Odim's Audio Engineering Corner

[odims_sphere]

From time to time, when I get bored I'm going to post different audio engingeering snippets in here both to help anyone out who doesn't know much about audio engineering and also to keep my own skills as sharp as possible. Most of what I'm going to put in here is for beginners though and by no means is anything I put here the only way of doing something, it's just what I've picked up. I also encourage anyone else to put your own tips and tricks in as well, for example as most of you know I have very little experiance with software. Everything I use is hardware based with the exception of my sequencer.

[glc]

Yay! I'll be here heaps. I love engineering.

[phee]

Dear recording guru...

Can you explain quantizing?

thanks

[glc]

as in signal processing, as opposed to midi I assume?

Isn't it the conversion of a signal into a bit stream, and then the mathematical process of filling in the gaps left by binary with something more meaningful?? It's been a while since I looked at the theory side of things...

[phee]

as in signal processing, as opposed to midi I assume?

Isn't it the conversion of a signal into a bit stream, and then the mathematical process of filling in the gaps left by binary with something more meaningful?? It's been a while since I looked at the theory side of things...

*ding ding ding*

[glc]

did I win something? or are you just bashing me over the head?

[phee]

did I win something? or are you just bashing me over the head?

You won a back shaving for just a nickle

[glc]

sweet! does that include my ass??

[glc]

Sorry Odims, back onto this most important topc................

[odims_sphere]

Compression on Kick Sounds.

A compressor limits the dynamic range of a signal.

It's used to bring up the quiet parts and roll back the loud parts of music. but there is more to it then that. Simply turning up the gain of a sound like say.. a kick drum would cause it to clip resulting in very a undesireable sound. So the problem is how do I bring the Kick up in my mix when it's already hitting 0 db? I don't want it to clip, but I can't hear it in the mix. The answer is compression.

Comressors work in several stages.

1. INPUT. To set the level in.

2. THRESHOLD. This sets how high the signal must reach before the compressor kicks in.

3. RATIO. This sets how much compression is applied in ratio to the Db rise in signal level above the Threshold.

4. ATTACK. This sets how fast the compressor kicks in once the Threshold has been breached.

5. RELEASE. This sets how fast the compressor lets go, once the input signal has dropped back below the threshold.

6. OUTPUT. Sets the output signal level.

Kick sounds are almost always compressed for the simple fact that the wave form of a kick has a very high inital spike for a very short period of time. By limiting the dynamic range of the spike we can bring up the overall gain of the sound. It's the spike in the wave form which has almost no tonal quality that is taking up all of the dynamic range, leaving the "meat" of the sound so buried in the mix that you cant hear it. Rather than turning down everything else it's best to just squash the spikes.

Every sound is different, but basic compressor theroy on kicks is pretty fast attack, High compression ratio and fast release with make up gain. By the numbers that would mean:

Attack 1-3ms

Release .2 sec

Ratio 5:1 - 10:1

That should make your kicks stand out in the mix.

Keep in mind I am writing for the beginner here, this is just the surface of what compression can do.

[glc]

Something else that's really intersting with compression, is if you use it in reverse. eg. if you are using a daw, reverse the signal you want to process. Then compress it as you'd normally set up your compression. The beauty of this method is REALLY SMOOTH intersting compression. you are setting the attack controls to the tail end of the original signal, working up to the peak, instead of the other way around. Then, apply the compression, and reverse the signal back to normal again.

Now keep in mind that this is just something interesting you can try, of course there are some issues like not being able to change controls during subsequent playback, and other issues, but it's something cool to try.

[odims_sphere]

as in signal processing, as opposed to midi I assume?

Isn't it the conversion of a signal into a bit stream, and then the mathematical process of filling in the gaps left by binary with something more meaningful?? It's been a while since I looked at the theory side of things...

And for the Midi answer, the very quick version: Fixing human error by putting all the key on and key off events on a spacific measure division specified by the programmer such as 1/4, 1/8th, 1/16th notes Etc..

