spectral analyzer

  • Thread starter Thread starter Krystof01
  • Start date Start date
K

Krystof01

New member
Hi,
I know this has been covered before a while back but are spectral analyzer's imperative for mixing (especially for home recordists without a refined ear and excellent monitors (to be able to actually visualise those hard to hear low bass frequencies rather than missing them)?

I mean to actually be able to visualise the actual frequencies and how they interact together so you can cut or boost accordingly can only be a good thing right?

The problem being I want one but I don't record with my computer and the rack based ones are either very expensive or not all that good.

Any thought?
Is it worth setting the computer up just for this one tool. There is an on-going thread at the moment about instrument seperation, surely this directly relates and is relevant (as I said particularly for the more inexperienced who wishes to view whats going on as opposed to rellying on inexperienced ears).

The only draw back I can think of is being dependent on it and try and formultaing your mix (are mixes formulated allready)?

I'm finished now,
What's the best way to incorporate Hi,
I know this has been covered before a while back but are spectral analyzer's imperative for mixing (especially for home recordists without excellent monitors and a refined ear for mixing)?

I mean to actually be able to visualise the actual frequencies and how they interact together so you can cut or boost accordingly can only be a good thing right?

The problem being I want one but I don't record with my computer and the rack based ones are either very expensive or not all that good.

Any thought?
Is it worth setting the computer up just for this one tool. There is an on-going thread at the moment about instrument seperation, surely this directly relates and is relevant (as I said particularly for the more inexperienced who wishes to view whats going on as opposed to rellying on inexperienced ears).

The only draw back I can think of is being dependent on it and try and formultaing your mix (are mixes formulated allready)?

I'm finished now,
What's the best way to be able to incorporate a spectral analyzer into my gear that doesn't comprise computer recording.

Thanks,
Krystof.
 
I think they're necessary for the beginner to intermediate when it comes to mixing. They just show you a lot of things you can't otherwise hear.

The best thing to do is play your individual tracks throught the analyzer and see what frequencies they are occupying. If two or more of your tracks are using up a lot of the same frequencies, then you're going to have to "make some room for them."

The reason certain tracks will sound good on their own, but then get lost when you add other tracks is because the tracks you are adding are causing conflicts. The bass drum, for example, pumps out some bass frequencies, right? Well, so is your bass guitar. Alot of the time, they're both trying to kick out a lot of the very same frequencies, so what you're getting is sort of a dual between the two instruments for those frequencies. They're battling it out down there pretty fiercely, but ultimately what happens is they cancel each other out and each loses their definition.

The way to look at eq-in individual tracks is you want to use it to help each track to "stand out of the crowd" so to speak. It's a lot like group psychology. When you throw a very large group of people together, a lot of times what will happen is people will change their behaviors in order to stand out more. In the world of mixing, that's essentially what you're trying to do.

Panning the instruments is the most obvious means, but in order to really get things done right, you need to establish their space on the frequency spectrum. In some instances, you may even have to surgically "carve out" some breathing room. Kind of like a surgeon trying to separate siamese twins.

The simplest and most obvious means of using a spectral analyzer is to help the bass guitar and drums to play nice together. View each track carefully and see what frequencies they're occupying. Usually you'll see the bass drum pumpin' up and down anywhere between 60-80 hz. The kick isn't a very stingy track, as it tends to be pretty satisfied with it's little space down there. The bass guitar, on the other hand, is kind of a hog, as you might notice. It likes to play around the whole entire space between 50 hz all the way up to 200 and sometimes even higher.

The first thing you'll want to do is figure out where the bass drum lies. Let's just assume for arguments' sake that it lives mostly at 80hz. Now, look at the bass guitar again throught the analyzer, and see how much action it has around that same region. If there seems to be a decent amount, then you're going to want to cut some of it. That's when you'll want to find 80 hz on your multiband EQ, and start cutting maybe 3-4 dbs from the bass guitar around there.

Without the specrum analyzer, you'll have to go by ear in order to figure out where that bass drum is.

Now you should be able to hear the kick drum a little better. Cut more if you have to, but as a rule, if 6 dbs isn't enough, then you probably have some other issues going on.

The next biggest source for conflict is usually the bass and the guitar. We used to have a really good thread going on this -- too bad it got lost. Anyway, someone mentioned that the job of the bass guitar should be to make the guitars sound "fuller." As much as you like the sound of your guitar tracks in solo mode, chances are it's got too much bass. And the job of handling the bass should be left to, well, the bass.

