wont ever start compressing, and the release doesn't really matter.

And if the release is very very long, it doesnt really compress it. It does lower in volume when the threshold is exceeded but after that it stays constant instead of returning back. I wouldnt really call that compression, rather some strange limiting.

And also if Im not wrong, to have extremely compressed audio (as loud as possible), you need both fast attack and fast release (with low threshold and high ratio of course), so mine was kind of a "counter-statement" to that.
 
Joza said:
Well.. depends. I guess you are more right.

But if attack is say 1000000s then it wont ever start compressing, and the release doesn't really matter.

And if the release is very very long, it doesnt really compress it. It does lower in volume when the threshold is exceeded but after that it stays constant instead of returning back. I wouldnt really call that compression, rather some strange limiting.

And also if Im not wrong, to have extremely compressed audio (as loud as possible), you need both fast attack and fast release (with low threshold and high ratio of course), so mine was kind of a "counter-statement" to that.
Click to expand...

Well, the whole concept of 10000000s attack time is never going to happen anyway :p

And, yes, if the release is extremely long (provided that the attack time is normal and the threshold and ratio are set normally), then it's not proper compression. But, we were referring to a release time in conjunction with a really really long attack time.

...i'm confusing myself... but trust me, it makes sense... :p
 
And where can I get me a compressor with an attack time of 11.5 days (1000000s) do API or Neve make one? Maybe SSL. :laughings:

I mean really, yes theoretically I suppose it's possible but why discuss a compression technique that would result in no compression on a compressor setting that does not exist on any compressor (hardawre or VST) in existence

How does that help anyone tame an overly "spikey" vocal?
 
philbagg said:
I know it's a long post
Click to expand...
:rolleyes: Phil, buddy, you have no idea what long is. ;) :D

I know nobody wants to read anything anymore, but for anyone interested in the subject of introductory compression in the kind of detail the subject needs yet explained in easy-to-understand terms, try hitting this page and clicking on "Compression Uncompressed".

G.
 
Bristol Posse said:
And where can I get me a compressor with an attack time of 11.5 days (1000000s) do API or Neve make one? Maybe SSL. :laughings:

I mean really, yes theoretically I suppose it's possible but why discuss a compression technique that would result in no compression on a compressor setting that does not exist on any compressor (hardawre or VST) in existence

How does that help anyone tame an overly "spikey" vocal?
Click to expand...

Because understanding in theory how something that doesn't exist and could never work helps us to understand in practice things that do exist and work...? :rolleyes:

Oh, I dunno... :laughings:

SouthSIDE Glen said:
I know nobody wants to read anything anymore, but for anyone interested in the subject of introductory compression in the kind of detail the subject needs yet explained in easy-to-understand terms, try hitting this page and clicking on "Compression Uncompressed".

G.
Click to expand...

I'm readin'... :D

Dr. V
 
Alright... so...

My next question is: Would you create a volume envelope for each wave sample that needs taming and save the sample again - or would you wait until all the samples are together in the sequencing project - and apply the automation on the finished thing... like, at the mastering stage?

Dr. V
 
Dr. Varney.

I think you're getting way to microscopic here. I wonder if you'd try to get this level had you been working with tape and an analog console ;)

In the end, by trying to "tame" each transient into some homogenic level that looks nice and orderly on the screen is likely to do you more harm than anything.

IMHO the best way to learn the relationship between attack and release times is to do the following:

Set the compressor's controls the following way:
Fastest attack
Fastest release
Lowest threshold
Leave makeup gain alone

Now play your audio through it. It will probably sound obliterated :D

Now, leave everything the same and slowly start turning the attack dial towards the slow side. Notice how the transients are starting to poke through?

Put attack back to its fastest setting and now slowly turn release towards the slow side. Notice how the material is kinda starting to get its "shape" back, albeit at much lower volume?

Now, look at the gain reduction meter in your compressor (which probably is reducing gain by some crazy amounts, maybe something like -40 to -50db or so with the current settings). Start moving the release back to the faster side to the point where the gain reduction returns to 0db (or close to it) before the next transient hits.

Adjust, listen, adjust, listen. Close the waveform window and don't bother with it. Adjust, listen, adjust, listen, until your ears get used to the relationships between attack and release settings.

Once you start understanding what these do, increase the threshold back to sane levels.
 
noisewreck said:
Dr. Varney.

I think you're getting way to microscopic here. I wonder if you'd try to get this level had you been working with tape and an analog console ;)

In the end, by trying to "tame" each transient into some homogenic level that looks nice and orderly on the screen is likely to do you more harm than anything.
Click to expand...

Fair assumption but I don't think I am (although saying it is good advice). If I were working with analogue tape, I'm not sure what I'd do. I suppose I'd have no choice but to apply compression afterwards - though, isn't digital editing supposed to give us more flexibility and control over the recording afterwards?

Approaching the task digitally allows choices which, to you, would appear as though I'm trying to "Tame each transient into some homogenic level..." and that's not really what I'm trying to achieve. Rises and falls in speech volume give this piece it's expression. I wouldn't want to destroy that by homogenizing the volume levels.

I can assure you, it's not my first priority to get everything looking nice and orderly on the screen (again, good advice). I'm trying to correct how it sounds, not the way it looks. Some of the spikes spoil an otherwise nice speech and it so happens, that when I look at them, I can see it in the waveform. Isn't that one of the given bonuses with digital editing? I suppose not everyone would agree. I guess it can easily distract someone new to sound editing.

The reasons I can think of that anyone would make these assumptions is because I opened my post with a visual illustration of how the waveform appears to my ears and complained that my waveforn didn't look as I expected it to. I'm aware that the viz gives me only a rough snapshot of the sonic reality but, short of making MP3s and getting you to listen, it seemed the easiest way to illustrate my problem.

Don't get me wrong... I'm not offended or anything. :) I just want to set the record straight, that I'm not trying to edit my sounds on visual information alone, or take the life and breath out of my actors' performance, just because technology allows.

Actually, I think I would benefit greatly from owning an analogue reel tape deck. I agree it might offer good extra training for the ears, but for the other hassles it would bring to this process.

Now I am beginning to think I'm trying to correct volume too early in the process. Maybe you are right about me getting too 'microscopic' at this stage. I've always felt as though most people in here do things more efficiently than I do - making less fuss about little things, to start with - then polishing things up afterwards. I'd like to gain the kind of confidence I see around me in here. That's why I come here.

That aside, I'm reading your instruction on compression use. I find that really helpful, thanks mate! :)

Regards

Dr. V
 
Last edited:
Doc, any chance you could post the clip? Or even a problematic part of it?
 
Oh, and you don't need analog tape to learn to use your ears, just try to stop focusing on the visual aspect altogether :)

Visuals, when mixing, shouldn't be used to judge a sound. They are useful when you can hear that something's wrong, and need to use a visual representation (such as waveform or frequency analyser) to help find the problem.
 
So look away from the screen and use your ears. When you hear a wave that sounds wrong, then fix it with the techniques given here...again, without looking.[/QUOTE]

Great advice here!

The ears don't lie. Compression can help fix the problem, but may soften the signal and ruin the tone leading to EQ's to fix it.
I would suggest using a limiter. The will only affect the volume and not the tone.

Good luck.
 
You must log in or register to reply here.
Back
Top