Skip to main content

Hi all,

I'm learning how to mix with DAW, and I need bit of help here.

First of all, here's a bit of info of what I have:
My tracks are completely made up of VST instruments (and I don't see much to record live tracks myself in near future). I don't have sophisticated room nor gear, so what I did was I tried to listen to as many speakers as I have and I can find (including headphones and earphones) at different playback system. Also with no gear (limited space), I have to choose software-based tools (EQ, compression, etc) for my work -I'm using Vienna Suite currently.
I have already tried to mix several pieces of music (my own) -most of them are orchestral, and some jazz. As much as possible on those works, I EQ-ed each and every individual instruments to clean them up -and hopefully make the sound nicer (for VSL instruments that I own, mostly I use Vienna Suite preset for applicable individual tracks).

My problem is this:
With lots of VSTi running on, the pieces I was (and am) working on quickly become muddy (with and without individual tracks EQ-ing, I've tried both).
I realize that, in many of the instances, (from many sound system, including a big one that I very seldom have access) if I cut an EQ on around 600 with big Q, the muddiness clears up. So that's what I did, so far so good, but then this problem always come up: I always end up with a piece that is clear in the mid, but piercing the ear on (I suppose) the high that makes my ears fatigue.
I have not been able to completely remove this annoying, piercing, high (and sometimes I don't even have clue on what's wrong -I'm sorry here, just a beginner willing to learn). Things that I tried includes cutting at 3000 and 5000 (sometimes 2000 and 6000, even 1000 or 1500 to get the lower harmonics).

I have read on the net that the highs are common problem in VST instruments, but knowing is one thing and solving the problem is another thing :(

So, what I want to ask are these:

  • How to clear up the muddiness without introducing the piercing high? Did my method pulling down at 600 acceptable? Or I have to think something else? (I read on some blogs that I have to take care of individual instrument on their shiny spots first, but hey, I have done that! -hopefully in good way though)
  • If I still end up with piercing high, what should I do to eliminate it?

Thanks

Comments

RemyRAD Fri, 10/28/2011 - 16:14

Well, firstly, it appears you're working purely with VST sampled instruments. Those are already expertly recorded and processed. Going to each individual track to clean them up and make them sound great really is not the way to do that. All things being equal, everything shouldn't be equal. So while you don't have the room to actually be able to record instruments, you are robbing yourself in the art of recording anything. While us recording engineers in the United States are really referred to as recording engineers/mixing engineers I think the British have a finer term called "balance engineers" because that's what we are really doing. And working in the real world, no one equalizes anything before hearing everything played back together as it was recorded. That's your first fundamental mistake. You need to balance your mix and play everything back without any equalization whatsoever. You need to come up with a good usable mix that way. That simply takes technique and experience. Upon playback of all sad instruments and a refinement of your mix, you can then go ahead, if needed, and add some equalization where needed. There is no law about 600 Hz. Where did you find that one? If that was based upon your playback systems, perhaps your playback systems are not refined enough to present to you proper sound. High frequency aberrations and grunge are just baby steps in learning how to actually mix well. Most instruments as a rule do not present flat frequency responses and neither should your music. Sometimes it's important to bandwidth limit certain instruments in order to make harmonic space for others and it's not all at 600 Hz. What you're basically trying to create here is what we call a SMILE CURVE which high fidelity equipment usually does with a loudness button engaged or the bass and treble controls turned up. It may sound good initially but usually causes ear fatigue within a short period of time. Just like you are experiencing. So everything cannot sound twinkly bright otherwise it all adds up to a high-frequency traffic jam for which there are no traffic control signals in use. So you have run the red light of high frequency usage. Those hit recording engineers were not automatically hit recording engineers when they started. But they did possess the ability to know where everything should be sonically. That's what you're learning right now. So when you tell the doctor " IT HURTS WHEN I DO THIS"... the doctor's response is usually " DON'T DO THAT".

One of the important things to do as a young engineer is to obtain as many quality recorded CDs by notable engineers that you can get your hands on. Those are your references that you need to emulate. Those are the references that you should play just prior to your mixing. Those are the references you should play in the middle and during your mixes. Because those are your references. And those references will better help you understand what is coming out of your monitor systems. So far it appears you're working without any references other than what you're listening to? It's fine that you are listening to your material on numerous different playback systems. That's always a good thing. But you still need some kind of reference monitors that you are extremely intimately familiar with, regardless of price or quality. Everything is 100% subjective in this business. No, everything does not need to be 20-20,000 Hz. That's simply the range of human hearing of somebody under 30 years of age. One of the most popular recording microphones of all time is the SHURE SM57/58 dynamic microphones. Those only have a response out to 17,000 Hz and yet virtually every guitar, every drum set has been recorded with those on nearly every platinum record. And we don't turn up for high-frequency response or suck out all of the midrange response just to accentuate the high-frequency nature of the music. That's just beginner slop engineering without technique. So this is something that must be learned through careful listening of other quality material. Your sampled stuff was already recorded in the $1 million studio with the $250,000 audio console into a $50,000 recorder. You're here to merely mix these quality samples together in a cohesive manner and that starts with first listening not equalizing for equalizing sake. Unfortunately that is what you're doing and you already know what you're getting, not what you want.

