Spectral analyzers and mixing

Chris Fallen

Chris Fallen

New member
Okay, this whole area about mixing is getting me really confused. I hope you guys can supply me with some answers and some basic or textbook information that'll help me out (and I know you will).

First off. I'm reading posts a lot and people are always saying 'cut off 4dbs off of the 500khz' and such. Someone posted about this. Another person talks about 'carving' out a section for the bass to fit with the drums and such. How do you go about carving out a section? How do you know what section to carve for what instruments? Is it just dropping the dbs in that area so there's not a large amount of that instrument in it and then filling it in with the other?

Another thing I want to ask about is spectral analyzers. How can I go about using one of these to make my mixes better and are there any free ones I can download?

Thank you.

Christopher
 
Christopher

I recently went through the spectrum analizer situation not long ago. They are mainly used to get your monitors in your control room flat. If you have too much 1K (or whatever) it will tell you. As for using one to mix, it's not a very good idea. A more knowledgeable engineer than myself told me this: If you try to match up your mix to the spectral analization of another mix, it won't work. Because how will you know if it's the vocals or the guitar adding that boost of 1.73 in the mid range?

I recommend you don't use it for mixing....but I could always be wrong.

As for the other frequency stuff, take a look at a 31 band equalizer. From left to right, it has all the frequencies these people are talking about from 20hz up to 20khz. For each frequency you can boost or cut certain db by moving it up or down. That is a Kindergarten explanation, but I wanted to be as specific as possible.

Every kick drum or bass guitar will sound different on every song & every style of music, so no one can tell you the "secret formula" to make your mix wonderful. However, if the kick drum is pumping out 250 hz, you don't want the bass pumping out a ton of the same frequency because then they're competing with eachother & you'll get too much of a boost in that low fequency range....it's a delicate balance, and will just take time....and 100 mixes!! Let each and every instrument have its own space by proper frequency placement, panning etc. That's the best advice I can give without sitting down and mixing with you.

Michael Angelo wasn't a great painter the first time he picked up a brush.......mixing is an art.

Bet of luck

Drummerbones
 
The same thing can apply to a live situation. A couple of weeks ago, my son's band was working on a new song.....things were going along well but the vocals just weren't cutting it. I sat and listened for a bit, then asked both guitarists to modify the tonal settings.......the improvement was immediate and everyone heard it. All of a sudden the guitars were out of the vocals' "space".

I have to thank these forums for that improvement, as it was here that I first learnt about giving everything it's own space where possible.

:cool:
 
I wouldnt try to use a 31 band grahic to "carve" space, since "carving" requires much more precision and control than you will ever get from a graphic.
THe idea is to help make instruments cut through and your mix more clear. In the standard homereccer mix, you can almost always improve the mix in general by cutting some at about 315... but you must have a parametric to do so. Making space in a bass guitar for the main frequencies in the kick drum, or vice versa, is pretty common. You can use an analyzer to see what those freqs are and compare serperate tracks to give you an idea of where to start. Sometimes it works, sometimes not. Whatever, When it helps, it helps. Ya know?
Chopping the bottoms off of guitars is pretty common too, and some people really cut into them, like at 100hz hi-pass or higher, which is scary at 1st, cutting into that much low end, but it can really open up the mix and make the bass instruemtns more clear. If you can swipe a db or 2 out of every track in the mud frequencies and compare the overall mix on an analyzer, you can really see it, too.
I'm babbling.
Later.
 
Chris, get ahold of a decent parametic eq.....takes turns with different instruments (guitar,bass,vocal,snare,kick,etc)....solo the track.....set up the Q pretty narrow with a big boost +10db......sweep thru the frequencies and HEAR what each frequency sounds like......then set it up for a cut and do the same.....this is a good exercise......

this will come in handy when you have a problematic track...say you have a guitar track that has an annoying honk to it.....set up you eq like above with the big boost and when you get to the point where that honk just jumps out and at its worst, youve found the offending frequency....cut it......
 
MASKING......

