Since there have been a good few questions on this in the past month or so, I thought I'd write up some basic info on getting your tunes to sound a bit more professional or real. There is no real one way to do this, and some of the coolest sounding mixes are way out of the box. However for starting out, this will be an easy way to grasp the concepts of mixing, and once you have the foundation it becomes it a lot easier to come up with crazy routing schemes and what not.
Sibelius and Finale do notation really well, but really lack the control most DAWs (Digital Audio Workstations) have in terms of editing, audio, MIDI control, mixing capabilities, and so on. You CAN get a great sound out of Sib/Finale, however it'll take you twice as long and you'll prob feel like pulling your hair out before you're done.
If you do not have a DAW and can't sport the money for one of the higher end systems, there are a good few alternatives out there with Reaper and Audacity. Audacity is great for quick audio editing, but it lacks MIDI support. You'd have to record in the live audio from your notation program instrument by instrument if you wanted full mix control, and by doing this you will induce a lot of noise from your sound card. Reaper is a great free/cheap option as well, and has almost every feature that Cubase/Logic offers. However Reaper can be a bit frustrating for someone who is not as versed in audio lingo, and will take a lot longer to setup than something like Cubase, Sonar, or Logic Express.
So in whatever program you choose, import the MIDI file from your notation software and setup your routing with VDL. You can find a very detailed way to do this in Jedi's tutorial here -
Pay attention to how you set your session up. I would highly encourage leaving the sample rate at 44.1khz and putting the bit depth up to 24bits. 24bit gives you A LOT more headroom for mixing and processing, which I will get into later.
Depending on how detailed you want to get, I suggest clearing out all velocity and volume information. Using automation will achieve a more human feel to the mix over relying on set dynamic markings. You can also convert the MIDI tracks to audio if you'd prefer to deal with that instead. Certain types of software will deal with MIDI and audio quite differently in terms of mixing precision.
When all of the routing is set, organize your tracks in a way that makes sense to you. Score order helps down the road.
Once we have everything set with routing and all effects cleared, set all of the faders in the software mixer to unity gain. A tendency is to start moving crap around right away without listening to what is going on first.
Sit down with your score and listen through the entire song a few times though, leaving at least 20-30 seconds of silence at the end to let your head clear and ears relax for a moment. This also gives you time to think about what you are hearing.
I view mixing as making a realistic listening situation for an event that never happened. Even with an album I record with live instruments, generally the band does not play the full song together; I track each instrument one by one. So to make someone believe that your recording is realistic, it helps to visualize what the ensemble looks like and where the sounds are coming from. We also have the luxury of using good samples, which generally sound very good on their own. Keep this in mind as you go through your mix, and I’ll reference this concept more later.
Now we’ll start with getting volume levels and panning. I view volume as depth, and panning as width. If at all possible, avoid using headphones. They do not give an accurate view of levels, panning, and depth. Even if you have crappy consumer monitors, they will help you more than headphones will. Also, be careful with how loud you're listening at. Even at a reasonable volume, constant listening for a few hours will wear your ears down a lot. You're not doing any damage, but like every other part on your body, they do get tired after a while from critical listening. You can hear a lot more at really quiet levels than you can loud. Your ears have a harder time discerning certain frequency areas at high volumes, and certain phenomena cause these areas to be masked. I do most of my mixing at a low speaking voice.
Go through the song and start setting some rough levels with the faders. Don't worry if one instrument suddenly becomes too loud or too quiet a few times, this is to only get a rough outline. When setting rough levels, pay attention to where you are placing the fader. Nine times out of ten, placing something above unity gain is generally too loud. There is only so much headroom that your mixer can handle volume wise, and if everything is over unity, you WILL clip the mixer and induce digital distortion into your mix. In the old analog days, it was desirable to push the meters a bit on a tape machine or console, since analog circuitry usually is much more forgiving to clips, and can add desired harmonics into the sound. Also, you had to deal with the inherent noise floor that tape and analog circuitry induced into the signal, so you had to put a louder signal in to mask the noise floor. Well, with digital we don’t have to deal with any of that, and digital tends to react very unfavorably to overly strong signals. So if you see any clip indicators going off (little red light on the meter somewhere), you need to turn things down a bit. It is better to air on the safe side and leave most of the meters in the green, with them tapping yellow for brief periods. This will become more important later.
