Objavljeno 18 maj 2008 - 07:10
en psihič z enga foruma si je uzeu cajt in razložu Musical Equalisation....splača se prebrat
I remember coming across this expression when I started studying the theory of sound in greater depth. I would hear people talking about how a specific production had really musical eq work done on it and I never quite grasped the definition of the term. However, as time has gone on I've developed my own understanding of what makes the application of EQ musical and how to achieve it.
Now, I am by no means an expert on the topic. But I thought I would share my insights and the techniques I've learnt to create very harmonic equalisations in my productions. I think this would be a really interesting topic for conversation around here and I'm very interested to hear what others have to say on the topic and if anything I'm saying isn't quite accurate, then I welcome any further clarification.
This is quite a detailed explanation so I'm going to write it in several parts. The first is my attempt at trying to explain what I am referring to when I describe an EQ's application as being musical.
Before I you read on, please forgive me for my crude laymen's explanation. I've always found this topic to be very difficult to understand and I am yet to find any freely available text on the matter which does any better than my attempts below. The only solid reference that I have on the topic as I write this is years and years of anecdotal information, countless hours of experimentation and the last remaining memories I have from my physics and music theory pre-exam cram sessions.
PART ONE: What does the term musical mean in reference to equalisation?
In simple terms, musicality in EQ refers to its application in a way that accentuates the harmonic content of the sounds being effected.
This is a process of identifying the frequency bands that are present which are in harmony with the overall musical characteristics of the sounds being used. A crude example that demonstrates this would be an A note with a root frequency of 110 Hertz; also the root of the western octave. Logically, you would find that a boost of a few decibels at approximately 220 Hertz would further enhance the musical characteristics of the note in question as this frequency is exactly an octave higher than the root frequency. Such a boost would accentuate the harmonic content which give the note its characteristic sound.
Obviously, the application of such techniques is much more complex than the above example. Such an example would only really work on a sound with very little harmonic distortions and only for A notes whose frequencies simply need to be doubled to achieve an increase of one octave.
Like this:
A0 55
A1 110
A2 220
A4 440
A5 880
A6 1760
A7 3520
A8 7040
A9 14080
Let's just say that every other note's frequency is determined by taking the frequency of the high A note in its octave and dividing it by its fraction of 12. I think I'll leave it at that because such a division involves nasty mathematical symbols which may cause your head to explode. However, if you are interested, read this:
Pitch (Music) - Labelling Pitches
I guess the point that I am trying to emphasise is that due to the exponential nature of our pitch perception, such an application is not linear and definitely not necessarily bound by a concrete set of rules. Therefore the accurate application of EQ in a way that is musical cannot simply be achieved by using mathematics. Paying attention to the numbers will only serve to give you a starting point in your search for the music.
The best tool you have at your disposal for identifying these frequencies are your ears and in part two I'll share a few little hacks I've learnt along the way that will let you quickly identify the musical content in your sounds and bring them out to full effect.
PART TWO: Further background and considerations to make before you begin
Okay, so in part one I think I managed to explain my definition of musical application of EQ and it's generated some great
conversation so far. This next part is going to look at some simple techniques you can use to apply your EQ in such a fashion. Now, before I get on to that, I'll need to provide a little more background so that this all makes sense.
There are two primary functions which you would use EQ for.
The first is using EQ for sonic treatment. I believe that this is EQ's most important and most difficult method of application. Its results should be very subtle and a lot of conscious ear training is required to properly observe them. Due to its obscurity, I believe that this is the least observed application of EQ in the circles we computer "producers" travel in. The fact that it's often overlooked is understandable because it's a slow repetitive process of observation and adjustment and we work in production environments that allow us to move very quickly from one stage to the next; and given the quality of sounds we have to begin with it's often easy to avoid using EQ to treat your sounds once you've arranged them. But your mixes will sound even better if you take the time to treat all of your sounds with EQ and limit their output to their most audible frequency bands.
The second primary application is using EQ for effect. This is a simple process in comparison to treatment. You tweak - you like - you keep. The results are obvious and applying EQ as an effect takes only as long as you take to move a few knobs and click a few buttons. I believe using EQ as an effect creates more problems than it solves. There are so many other tools available in synthesis that allow extreme alterations in sound and most of these are far more effective. I usually only stick to the most basic applications of EQ as an effect which is using single filters - High/Low/Band Pass Filtering - as these are essential to the type of music that we create.
The musical application of EQ is more a process of sonic treatment than EQ as an effect. I recommend that you apply EQ in this manner in addition to a more practical sonic treatment which would deal with problem frequencies, isolation, gain structure and the like. Typically, I find that it's best to apply standard sonic treatment before the musical treatment because this ensures that any undesirable frequencies are removed or limited and thus are not being boosted any further.