So basically, you're going to have to start cutting some of the bass from your guitar track to where you're probably not going to like the sound of it when soloing the track. Don't worry, because when you add the bass track, it should all of a sudden sound like it's fuller. That's the idea we're after. Look at the guitar tracks on your spectrum analyzer, and see what freq's they're using. Now look at the bass tracks again. Drawing a little graph for both is a good idea.

Now, wherever you see overlap between the two is where you're going to have to start compromizing. Most often this means rolling off anything below, say, 200 hz on the guitar. Maybe you can compromize and start shelving off everything above 150 on the bass guitar as well. The idea, again, is to carve out a nice little "pocket" for each instrument.

For the students of Euclidean Geometry, imagine that each of your tracks is a separate plane (collection of points going off to infinite in all directions). What is causing your problems is where your tracks are INTERSECTING. Find out where the lines of intersection are and get rid of them so they aren't intersecting anymore. That's probably a little drastic, but you get the point.

Another analogy I like to use is that of the characature. When eq'ing, think of it like this. Think of your tracks as a bunch of drawings thrown together on a board. With so many different drawings thrown together, it's beginning to look like a big, messy collage. It's hard for you to view each individual picture, since it's just sort of thrown in with the rest of them, losing it's individuality.

Now, your job as the artist is to make each drawing stand out on it's own, as well as blend in with the others. It's kind of tricky. One idea to make each picture stand out is to turn it in to a "characature" of itself. Exaggerate what makes it unique, and de-emphasize what makes it similar to the other pictures. If it has big ears, then make it's ears even bigger. If it has a small nose (in comparison to the others) then make that even smaller. If it's dark, then make it even darker, and if it's dull make it duller. And if it doesn't really do anything for the rest of the picture/collage, then consider getting rid of it altogether.
 
Last edited:
Chessrock, great post and I hope you plan to continue. Two questions for ya. As long as we're on the subject of guitars, if I've double tracked the guitars hard left and right with virually the same amp and mic settings do I need to do any eqing to seperate them or being that they're panned, will they sit ok? Question 2, do you like to bring up the vocals before or after the guitars while mixing? thanks
 
chessrock; nice analogies. I do indeed understand the concept of the spectral analyzer (from now on lets call it the SA, it's one of those horrible ones to spell); the thing I'm trying to get at more than anything are the possibilities open; what hardware and what software options are available.

Chessrock, it seems like you have experience with them, do you have/use one.

The problem as stated in the first post is that I don't record into my computer and the only ones I've seen so far (apart from rack ones in the £thousands) are software ones.

Is there more of a DIY method/approach. Like you said about graphing each out individually, with the software ones it's all done for you from what I've seen; so does that mean there's another way of reading these frequencies.

Ever since I understood the concept of the SA's I knew that I needed one (and understand the concept) but have been unable to locate anything I could practically use to do the job; any ideas.
 
I have been reading a fair bit about the SA's and it seems very clear that software on the computer is the only means for quality/cost.

So come on then; lets have some software links for SA's (prefereably SA's without much else), one that has multiple graphs, real time, wide a range as possible etc......
 
The very best of the best are made by these guys:

http://www.audio-software.com/

Krystof, the only ones I've ever used are software-based.

I've used the PAS and I now use brainspawn. The brainspawn I particularly like, because it has a time-limited demo version. In other words, you can view it for only so long before it turns off. The reason I like this is because: a) I don't have to pay for it. and b) it keeps me from getting too obsessive about it. Time limits are good for me, because I tend to be a perfectionist, spending too much time staring at the analyzer and not enough using my ears. :)

Joedirt, as far as double-tracking guitar goes, it's really a matter of taste. You don't have to worry about them sharing the same frequencies because they're going to anyway . . . that's kind of the idea behind double tracking. Although eq-ing them slightly different and hard panning is a good way to make them sound bigger and fuller. You could even go as far as to mic a guitar cab with one track, and use a POD on the other and see if you like that. My favorite trick is to play the same chords, but on different positions on the fret board and mix those two together and pan.

I get the feeling though, that I'm getting off of the subject and perhaps that belongs on another thread. :)

As for the second question, it's really about personal preference. There is a school of thought that says it's a better idea to eq everything else "around the voice," since it is your most important track in most cases. In that case, you would want to bring up the voice first and then everything else. And it's some great advice, because it forces you to focus your attention on what most people listen to the most.

Personally, though, I like to bring the vocals in later. What I've been finding myself doing a lot more lately is I will first mix and almost get to the point of "mastering" the music on it's own, pretending there isn't even going to be a vocal. Once I have the music mixed to two tracks, I will then start carving some of the mids out of it, but only during the parts where there are vocals. That way, it's pretty seamless, and allows the vocals and music to sort of "share time."