Once you have some basic levels set, it helps to group your tracks together if you have a large session. By making mix groups, you can control the overall volume levels of the battery, marimbas, vibes, and so on without having to try and turn up each individual track 40 times over. You can also do group processing this way as well. Consult your manual on how to set up group faders for your software.
I feel that the two most important voices in a song are usually the rhythmic and melodic instruments. If you were to remove the harmony and background parts from a song, it should be able to stand on it’s own in a basic skeleton form. This is why I usually tend to start setting levels for these two sections, because once you have them gelling together, it makes it a lot easier to place everything else around them.
So starting with your rhythmic tracks, get a good overall setting with volume levels. Again don’t worry too much about short sections that need to come up or down, we can automate these later. Continue on doing this for each group until you get a good overall dynamic feel for the song.
Next, we’ll move on to panning. Panning adds a lot of realism, separates each instrument into its own area, and can solve a lot of frequency issues without having to turn to processing. Some people like to pan from the group’s perspective and others from the audience’s. This is personal preference and I suggest playing around with it some. Once you decide on your scheme, start placing instruments where you imagine them to be in your virtual space. For synthesizers, I like to have them placed out a bit wider since they are usually coming from a set of monitors. Keep in mind that you might not want to place anything full panned left or right, since that area is usually best for placing ambient effects such as reverb or delays. Also, it can give the uncomfortable feeling of being surrounded by the ensemble if you place too much out there.
Once you have rough panning and levels set, now we’ll look to any frequency issues that you might have with your song. Lookup how to set your plugins to a Pre-Fader mode. This places the plugin BEFORE the fader (volume controls) in the mixer. Reason being, is that most plugins like compressors rely on input volume to work. That means if the plugin is placed after any automation or volume changes, it will change how the plugin reacts to the incoming volume, and will give you an inconsistent sound.
Remember that there are three major rules that apply to EQ.
1) Use subtractive EQ to make things fit. 2) Use additive EQ to make things sound different. 3) You can't add any frequencies that are not there. There are a wide variety of EQ’s that come with most DAWs. Start out with a 4 or 5 band parametric EQ.
There are a few different bands that audible audio frequencies live in- 1) Sub bass: 16hz-60hz This area is a power area that is more felt than heard. Too much will make you mix sound muddy. 2) Bass: 60hz-250hz Contains the fundamental sounds of the rhythm section. A good blend will make the mix sound fat, too much will make it boomy. 3) Low mids-250hz-2k This area is the low order harmonics of most instruments. Boosting too much in these areas causes your mix to get very horn or tin-like. 4) High mids: 2k-4k Very sibilant area. 5) Presence: 4k-6k Clarity and definition area for most melodic instruments, boosting will make things seem closer. 6) Brilliance: 6k-16k Controls the brilliance and clarity of the mix. Too much will make things raspy.
You'll find that the majority of your problems come from areas in the 200hz-700hz, and from 1k-3k ranges. These areas are buildup ranges, where most instruments’ fundamental ranges are. Remember that a little goes a long way, a 2dB boost is a lot, and usually you only need to cut 1-3dB to make a major difference.
Any problems in the 200hz-700hz range are what you would usually describe as "muddy" or "boxy". Instruments in this area are any type of pad synth or low end keys, bass drums, low to mid octave marimba, low octave vibes, concert bass, bass guitar, low end guitar, low end strings, fundamental snare pitch, and the 3 and 4 drums on a set of quads.
In the 1k-3k areas, problems are usually described as sibilant, or over-present. This is where the mid range pitch of a snare, top basses, upper quads drums, higher end marimba, vibes, xylo, and so on live. This is also the area that is most sensitive to human hearing, and a buildup in this area causes an uncomfortable feeling (try it, open up an EQ and crank on 1-3k and see how your head feels after listening to it for a bit).