[CixWicked]

And for the Midi answer, the very quick version: Fixing human error by putting all the key on and key off events on a spacific measure division specified by the programmer such as 1/4, 1/8th, 1/16th notes Etc..

Wow, I'm actually understanding this crap... er stuff... STUFF I meant stuff...

* Runs away *

CixWicked

[odims_sphere]

Wow, I'm actually understanding this crap... er stuff... STUFF I meant stuff...

* Runs away *

CixWicked

What would you like to know?

[Scales]

Compression on Kick Sounds.

I'm a software geek. But this is interesting stuff. After that sentence my brain stopped working properly, but I'll come back and read through it sometime till I understand it, lol. The terms were familiar.

[DeadBurgerKing (10)]

I dunno why my comment went here. I went to Artifical and lether, went to add reply then it ended up here...that was WEIRD

[glc]

Here's a, eq tip for you tweakheads...

If you need to eq, try cutting instead of boosting. for instance, if you want more bass, cut the treble etc...

This usually makes a much more appealing and natural sound while also preserving your precious headroom.

[Fierce Critter]

You hit upon this a little bit when conversing with me at CC. But please elaborate, in as simple terms as you can. I'm only using Cubase and Cool Edit Pro for making my songs, so I'm not well versed in the depths of engineering.

This is regarding volumes. I have trouble figuring out if I've got the volumes of the parts of my songs right. Usually what I do is complete the instrumental parts first, and make sure each part is at it's proper "place". Such as, the lead synth part is going to be slightly louder than the bass part, etc. Thusfar, I think I'm doing pretty good in that area.

But I have two problems. First, I have a HELL of a time figuring out what volume my vocals need to be set at so as to blend/mesh with the instrumentals, yet stand forward just enough that they're prominent, but not so loud as to be obnoxiously off from the instrumental. Usually what I have to do is burn an MP3 or wav copy to disk, and take it into my car where my ear is best attuned to levels. Even then, it's not always easy to figure out if I got it right. A prime example - "Ever, Mary", the song I have on the DGN comp. Turns out the version I submitted had vocals that were way too low. I've fixed that for future use, but I'm still not sure I got it right.

Then, there's the second problem. You mentioned at CC something about volumes and the zero point. When I am working on my instrumental levels, usually my lowest instrument is set at or around 0. Then they go up from there (I start with 0 because that's the default on Cubase). But if I understood you correctly, 0 should be the HIGHEST volume, to leave room for tweaking by the final mixing.

Can you elaborate on that and correct me if I misunderstood you?

[Scales]

Is compression best to be used on shorter, louder sounds like kicks? I use it rarely, since I'm not making club-playable music, but I'm just wondering how much or where it's best to be used. Also wheres the best place to put it software-wise? Such as they say if your going to tweak your sound in fx by cutting with parametric eq, you do it first, and reverb should always be last (from what I've read).

[odims_sphere]

You hit upon this a little bit when conversing with me at CC. But please elaborate, in as simple terms as you can. I'm only using Cubase and Cool Edit Pro for making my songs, so I'm not well versed in the depths of engineering.

This is regarding volumes. I have trouble figuring out if I've got the volumes of the parts of my songs right. Usually what I do is complete the instrumental parts first, and make sure each part is at it's proper "place". Such as, the lead synth part is going to be slightly louder than the bass part, etc. Thusfar, I think I'm doing pretty good in that area.

But I have two problems. First, I have a HELL of a time figuring out what volume my vocals need to be set at so as to blend/mesh with the instrumentals, yet stand forward just enough that they're prominent, but not so loud as to be obnoxiously off from the instrumental. Usually what I have to do is burn an MP3 or wav copy to disk, and take it into my car where my ear is best attuned to levels. Even then, it's not always easy to figure out if I got it right. A prime example - "Ever, Mary", the song I have on the DGN comp. Turns out the version I submitted had vocals that were way too low. I've fixed that for future use, but I'm still not sure I got it right.