Of course, there are certainly ways to deal with product with excessive high frequencies. Limiters are frequently used in this capacity that have the capabilities of side chaining. A sidechain insert on a limiter (hardware or software) allows one to tailor how the limiter will respond to different frequencies. If one connects an equalizer and boosts the higher frequencies on the equalizer in the sidechain, the limiter will be more sensitive to high-frequency material. This will cause greater gain reduction to occur to the high frequencies and not greatly affecting the lower frequencies. Every FM & TV station utilizes these. They utilize them for different reasons from yours but they are still necessary. Back in the analog days of cutting records these high-frequency limiters also had to be utilized since there is a high-frequency preemphasis curve built it into all recording systems for disk & analog tape along with FM & TV. So these higher frequencies had to be tamed a little more vigorously through high-frequency limiters. Another type of limiter reduces sibilance on vocals and works particularly well on female vocalists since their natural harmonic content is higher than that of most males. Recorded sound can take a nice little squeak and turn it into a raucous obnoxious zip. That's because electrical recording as we know it is completely unnatural. So we have to have special plans and techniques to make our music listenable. And that means not listening with your eyes but with your ears. Making a recording may be scientific but not necessarily clinical.

I'll wrap this up by saying to fix your mistakes, you could slather it with high-frequency limiting. Otherwise, you could just learn how to mix correctly. I mean do you really think that Stevie Wonder ever looked at a response curve? Do you think he ever saw a waveform? Do you think that he looks for pretty knobs, dials and lights? He mixes the way you're supposed to mix. I've even watched Stevie plug in a recording studio's worth of equipment, all by himself in the top half floor of the Omni Shoram Hotel suite in Washington DC. That's because he has acquired magical vision and can see with his ears. This is not something that is taught easily to entry level folks like yourself. It takes years. It's like the old musicians joke of how to get to Carnegie Hall, that's easy. Practice, practice, practice. And you'll get there. Plenty of folks like yourself are producing stellar mixes in their bedrooms today. You'll be doing the same soon.

Ear on rear
Mx. Remy Ann David

SilverPiano Tue, 11/01/2011 - 02:41

Hi Remy,

Thanks for your reply and solution. I have looked into one particular music that I'm working on again, the analyzer clearly shows up that the 4k is way too much up -so for quick solution I shelve out some and the piercing high dims away.
For the long and permanent solution, I do agree with you to correctly learn to mix and train my ears -and this is what I aim for when I asked if my (first) method is correct or not.

Thanks again Remy!

RemyRAD Wed, 11/02/2011 - 12:46

You really can't judge things by the frequency response spectrum display. It does give you an overall look at what you might be listening to you and from what you've described, somebody is screwing up your recordings by adding too much presence boost in too many places. I use the spectrum analyzer to look at the overall meat and potatoes and to spot any particular specialized problems. So while you indicate what kind of software you are using, that doesn't indicate what kind of source capture equipment you are using? If you are using your computer's built-in soundcard apparatus, you can expect dismal results like you described. If you are using a quality computer audio interface the problem is in your technique. So your question is a little bit like the doctor joke, " it hurts when I do this..." if you go to your auto mechanic and tell them your car won't work what are they going to be able to tell you if you're not there with your car? So your problem isn't with your software I can assure you of that. It may be your technique or it may be a $15 built-in computer sound card device? We have no way of actually helping you with your vague descriptions. Real-time spectrum analyzers don't tell you what anything is going to sound like. We don't hear a flat response but it's important that our equipment does. That's our baseline reference. So crappy equipment can be equated to severe hearing impairment. That is to say, equipment of that caliber relates your recording to the real world much in the way a person with a severe hearing impairment may perceive it. Or so to speak. Just because your Toyota Corolla has a stick shift doesn't mean that it can perform like a Lamborghini with a stick shift. The basics still hold true but the quality level is all different and so is the performance that is perceived by everybody. And in the same aspect, my two cylinder Honda motorcycle doesn't sound like a two cylinder Harley motorcycle. Even though it still has two cylinders and two exhaust pipes. And it was a lot less expensive. So it neither sounds like nor performs like a Harley. I can't make it into a Harley even though it can be more modified like a Harley. It'll never be a Harley. So I either have to drive I have happily or not drive at all. Strangely enough even though it's not a Harley, I still love it. It still performs well for my purposes even if I don't look as cool as folks on Harley's. Does that make it less then desirable or unusable? No it doesn't. I can get just as easily killed on my Honda as I could on a Harley. Is it going to outperform a Harley? Hardly. But strangely enough, it still works with just 2 tires. But only seems to work well that way when I am on top of it. So what's the best remote control I can purchase for my motorcycle so I don't have to be killed when I'm going places? Answer:... my brain along with its built-in damage. So I have to make the best of what I have to work with which is damaged goods. It's okay nobody knows how damaged since I'm good at workarounds.

I think more my brain just fell on the floor? Did you see it anywhere?...?
I think it's me?