Musical appplication of EQ is an exercise in emphasising the harmonics of a sound.
There are two types of frequencies that determine the pitch and character of every sound; its root harmonies and its formant frequency.
Root harmonies are directly related to a fundamental frequency which is a note's musical pitch (A4=440Hz) and these are the same for every sound that is tuned to a traditional musical scale. It is fair to assume that the frequencies that are in key with your composition are the ones that you should be boosting in your mix. Theoretically, these are the frequencies which define your music. But in practise, you should use those frequencies as a starting point as these root frequencies only determine the perceived pitch of your sound. It is a sound's formant frequencies that determine its character.
A sound's formant is occurs where the main concentration of acoustic energy occurs. Put simply, the formant frequency is the frequency which contains the body of a sound's energy and determines its character. This is not necessarily the same as the fundamental frequency and is typically different for each sound you will encounter. Formant frequencies are the result of harmonic distortion which is why they vary from sound to sound.
Formants occur because every layer of harmonics present in a sound resonate with every other which then causes a slight variance to these frequencies. Resonance, in a physical sense, is simply a process of absorption where one mass absorbs the energy of another. The physical effect of this absorption on each harmony present is a slight shift in frequency. This explanation is crude at best and does not really explain the reason for these changes. However, I hope it gives you a basic understanding of the effect that harmonic distortion has on a sound. The main thing to remember is that the resonance of harmonic distortion has a physical effect on the sound that alters its frequency characteristics.
A further explanation of formant frequencies would merit an entirely new thread and I'd have to do a lot more research. The majority of material you'll find on formants deals with vocal formants. There is one good discussion available here if you're interested.
In my experience, formant frequencies are usually in close proximity to the root harmonics of a sound. This is why I suggest using these as a guideline when you're applying your EQ. It's not essential that you understand the exact nature of formants or root frequencies to continue this exercise.
Before you start
The technique I'm going to outline does not deal with precise adjustments on specific frequencies. Wide adjustments are the best because if you're working with music, it's most likely you'll be dealing with a range of frequencies which vary from note to note. Of course, if you're dealing with fixed-pitch elements, you can have a lot of fun with narrow bandwidth adjustments and if you spend time tuning your percussion, for instance, you'll find that such adjustments will enhance this process.
Although the end results of this process are very subtle. The process of identifying musical frequencies involves using extremely narrow bandwidths and very high gain settings. Relying on speakers for monitoring will yield less accurate results as the room modes present in your room will colour the output. Filtering using such high gains and narrow bandwidths will increase the effect of these room modes. For this reason, I suggest you monitor with a good set of headphones that have a fairly flat frequency response. Using headphones ensures better accuracy because the output arrives directly at your ears with no additional colouration. Their use is even more essential if your room has little or no acoustic treatment and even more so if it's square.
PART THREE: Practical techniques for applying musical EQ to a single sound that is a tuned instrument
The first technique that I will describe in part three is applied only to a single sound source and is best suited to dealing with a pitched instrument. That is, an instrument which is tuned to various musical notes. The steps I describe here are used in the other applications which I will outline in part four.
There are three main steps that I take during this process:
1 Find the center of a sound's musical frequency bands
2 Determine the width of these frequency bands
3 Boost the frequency bands
You're welcome to use more complex analysis tools like spectrum and frequency analysers while you're undertaking these steps. However, I've found that such tools hinder, rather than assist, the process. The best tools you have at your disposal ar'e your ears and if you trust what they are telling you then you'll get the best results.
Now, if my explanations of roots and formants in part two weren't entirely clear, and are adding to your confusion, then just focus on these points:
1 - Points of harmony in every sound specific to its root frequency that are related to its perceived pitch
2 - There are other points of harmony within a sound that are not necessarily specific to its root frequency which relate to its character and differ from sound to sound.
Like I said, you'll be relying on your ears more than anything else and, if you've got a good pair, you won't need to rely on anything other than what you hear.
1 - Finding the center of the musical frequency bands
I've already said that the best place to start searching for the music is at a sound's root frequency. I recommend starting to search using a section of audio that primarily contains the root notes of your composition's key signature.
In my case, such sections are usually found when a sound is introduced as my composition style typically introduces each element in a more rhythmic arrangement containing said root note and very little others. If your composition style differs from this or if your material is more melodic, then you should either find a point of less melody or create such a loop for the purpose of this exercise.
Once you've chosen or created such a section, take the following steps:
Insert a parametric EQ with variable bandwidth and frequency.
Engage one of the EQ's filters and set its frequency according to your root note.