Ummmm, you do know what "masking" is, right? If you do, then ignore the following paragraph. I just noticed that no one mentioned it. If you don't know what it is, then it's probably no wonder you're confused.

The whole reason that instruments "compete for the same space" is because of a phenomenon known as masking. It goes as this: Given two sounds in a similar frequency range, your ear will only hear the louder of the two. That's why instruments "cover up" eachother. And that's why there is "space" in mixes. Wonderful masking..... you can hear a guitar covering up the vocals, and your first reaction would be to say, "The guitars are too loud!" Then after experience, you learn, "Oh, if you just cut a little of the guitars between 1-2khz, the vocals are understandable, and I don't have to lower the volume of guitars. Thus, they "fit" together.

If I could some up mixing into one word, I would describe it as "relativity"
 
Gidge said:
Chris, get ahold of a decent parametic eq.....takes turns with different instruments (guitar,bass,vocal,snare,kick,etc)....solo the track.....set up the Q pretty narrow with a big boost +10db......sweep thru the frequencies and HEAR what each frequency sounds like......then set it up for a cut and do the same.....this is a good exercise......

this will come in handy when you have a problematic track...say you have a guitar track that has an annoying honk to it.....set up you eq like above with the big boost and when you get to the point where that honk just jumps out and at its worst, youve found the offending frequency....cut it......
Click to expand...
checking the room for sonic problems is a lot more complicated than just sweeping a sine wave. this method does not take into account many room issues such as early reflections (comb-filtering), room modes, speaker non-linearities, possible parametric-EQ non-linear phase problems, etc.... i would never recommend this approach and i would not recommend placing a parametric EQ on your monitoring system but if you do make sure it's a hi quality EQ with linear phase.

unfortunately, it's going to take some study to get your speaker/room response where you want it. i recommend you download ETF and let it analyze your setup. it'll help you fix the problem areas with a scalpel and not with a sledge hammer.

now using a Spectrum Analyzer on your mix really isn't telling you much of anything. it's neat to look at, but aside from that there's little coorrelation between the display and what you're hearing due to the speaker/room interaction, room size, speaker placement, etc.... unless your room has been EQ'd with treatments, the Spectrum Analyzer is of little value.
 
Sonixx, i wasnt recommending that to check the room for sonic problmens......it an exercise to learn what different frequencies sound like on certain instruments....nothing more/nothing less....
 
Chris,

It saddens me that so many people have been dodging this subject and leading you so far astray.

"Carving out" Frequencies:

If you've ever listened to one of your mixes and said to yourself: "Why does the bass sound so muddy, as opposed to in-your-face fat and puchy?"

Usually, there are two things going on:

1) You have too much unnecessary bass content in the extreme low frequencies. The first thing you should do is roll off all frequencies below, say, 60 hz.

Most eq's should have a rolloff feature that will allow you set where you want them to be rolled off -- rolloff just means that everything below that frequency is just gradually reduced in a slope as opposed to just having it drop off like the edge of a cliff.

2) You probably have some conflicts going on between the various instruments. Ideally, each instrument should occupy it's own "spot" in the frequency spectrum. Naturally, there will be some overlapping where an instrument or two are both occupying a lot of that frequency range. Bass Guitar and bass drums are probably the most common example.

What you want is for each of the instruments to play nice instead of trying to hog up all of a particular freq. range. Suppose the bass drum is occupying mostly the 60-70 hz range. Yet, the bass guitar is also pumping out a lot of energy in that very range, as well as a lot of energy on up to 150 and even 200 hz, as well.

In this instance, you're going to want to scold out your bass guitar for being a big frequency hog. Your bass drum isn't being greedy -- it's happy with it's little 65 hz slice of the pie. It's time to make the bass guitar compromize a bit. So what you will do is "carve out," or cut some low eq from the bass guitar around 65 hz. By doing this, you're giving the kick drum some of it's own "territory" where it can breathe free without having the bass guitar jumping all over it.