So as you listen to each problem area, if things sound funky, open up a EQ. Place it on the instrument you think is causing the problem, then start slowing sweeping the frequencies. Start by boosting about 7-8dB, making the Q (bandwidth) fairly sharp, then sweep the frequency control (slowly turning) from 50hz or so on until you hear something that sounds really bad. That is most likely the cause of the problem. Leave the Q and frequency where they are, then attenuate that area by about 2-3dB. Continue on doing this until the area cleans up.
Next, we’ll touch on compression.
Place any EQ plugin BEFORE any compressor, since you really don't want to bring out any problem areas before they get fixed. A compressor will usually increase the presence of problem areas, and its generally better to remove those first.
A compressor is an effect that limits the dynamic ratio of an instrument. So if you set a compressor to a 3:1 ratio, for every 3db that passes through the compressor will allow 1dB out. A compressor can make an instrument punchier and help you control an overly dynamic part. However, too much compression will cause an instrument to lack clarity, get shoved to the back of the mix, and kill off any dynamic range that was there. Adding a slight amount (like a 2:1 or 3:1 ratio) to marimbas, bass drums, snares, quads, bass guitar, and synths will make them sound and feel much punchier and upfront. Open up a compressor on the bassline track, set a small ratio, and slowly pull down the threshold control until you see the meter starting to pump slightly. I usually like a fast attack and slow release, which causes the instrument to get a slight pulse to it. I view compression as less of a dynamic and sonic tool, and more of a feel type effect.
Now that we have basic levels and any frequency issues resolved, let’s fine tune what we have. Automation is a great way to make dynamic changes with your mix. Using your score, now you can go in and draw your crescendos into your score. Consult your manual on how to add automation. Most programs should allow you to either do it by moving the fader on a live pass, or go in and draw velocity changes with the MIDI editor.
Once you have all of your dynamics fine tuned, listen through the song a number of times over. If you need to make a change, make a small one. Nine times out of ten, small tweeks make a huge difference.
If your mix needs a bit more ambiance and depth, check out some reverb. Like any effect, too much will ruin your mix, and make it sound like you are listening to a song in the bathroom of a club. Reverb does not usually work as well as an insert like EQ and compression do. Reverbs also tend to eat up A LOT of processing power, so instead of using 5 of them, using one will usually do just fine. Reverbs are used to place an instrument in a sort of physical space.
Set up a reverb on an aux track or aux buss. Select an algorithm that you like (medium room, medium hall, whatever) and turn the fader on the track down very low. Next, use an aux send on more transient instruments, such as battery instruments or other drums. Send a little bit to the reverb track and see how it sounds. Slowly add more until you get a good mix. Also, pan the reverb all the way left and right. This creates an aura of ambience. Reverb also works well on ambient tracks, like synths and so on, and can give marimbas and vibes a nice roundness.
Once you have the full mix done, you might notice that the full mix sounds kind of quiet. A Limiter plugin on the stereo buss adds some interesting qualities. A limiter is basically a brick wall compressor, and a little of this will give your song more oomph and volume makeup. Like before, too much will destroy the tune.
Throwing a mastering EQ over the stereo buss (master fader) can help a lot. Boost a tad at 40-80hz, cut a bit around 250hz, cut again at about 700hz, slight cut at 1-3k, slight boost at 8k, slight boost at 16k, and a slight cut at about 18k if things are too bright.
It also helps to have a reference track going at this point. Pick out a favorite recording of yours that somewhat fits the genre and feel of your song. Compare your work to it. This also gives you a good idea of where you are overall volume wise.
When you are happy with your final product, bounce (or mixdown, whatever lingo your DAW uses) down to a 44.1khz/16-bit AIFF or WAV file (Redbook CD standard). Most will include an MP3 option as well, but it works just as easy to convert it in iTunes or something else.
It'll take a lot of patience and time at first. Most of the rock tunes that I mix average about 40 tracks, and I take about 8-12 hours to mix one song. About 6 hours of that is spent listening, and doing very little knob turning on the console.