Again, for this I pull out the "SE" to find out what frequencies the voice is using up. My own voice generally falls between 200 and 1000 hz, with the bulk between 200 and 500. Now for the vocal tracks, everything below 200 either gets chopped off, or better yet compressed, so as not to interfere with the snare drum (which is usually around 200 hz) If you have a multiband compressor, that tends to work even better than simply "chopping" away frequencies like fat on a steak or mold on cheeze. We should really get in to multiband compressing a little later, but for now, if you have one, then do what you can to learn how it works.

Okay, so after I figure out that my voice is taking up the 200-500 range pretty heavily, I will go back to my music tracks and see what's going on in that region. Sometimes, I've actually gotten lucky and noticed a little "pocket" in that general region. Particularly, with my tunes, I often find a dip right around 350. If that's the case, then I'll try boosting 350 hz on the vocal track to "fill that gap" and see if I like it. If I notice any other areas that look like they could be trouble, then I'll either carve out a little "dip" in that region using eq, or I'll compress in certain instances with the multi-band comp.

There really are certain instances where the multiband comp is far superior to eq'ing, so remind me to get in to that a little later.
 
Hey Chessrock...

I've got the Brainspawn pro analyzer as well but someone here found it not to be accurate....

Have you compared the PAS to the Brainspawn?

Curious as to the outcome. The only thing I have noticed it is it appears to have a slight latency characteristic.

zip >>
 
Actually, it isn't as fast when I use Cool Edit Pro, but it seems to have far less latency with Sound Forge. Funny. Other than the latency, it is afully accurate, especially the pro version. Oftentimes it's the eq that isn't accurate. Either way, it's a matter of playing with it and getting used to using it. One thing I've noticed with PAS is that it tends to exaggerate the higher frequencies so when I use it to "master," everything comes out sounding too bassy. Other than that, it is deadly accurate for mixing purposes.
 
Cool...thanks.

I've been using the pro version (now 120 band :)) in Logic - after a chain of plugins...

that might explain the latency.

zip >>
 
Chessrock; firstly concerning the PAS it states:

Spectrum +3dB for an overall improved monitoring of signals (less bass and more highs!)

Does this explain the:

One thing I've noticed with PAS is that it tends to exaggerate the higher frequencies so when I use it to "master," everything comes out sounding too bassy

Next, there is the choise between SA pro, SA basic and SA live, the basic is just that but maybe enough; I don't really get the difference in the other two, can anyone explain:

http://www.audio-software.com/

---------------------------------------------

I'm thinking of maybe getting a laptop off ebay just for this SA software ( don't know which one yet (I really dont want to use my computer with any part of the recording process cus it always fucks up and has errors etc.)).
 
Will any of those work as Logic plugs that you can just stick on a bus?
 
No prob man...

The new pro version has some cool compare functions and 120 bands...

...you point the mouse at the "trouble spot" and it tells you the frequency.

Works pretty well....
 
I just got an email add from the guys at brainspawn the other day. Apparently they have an evern newer version with a "tone generator" or something like that. I'll have to check it out. I wonder if they know I've still been using their demo version all this time. :) Man am I cheap.
 
Hey Chess....

The pro version is WELL worth the 35... ;) Screw the demo...step up to the plate man! :D:p

zip >>
 
Alright the brainspawn it is. I know virtually nothing about computers, the system requirements for the software are:
----------------------------------------
SpectR-Pro
System Requirements

DirectX Version:
Windows 98SE, ME, NT 4.0, 2000 or XP
DirectX host application
DirectX 6.0 (8.1 recommended)
600 Mhz or faster processor
64 MB RAM (128MB recommended)

SAWStudio Version:
Windows NT 4.0, 2000 or XP
SAWStudio
600 Mhz or faster processor

------------------------------------------
Can someone please explain exactly what I would need to be able to use it, especially what lead/cable you need to plug into the computer and where.

Thanks,
Krystof.
 
No cables at all...

If you use any program which supports DirectX plugins (n-track,CoolEdit,SoundForge,Cakewalk,Logic,Waves etc...) it can be inserted into a channel strip, aux bus or master bus (multitrack program) OR simply selected in a stereo editing program.

Do you have any of these "host" programs?

zip >>
 
do you know of any program that does 3D spectrograms for Mac?
I remember there was an old one called Soundedit that did Volume vs. Frequency vs. Time.
I found that it was great for looking at kick vs bass in a sample basis, not for a whole track. This way you see time frozen, and that helped me to understand sounds.
There were some that made it with just two axis and colours for intensity, but I found the 3 axis ones better.

Thanks in advance.
 
Back
Top