Bass guitar before carving:

. . . /////////////////////////
0 hz . . . . . 65 . . . . . . 200hz


Bass guitar after carving:

. . . ///////// . . . ///////////
. . . . . . . . /////. . . . . . .
0 hz . . . . 65 . . . . . . . 200hz


The same can be said of the accoustic guitar. When you think of the acc. guitar, you don't think of a big, boomy sounding instrument, yet alot of the time, it kinda' wants to be one. And in the process, it winds up treading on the bass guitar's territory around 150 - 200 hz. Now the accoustic guitar doesn't need these frequencies as much as the bass, so a good idea is to roll off everything below this range in order to give the bass some of it's own breathing room.

Spectrum Analyzers:

Spectrum analyzers will give you a graphical display of what frequencies are being occupied by a given instrument. Say you do a freq. analysis of the bass guitar, for example. Now suppose this analysis reveals that the bass guitar actually has very little going on in the 60-80 hz range.

Now you know that you were wrong, and that the bass guitar actually ISN'T the culprit in this instance. After apologizing profusely to the bass guitar for calling it a bass hog, you move on to the bass drum. After doing a freq. analysis on the bass drum track, you notice something interesting: Every time the bass drum is hit, you see the 65 hz light up, along with a whole lot of stuff going on around 20 hz. You have now spotted your culprit. The bass drum doesn't need all that 20 hz action. That's probably what's stealing a lot of your bass energy. So you quickly cut it . . . and bingo! Your mix is now cleaner and punchier in the bass region.

This is all very over-simplified, but I hope it at least gives you the gist.
 
Thank you a lot lot lot lot lot for that example, chessrock. I'm still a newbie, so when someone spells everything out for me, it helps. Thank you.

Christopher
 
Chessrock, good, helpful stuff. Actually I've got the the bottom end sounding good now. Do you have any carving tips for the midrange? And what about topend? Does rolling off or cleaning up high end stuff help?
 
Chris,

If you get a spectral analyzer it can help you to SEE what you are hearing.

If you are using a software based recorder (DirectX, VST, etc...) you can add the spec analyzer to your virtual track's FX insert and then you can see where the frequencies are.

One thing i like to do (once i get my mix levels right, but before i start panning) is to mute all the tracks except the kick and bass, put spec analyzers on both tracks and on the output channel. this way i can SEE the individual tracks and their summed result output. it's helped me to learn to HEAR how scooping frequencies (using a parametric EQ) from the bass can help the kick sit in the low end right. of course, once i do this i normally end up lowering the volume of the kick some because it's not fighting the bass for it's territory anymore.

i don't recommend using a graphic EQ for scooping because you can't control the shape, ie.. width, ie.. Q of the scoop.
 
Wow. Long post. I need to get a life . . .

joedirt said:
Chessrock, good, helpful stuff. Actually I've got the the bottom end sounding good now. Do you have any carving tips for the midrange? And what about topend? Does rolling off or cleaning up high end stuff help?
Click to expand...

As far as the midrange stuff goes, you're probably going to have some conflicts going on between the vocals and electric guitar. It's invevitable, but it's not really a bad thing. The same principles will apply here.

When working with vocals, keep in mind that they are usually the focal point of a song, so it's okay to let them hog up a lot of the midrange. Let them be stingy. Voices sound much fuller when they are able to resonate a wide, balanced freq. spectrum, so do as little cutting as possible here.

With this in mind, what I am about to say will seem to contradict things, but it really doesn't: In order to make the voice sound fuller, I usually look for the most prevelant frequencies that stick out in the spectrum analyzer, and then eq them down a few decibles. What that will do is make the "spectrum" of that voice wider by "leveling it out." Now the frequencies in that person's voice that normally aren't as prevelant can be heard better. Remember, though, that your ears should be your final judge on this. Sometimes these "main frequencies" I'm telling you to cut could actually sound better if left alone. Maybe they give the person's voice it's unique character, or they help it cut through the mix better.

There's not much to worry about with the guitar and voice. They sound sonically different enough that there's no real danger that you won't be able to distinguish the two. Just do your best to eq down any frequencies that tread too heavily on the voice's turf. Also beware of large freq. jumps durning times that they do meet on a common freq.