Legacy Forum Post
Sibelius and Finale do notation really well, but really lack the control most DAWs (Digital Audio Workstations) have in terms of editing, audio, MIDI control, mixing capabilities, and so on. You CAN get a great sound out of Sib/Finale, however it'll take you twice as long and you'll prob feel like pulling your hair out before you're done.
If you do not have a DAW and can't sport the money for one of the higher end systems, there are a good few alternatives out there with Reaper and Audacity. Audacity is great for quick audio editing, but it lacks MIDI support. You'd have to record in the live audio from your notation program instrument by instrument if you wanted full mix control, and by doing this you will induce a lot of noise from your sound card. Reaper is a great free/cheap option as well, and has almost every feature that Cubase/Logic offers. However Reaper can be a bit frustrating for someone who is not as versed in audio lingo, and will take a lot longer to setup than something like Cubase, Sonar, or Logic Express.
So in whatever program you choose, import the MIDI file from your notation software and setup your routing with VDL. You can find a very detailed way to do this in Jedi's tutorial here -
[url]https://www.tapspace.com/support/tutorials/jessem_vdl2_tutorial/index.html[/url]
Pay attention to how you set your session up. I would highly encourage leaving the sample rate at 44.1khz and putting the bit depth up to 24bits. 24bit gives you A LOT more headroom for mixing and processing, which I will get into later.
Depending on how detailed you want to get, I suggest clearing out all velocity and volume information. Using automation will achieve a more human feel to the mix over relying on set dynamic markings. You can also convert the MIDI tracks to audio if you'd prefer to deal with that instead. Certain types of software will deal with MIDI and audio quite differently in terms of mixing precision.
When all of the routing is set, organize your tracks in a way that makes sense to you. Score order helps down the road.
Once we have everything set with routing and all effects cleared, set all of the faders in the software mixer to unity gain. A tendency is to start moving crap around right away without listening to what is going on first.
Sit down with your score and listen through the entire song a few times though, leaving at least 20-30 seconds of silence at the end to let your head clear and ears relax for a moment. This also gives you time to think about what you are hearing.
I view mixing as making a realistic listening situation for an event that never happened. Even with an album I record with live instruments, generally the band does not play the full song together; I track each instrument one by one. So to make someone believe that your recording is realistic, it helps to visualize what the ensemble looks like and where the sounds are coming from. We also have the luxury of using good samples, which generally sound very good on their own. Keep this in mind as you go through your mix, and I’ll reference this concept more later.
Now we’ll start with getting volume levels and panning. I view volume as depth, and panning as width. If at all possible, avoid using headphones. They do not give an accurate view of levels, panning, and depth. Even if you have crappy consumer monitors, they will help you more than headphones will. Also, be careful with how loud you're listening at. Even at a reasonable volume, constant listening for a few hours will wear your ears down a lot. You're not doing any damage, but like every other part on your body, they do get tired after a while from critical listening. You can hear a lot more at really quiet levels than you can loud. Your ears have a harder time discerning certain frequency areas at high volumes, and certain phenomena cause these areas to be masked. I do most of my mixing at a low speaking voice.
Go through the song and start setting some rough levels with the faders. Don't worry if one instrument suddenly becomes too loud or too quiet a few times, this is to only get a rough outline. When setting rough levels, pay attention to where you are placing the fader. Nine times out of ten, placing something above unity gain is generally too loud. There is only so much headroom that your mixer can handle volume wise, and if everything is over unity, you WILL clip the mixer and induce digital distortion into your mix. In the old analog days, it was desirable to push the meters a bit on a tape machine or console, since analog circuitry usually is much more forgiving to clips, and can add desired harmonics into the sound. Also, you had to deal with the inherent noise floor that tape and analog circuitry induced into the signal, so you had to put a louder signal in to mask the noise floor. Well, with digital we don’t have to deal with any of that, and digital tends to react very unfavorably to overly strong signals. So if you see any clip indicators going off (little red light on the meter somewhere), you need to turn things down a bit. It is better to air on the safe side and leave most of the meters in the green, with them tapping yellow for brief periods. This will become more important later.