Set the filter's bandwidth to a very narrow setting.
Boost its gain to the maximum setting.
Now this won't very nice at all and you'll be redlining all over the place (another reason you should use headphones). But you will notice that the sound's output will now be dominated by a single frequency. Now you want to set about tuning the filter by employing a repetitive process of elimination that involves comparison and adjustment.
Listen to the pitch of that dominant frequency and compare it to the pitch of the musical content.
If the dominant frequency sounds flat in relation to the musical content, make very small increases to the filter's frequency.
If the dominant frequency sounds sharp in relation to the musical content, make very small decreases to the filter's frequency.
Listen and compare the dominant frequency to the musical content.
Take note the frequency you've arrived at and make smaller adjustments each time before making your next adjustment.
Repeat the process accordingly making smaller adjustments each time
It is essential that you note the frequencies you have already tried before making your next adjustment. This allows eliminate them as a possibility and narrow down your range of possible frequencies to try for your next adjustment.
Eventually, you will arrive at a frequency that is in tune with your musical content. You will know when you reach it because the perceived pitch of your music will be incredibly clear in relation to the overall tone and character of the sound. The person who described this method to me put it best when he said, "You'll know when you've found this middle band because the sound will just jump out at you."
Locating this frequency means you've found the centre of your musical frequency bands and you're past the hardest part.
2 Determine the width of these frequency bands
This next step in the process involves bringing the filter's gain back to an acceptable level and increasing its bandwidth to boost the surrounding frequencies that will be present in nearby musical notes. The settings you apply here are a matter of personal preference and you will determine these by further comparison. For instance, my definition acceptable levels for EQ is no more than +2 to +2.5dB. If you have different ideas on this, go with what you know.
You may continue using the same looped section that you used to find your tuning if there are already some other notes present. If not, you may wish to find another section with a few more that you will be using. I recommend limiting your range for now to around +/- 3 tones (ie: Bb to C to Eb) as anything above this is getting closer getting closer to the circle of fifths (ie: F to C to G) and would better be dealt with using a separate filter.Reduce the gain of the filter to approximately twice the acceptable level - for me, that's about +4dB
Gradually increase the filter's bandwidth to encompass the surrounding frequencies
Once you have set a relatively broad bandwidth compare the EQ's output with dry output
Broaden or narrow the bandwidth and continue comparing until you're happy with the results
3 Boost the frequency bands
Once you've settled on a bandwidth all that is left to do with this filter is to set its gain to a level which you feel is most effective at bringing out the musical tones in the sound. But you will keep coming back to this with each layer of EQ you apply.
Repeating the process
You'll want to repeat the process using additional filters. You would firstly focus on other musical keys present in your composition, then harmonics, and finally working on your upper harmonics.
I would recommend starting by repeating the above process in the context of musical content that is a fifth above or below as these are often the most common key variations used within an arrangement and typically allow you to cover the majority of notes present in your composition's root key signature. For the first harmonic, try by beginning with a starting frequency approximately double the one that you settled with when dealing with your root note. For each harmonic that follows you would logically start with a frequency double that of the one you chose for the previous harmonic. For the high frequencies (say above 5-8 kHz) I'd also try adding a high shelf EQ just above or below the filter you have used for this with a slight boost. Personally, I find this just gives a crisper sound.
Each time you repeat the process remember to ensure that every other filter is off while you are finding your centre frequency as their effect will colour the results. Once you've settled on your frequency and bandwidth you should then reference it with the other filters and adjust your gain structure to suit.
Compare, Adjust, Eliminate
If you're unfamiliar with the process of comparison and adjustment then you should consider the process of tuning a string by ear or matching the beat of a record. Both are similar processes which may give you a better understanding:
When tuning a stringed instrument by ear, you start at the bottom string. If it's in tune, you can then tune the string above it by placing your finger on the bottom string at the point along the fretboard which produces the same open note of the string above it. You then play this note and the open note of the next string and you COMPARE the two. If the open note is not in tune you would then ADJUST it by tightening the string if it's flat or loosening the string if it's sharp. You would then repeat the process until both strings are in tune with each other.
Beat-matching records is also a process of elimination. You start the playback on your cued track and determine if it's faster or slower. Before you adjust the pitch control you'd make a note its current value - let's say it's 4%. Assuming that cued playback is faster, you would then decrease the pitch control -let's say to 2% - start the playback again, and compare. Assume that it's now slower meaning you'd need to increase the pitch. Your next adjustment will be smaller and more accurate because you now know that the correct speed is more than 2% and less than 4%. If you set it at 3% you can then eliminate half of this range as being incorrect.
Music is what feelings sound like