What's most important for the guitar is to cut out some of the bass below 150 hz, and maybe even 200. That's the bass' territory, so think of the two as almost one instrument. Let the bass make the guitar sound fuller. Johnboy Walton said that in another thread, and I think I'm going to steal that one from now on.

The guitar is always the easiest thing to hear anyway, and it gets enough help from the bass, so feel free to carve away on it if you absolutely have to. :) Can you tell I'm a bass player yet? Another approach I might take is to mix down just the music itself first, and then carve out a space in that for the vocals. I rarely have to do this, but sometimes find it necessary when the voice is having a real hard time cutting through.

The only danger you're going to run in to on the high end are siblant vocalists and/or swooshy or harsh cymbals. The best cure for this, actually, is preventative medicine. Just use good microphones and be happy during mixdown when your life is made 100 times easier. Actually if YOU USE THE RIGHT MICS and put them in the right spots, you can pretty much DISREGARD THIS ENTIRE THREAD!

A lot of cheaper models like the Rode NT1 or the Marshall 2001 have a serious bump in the higher freq's around 12khz. This can sound really nice on some things but really bad on others. It can sound fine on acc. guitar. It can awesome on some vocalists. Or it can sound really AWFUL on more siblant voices. In my opinion it ALWAYS sounds BAD on drums and/or tamborine (makes hi-hats and tambo sound like fingernails on a chalboard). Harsh preamps can have a similar effect.

If you abuse these types of mics in that fashion, I recommend doing some serious gouging out around the 12 khz range on those tracks. Or you can pull out the microphone modeler and model it to an earthworks or an octava mc012 . . . something with a flatter response in the higher frequencies. Ideally, you should use the behringer ecm-8000s, oktava mc-012's or Marshall mxl 6003's when tracking drums, acc. guitar, or tamborine, anyway.

There's no excuse when you can get them as cheap as they are these days.

So to FINALLY answer your last question: If you track correctly and use good mics and neutral-sounding preamps, then you should be able to let the high-end get as loud as the rest of the mix. If not, then it's better to cut out anything that may sound gritty or that will tire your ears out over time.

On a final note, keep in mind that the area where our hearing is most sensitive is going to be around 2,000 to 4,000 hz. It's okay if there is a dip in this region, because our hearing will make up for it. The bulk of a song's energy will often be around 65 - 80 hz (bass drum), and 200 hz (snare). It's okay if they peek a little louder than the rest of the mix. The midrange should have the most consistant action going on, while the high end is there to provide crispness and air.
 
Last edited:
Here's a ridiculously simple addition to the conversation:

Pan your guitars away from the center giving the vocals their space in the middle. There's a LOT more to it than that with frequency issues, but it looks like you're already pointed in the right direction on that. Creating space in a mix is more than frequencies. Reverb for depth on certain thing, panning and EQ to create seperation.....

Every element effects the other, and it's a careful mix of all of them. Kinda like baking a cake!

I'm glad I can record cause I sure as hell can cook! :)
 
Thanks a mill. Things are sounding much better allready. I'm just trying to get this stuff sounding as good as I can in mono before I pan and add depth (verb and delay) I do have another question while we're on the topic. If my snare is at 450 and my hats 3600 (an industrial sample) should I do any carving for them? Or because they're not similar sounding to any other instruments, is this not necessasary?
 
Tough to say without actually hearing it. It would also depend on what other instruments/voices are in the mix. If you 've got any elect. guitar going on, you might need to carve a little out of that so you can hear the snare better if that's at 450.

450's a little high up there for snare. Any way you can add some lower frequencies (around the 200-300 hz range) to beef it up will really help. If you can do it without adding mud, that is. :)
 
Yeah, I've got a clean, chorused guitar and somewhat of a synth pad going on as well. Maybe I'll scoop a little 450 on these and a/b them. As far as the vocals, I'm assuming that you try to leave those alone as far as carving for the snare? Thanks for your advice, my mixes are getting better and better!
 
You must log in or register to reply here.
Back
Top