Once you have some basic levels set, it helps to group your tracks together if you have a large session. By making mix groups, you can control the overall volume levels of the battery, marimbas, vibes, and so on without having to try and turn up each individual track 40 times over. You can also do group processing this way as well. Consult your manual on how to set up group faders for your software.
I feel that the two most important voices in a song are usually the rhythmic and melodic instruments. If you were to remove the harmony and background parts from a song, it should be able to stand on it’s own in a basic skeleton form. This is why I usually tend to start setting levels for these two sections, because once you have them gelling together, it makes it a lot easier to place everything else around them.
So starting with your rhythmic tracks, get a good overall setting with volume levels. Again don’t worry too much about short sections that need to come up or down, we can automate these later. Continue on doing this for each group until you get a good overall dynamic feel for the song.
Next, we’ll move on to panning. Panning adds a lot of realism, separates each instrument into its own area, and can solve a lot of frequency issues without having to turn to processing. Some people like to pan from the group’s perspective and others from the audience’s. This is personal preference and I suggest playing around with it some. Once you decide on your scheme, start placing instruments where you imagine them to be in your virtual space. For synthesizers, I like to have them placed out a bit wider since they are usually coming from a set of monitors. Keep in mind that you might not want to place anything full panned left or right, since that area is usually best for placing ambient effects such as reverb or delays. Also, it can give the uncomfortable feeling of being surrounded by the ensemble if you place too much out there.
Once you have rough panning and levels set, now we’ll look to any frequency issues that you might have with your song.
Lookup how to set your plugins to a Pre-Fader mode. This places the plugin BEFORE the fader (volume controls) in the mixer. Reason being, is that most plugins like compressors rely on input volume to work. That means if the plugin is placed after any automation or volume changes, it will change how the plugin reacts to the incoming volume, and will give you an inconsistent sound.
Remember that there are three major rules that apply to EQ.
1) Use subtractive EQ to make things fit.
2) Use additive EQ to make things sound different.
3) You can't add any frequencies that are not there.
There are a wide variety of EQ’s that come with most DAWs. Start out with a 4 or 5 band parametric EQ.
There are a few different bands that audible audio frequencies live in-
1) Sub bass: 16hz-60hz
This area is a power area that is more felt than heard. Too much will make you mix sound muddy.
2) Bass: 60hz-250hz
Contains the fundamental sounds of the rhythm section. A good blend will make the mix sound fat, too much will make it boomy.
3) Low mids-250hz-2k
This area is the low order harmonics of most instruments. Boosting too much in these areas causes your mix to get very horn or tin-like.
4) High mids: 2k-4k
Very sibilant area.
5) Presence: 4k-6k
Clarity and definition area for most melodic instruments, boosting will make things seem closer.
6) Brilliance: 6k-16k
Controls the brilliance and clarity of the mix. Too much will make things raspy.
You'll find that the majority of your problems come from areas in the 200hz-700hz, and from 1k-3k ranges. These areas are buildup ranges, where most instruments’ fundamental ranges are. Remember that a little goes a long way, a 2dB boost is a lot, and usually you only need to cut 1-3dB to make a major difference.
Any problems in the 200hz-700hz range are what you would usually describe as "muddy" or "boxy". Instruments in this area are any type of pad synth or low end keys, bass drums, low to mid octave marimba, low octave vibes, concert bass, bass guitar, low end guitar, low end strings, fundamental snare pitch, and the 3 and 4 drums on a set of quads.
In the 1k-3k areas, problems are usually described as sibilant, or over-present. This is where the mid range pitch of a snare, top basses, upper quads drums, higher end marimba, vibes, xylo, and so on live. This is also the area that is most sensitive to human hearing, and a buildup in this area causes an uncomfortable feeling (try it, open up an EQ and crank on 1-3k and see how your head feels after listening to it for a bit).