Then, there's the second problem. You mentioned at CC something about volumes and the zero point. When I am working on my instrumental levels, usually my lowest instrument is set at or around 0. Then they go up from there (I start with 0 because that's the default on Cubase). But if I understood you correctly, 0 should be the HIGHEST volume, to leave room for tweaking by the final mixing.

Can you elaborate on that and correct me if I misunderstood you?

Volume is interesting, there are so many points where you can turn up and turn down volumes. You have your source volume, (synth), In-put gain in to a mixer, (trim), Fader gain which goes from infinity (off) all the way up to, and in the case of analog past zero db. And almost all processing gear, weather hardware or plugins have in-put and out-put gains. It makes it really tricky to keep everything under control. First and foremost you should keep your source volume as high as possible with out distorting, this is going to help down the line with your noise floor, or the ambiant background noise that ALL sound sources create. Trim, or mixer in-put gain should be avoided as much as possible on hardware mixers, both analog and digital as they are notoriously noisey. If your mixer allows headroom, or (going past zero db), a basic rule of thumb is to never go above +7db if you have to do that you might want to consider compression, or look and see if there is another problem in the signal path. As far as Digital mixing goes you cannot ever go above 0db without getting digital clipping. Some digital mixers, or software mixers have what they call headroom, but in actuality they just roll back their numbers essentially making "0" actually "-4", or what ever they decide they want it to be. It's like in This is Spinal Tap "We got 11" it's still ten, it just says eleven on the dial. Same thing in the digital world. It doesn't really matter what numbers they put on there if it passes true zero it will clip.

I've had very little experiance with Cubase but just looking at a pic I found online it has a mixer emulater, should look something like this:

It's modeling an analog mixer. The faders at the bottom are where you should be making all your "volume" changes but before you get there you need to balence your sounds. You have to keep in mind that in this phase the mix is going to sound like shit. First you set all the faders to the little line about 3/4 of the way up on the fader throw. That's where Cubase has decided "0" should be. Once the faders are there you either raise or lower the source output so that your VU meters, (the light bars going up and down next to the faders), are also hitting "0". Now your sounds are balenced and that's going to make mixing a heck of a lot easier. In general it's best to try and turn stuff down rather than up then make up the gain in the main mix. It's not always possible to do it that way but you should always start that way.

As far as vocals go... I started writing a big long instructional outline on how to do vocals, but I found this which explaines it really well. I think the main point they try to hit on is that good recording technique makes for easier mixing. A good vocal recording will definatly make it easier to blend smoothly into the rest of the mix. Beyond good recording technique it's just personal taste. Most vocalsists want the vocals louder and vice versa for musicians, that's normally why there is an engineer and producer.

Understandibly though most people here cannot afford engineers and producers and so the best advice I can give for getting levels right is exactly what you've been doing. Make a mix, burn it then listen to it in as many different systems as possible, take notes, make changes and do it again. I'll adverage about 7 different mixes before I get the one I want. You may also want to think about running your vox through a sonic maximizer. Hardware based is prefered but they make plug-in versions as well. They do a really good job fo smoothing out the waveforms of sound.

[Homicidalheathen]

Thanks for this......I may have questions for you soon.....and we may have a p/t job available soon for someone with knowlege like this also.

[odims_sphere]

Thanks for this......I may have questions for you soon.....and we may have a p/t job available soon for someone with knowlege like this also.

That sounds interesting.. A freelance musical engineer consultant.. Hrmmm... *ponders*

[Homicidalheathen]

anyone on here using sonair 4.0 (cakewalk)

[CixWicked]

anyone on here using sonair 4.0 (cakewalk)

Phee and I are, for your 'Concept Exit' project. I'm sure one of us could answer some questions... maybe not, but I'm all ears.

CixWicked

[Homicidalheathen]

its ok I think I got the answer off of a forum but will write ya if it doesn't work thanks.