Guitar | Bass | Keyboard | Microphones | Mixers | Audio Interfaces | Monitors | Sequencers | Soft Synths | Live Sound | Drums | Club | Accessories | Blowouts
mixing "in the box" and "out of the box" in the
|This is the Mackie 1402. Note that the ALT 3/4 bus (the third fader from the far right) is routed to the audio interface. By pressing the ALT 3/4 button above each fader, that channel is routed direct to the audio interface.|
Connecting the board back to
All recording mixers have at least one extra bus, in addition to the L/R bus (the main outs). On small mixers this is called the ALT 3/4 bus. On a 4 bus mixer, it may be called Bus 1/2 and bus 3/4. On an 8 bus mixer you may have 1/2, 3/4, 5/6, 7/8. You don't need to connect all your busses back to the audio interface. Just one will do in all but the more exotic configs.
Important Note: 0dbVU (where the green LEDS turn amber in a hardware mixer) and 0dbFS (full scale) where the LEDS turn red in a software mixer are NOT the same and they do not and should not "match". 0dbVU is an electrical measure of the analog signal--it is the level where the signal is at is optimum.
You still have headroom over 0dbVU in a hardware mixer, sometimes up to +10 without any audible distortion (even higher in quality mixers). However, 0dbFS is the highest point of the digital scale where a signal can be measured. It is technically impossible to exceed 0dbFS. To understand this is to understand the critical relationship between analog and digital signals.
You set the mixer trims so your instruments peg around 0dbVU on the analog mixer. You take the output of the channel to the audio interface and record it as is. You do not need to boost the signal so it gets near 0dbFS on the software mixer. It may come in significantly lower. That is OK. When you send it back to the analog mixer after processing it will come in at a good average level. You don't need to normalize the waveform or boost it. You can adjust the trims to make the peak (highest) level coming out of the daw get back near 0dbVU in the hardware mixer. It should not be far off. While recording, instead of thinking "how do i make the level hotter?" think of "how do I make sure I have plenty of headroom for processing later". Don't push the level.
Role of the software mixer. The software mixer is not the final mixer but a "pre-mixer". That is, it treats the signals with virtual compressors, eq, fx for each internal channel and outgoing bus. This mix will contain 1. all of your audio tracks, 2. all of your soft synth and soft sampler tracks, an optional effects bus. At the hardware mixer, you "fine tune" these busses with another layer of eq (if necessary), and can use sends and returns to add bits of reverb and other effects. The goal here is simply to add polish to tracks that should already be sounding pretty good coming from the software mixer.
In the software mixer you will send different instruments to different output busses. For example, you might send all the vocals to bus 1/2. On the software bus, you will probably add reverb, maybe patched after a slight delay. When it arrives at the hardware mixer, you touch up the eq, maybe send just a tad to a hardware reverb on the sends and returns, just for a bit more depth. Continuing on, all the soft synths and sampler might be broken down into those you want dry and those you want effected heavily. In the software mixer, send your dry stuff to bus 3/4. That might include bass, drums. Bus 5/6 might be synths. 7/8 might be wilder effects like stuff heavily laden with plugins. When 7/8 come into the board, you can further ram it down the effects sends to even more FX processing to get some really unearthly sounds. Or if you have a bus leftover, consider dedicating it just to a kick drum. How you divide the pie is up to you. Of course, it depends on the piece.
Note: You can also configure what is known as a "stem mix" this way. You might have 50 tracks in your sequencer and render these to 8 stems or groups. Here you would mix down each stem in the sequencer so you have 8 audio files feeding your 8 mixer channels. A stem is any meaningful grouping of tracks that you want to level as a group. You could have a drum stem, a vocal stem, a rhythm section stem, etc.
Using this approach, level setting is not critical in the software mixer, you can go with your ears with what sounds good. If you are getting audible distortion, notch stuff down a bit. The Master out fader of the software mixer can be anywhere, it does not have to sit at at the 0db marker like it does in the software only mixer configuration. What you have to do to get the best sounding signal at your hardware board is to solo each channel and adjust the trim so the peak of the audio hits 0db This assures that the board will add minimal distortion in the mix. While you are soloing these channels, listen carefully for distortion coming from the software mixer. A few "overs" on the software mixer are not critical. You can probably get away with 3-4 db before you start hearing distortion. Do this for all the channels. Naturally you reference your hardware synths the same way so they peak at 0dbVU. Then you move the hardware faders till it locks into a great image, or at least close to one. Then comes the back and forth process of subtle tweaks on both the software and hardware mixers as you add the touch of beauty to your creation.
The fun begins when you press start on the sequencer. Your fingers nudge faders, turn up effect sends, gently move the master fader. I like to apply a number of cool tricks, such as pulsing the faders to the beat (moving faders a tiny amount up and down to the beat to give it "drive"), opposite fades (sweep one channel out and you glide another in, creating "distance" by turning up the reverb send as you fade out, creating "blooms" by fading up and turning up an effect just slightly later. These are things the hand can do intuitively that takes lots of practice to do with automation. The thing about real time mixes, for those of you who never did one, is that you don't always get it right on the first pass. You do it over and over again and each time you strive to make it better.