So as you listen to each problem area, if things sound funky, open up a EQ. Place it on the instrument you think is causing the problem, then start slowing sweeping the frequencies. Start by boosting about 7-8dB, making the Q (bandwidth) fairly sharp, then sweep the frequency control (slowly turning) from 50hz or so on until you hear something that sounds really bad. That is most likely the cause of the problem. Leave the Q and frequency where they are, then attenuate that area by about 2-3dB. Continue on doing this until the area cleans up.
Next, we’ll touch on compression.
Place any EQ plugin BEFORE any compressor, since you really don't want to bring out any problem areas before they get fixed. A compressor will usually increase the presence of problem areas, and its generally better to remove those first.
A compressor is an effect that limits the dynamic ratio of an instrument. So if you set a compressor to a 3:1 ratio, for every 3db that passes through the compressor will allow 1dB out. A compressor can make an instrument punchier and help you control an overly dynamic part. However, too much compression will cause an instrument to lack clarity, get shoved to the back of the mix, and kill off any dynamic range that was there.
Adding a slight amount (like a 2:1 or 3:1 ratio) to marimbas, bass drums, snares, quads, bass guitar, and synths will make them sound and feel much punchier and upfront. Open up a compressor on the bassline track, set a small ratio, and slowly pull down the threshold control until you see the meter starting to pump slightly. I usually like a fast attack and slow release, which causes the instrument to get a slight pulse to it. I view compression as less of a dynamic and sonic tool, and more of a feel type effect.
Now that we have basic levels and any frequency issues resolved, let’s fine tune what we have. Automation is a great way to make dynamic changes with your mix. Using your score, now you can go in and draw your crescendos into your score. Consult your manual on how to add automation. Most programs should allow you to either do it by moving the fader on a live pass, or go in and draw velocity changes with the MIDI editor.
Once you have all of your dynamics fine tuned, listen through the song a number of times over. If you need to make a change, make a small one. Nine times out of ten, small tweeks make a huge difference.
If your mix needs a bit more ambiance and depth, check out some reverb. Like any effect, too much will ruin your mix, and make it sound like you are listening to a song in the bathroom of a club. Reverb does not usually work as well as an insert like EQ and compression do. Reverbs also tend to eat up A LOT of processing power, so instead of using 5 of them, using one will usually do just fine. Reverbs are used to place an instrument in a sort of physical space.
Set up a reverb on an aux track or aux buss. Select an algorithm that you like (medium room, medium hall, whatever) and turn the fader on the track down very low. Next, use an aux send on more transient instruments, such as battery instruments or other drums. Send a little bit to the reverb track and see how it sounds. Slowly add more until you get a good mix. Also, pan the reverb all the way left and right. This creates an aura of ambience. Reverb also works well on ambient tracks, like synths and so on, and can give marimbas and vibes a nice roundness.
Once you have the full mix done, you might notice that the full mix sounds kind of quiet. A Limiter plugin on the stereo buss adds some interesting qualities. A limiter is basically a brick wall compressor, and a little of this will give your song more oomph and volume makeup. Like before, too much will destroy the tune.
Throwing a mastering EQ over the stereo buss (master fader) can help a lot. Boost a tad at 40-80hz, cut a bit around 250hz, cut again at about 700hz, slight cut at 1-3k, slight boost at 8k, slight boost at 16k, and a slight cut at about 18k if things are too bright.
It also helps to have a reference track going at this point. Pick out a favorite recording of yours that somewhat fits the genre and feel of your song. Compare your work to it. This also gives you a good idea of where you are overall volume wise.
When you are happy with your final product, bounce (or mixdown, whatever lingo your DAW uses) down to a 44.1khz/16-bit AIFF or WAV file (Redbook CD standard). Most will include an MP3 option as well, but it works just as easy to convert it in iTunes or something else.
It'll take a lot of patience and time at first. Most of the rock tunes that I mix average about 40 tracks, and I take about 8-12 hours to mix one song. About 6 hours of that is spent listening, and doing very little knob turning on the console.