You can mix to a high resolution digital recorder like an Alesis Masterlink, DAT, tape decks, reel to reel or back to the analog inputs of the audio interface if your computer still has enough CPU to do it. However, if you go back to the same computer you will be stuck at the same sample rate and bit depth. A second computer works well, and can allow you to record at 24/96.
The final level of the mix as it re-enters the digital domain will be down significantly from 0dbFS. If you output 0dbVU from the hardware mixer your software level will be somewhere between -14 to -18 dbFS. This is normal. You can boost it a little more at the board if your board has good headroom, but leave some cushion for mastering later. If you want to master while mixing as some of us in a hurry like to do, you would patch a mastering processor over the outputs and run that to the inputs of the audio interface.
- On channels that hold your mics and synths, all you need to do is press the ALT3/4 bus button to pipe the output into the audio interface for recording.
- Your final mix is tweaked at the board, on real knobs and faders, with real eq and effects processors
- Software mix levels are not so critical to get technically correct
- The mix is analog. You avoided errors caused by digital summing. It can be sweeter, smoother and warmer than a purely digital mix.
- The mix has the human touch created by your own hands.
If you have a hardware mixer, (not required) you may set it up as follows. As before, your ALT 3/4 bus connects to the audio interface or soundcard. Your main outs connect to your monitoring system. We're going to assume this is a 12 track board, for simplicity, but it will also work with massive consoles. Connect the Outs 1-2 of the audio interface to hardware mixer channels 11/12. We are only going to use this bus for monitoring so you can hear what's going on in this all digital, software mix. Put the trims on these last 2 channels at their line level rated setting (-10, usually) Set the faders at 0db and the pans 100% Left and Right. You will never touch these again. Channels 1-10 hold your mics and hardware synths. For each of these channels press down the ALT 3/4 bus button. That's right on ALL of them except ch 11-12 which are the only two channels going to the main mix on the hardware mixer. What this does is it routes all your mics and synths direct to the audio interface where they can be controlled totally by the software mixer, and recorded as audio tracks. If you want to use your outboard effects at this part of the process you can remove them from the returns and patch them into unused channels.
Side Note: If you have a full featured "expensive" hardware mixer, you may be able to route the FX returns to other busses with a switch and get them to the software mixer that way. But on most small mixers, you will have to bring them into unused channels that have the ALT 3/4 switch.
Those of you with 8x8 audio interfaces and 4 or 8 bus hardware mixers will benefit from being able to send different groups to different audio inputs. The person with just a small 2x2 audio interface has no option other than recording all tracks as audio tracks. That is all your synths are going to have be recorded as audio, as well as all your audio tracks. Those with at least 4 audio inputs can use input 1/2 for recording and use input 3/4 as an audio-thru channel. In Logic 5 for example, these are called input objects. What the input object does is allow you to route your midi controlled synths through the audio interface. This allows you to control the volume of these synths directly in the software mixer without having to record them as wave files. You generally raise your alt 3/4 fader on the hardware mixer to maximum and set the levels of the individual hardware channel faders all at 0db.As the audio comes through the interface, you control it with the fader on the input object. One fader for my rack of 10 synths? Are you mad?! Hehe. Right now you are probably scratching your head wondering why the Tweak would say this logic defying rule. Stay with me. Because we are using the software mixer, we are controlling each MIDI track with midi volume and pan controls. Remember, the software controls are the important ones.
OK, time for a question. You, digihead, in the back...
DigiHead: But Tweak! How can you add plugins and effects to all those midi tracks coming in one bus? It'll all sound like mush!
Tweak: Your absolutely right. Gold star for you! You can't. When you want something to be effected by plugins you record it to a seperate audio track and plug away. But some synths have good sounding FX out of the box. You can let these stay analog all the way through till the final bounce.
Here your hardware mixer is the sub-mixer or pre-mixer. The software mixer is the final mixer everything goes through. All the audio tracks, each with its own plugins, all the midi tracks, controlled with MIDI controllers, all the soft synth/sampler tracks, all the software buses and input objects are all controlled on the computer screen. The final mix will not go to an outboard recorder (well, it can if you want) but will be a "bounce" a rendering to a single stereo wave file. The hardware channel controls just sit there and pass gain, you never need to touch them. You can use your external FX as before on your hardware mixer. As you get close to mix down stage you might want to record several tracks as audio. When you want to record MIDI to audio you simply mute all the midi tracks except the ones you want to record (or use solo in the software mixer). Once you record to audio you can delete or permanently mute the MIDI data. (I put it in a muted folder in logic).
|Logic Control has automated motorized faders that are touch sensitive|
The control surface is not a mixer per se. Audio does not go through it. What the control surface is is a midi controller that controls you software mixer perfectly. You can control any MIDI track, audio track, audio instrument track and most plugins with automation you can record in real time, and later edit on the screen. The Control surface matched to a sequencer gives exacting control over all MIDI and audio sources. It is not required to have one doing this software mix configuration, but it is nice. Your control surface firmware interacts with the software so when you move a fader on the surface it moves on the screen and vice versa. Thankfully, there are only 2 cables to patch this up, just MIDI in and Out cables connected to you MIDI interface. The cool thing about control surfaces is they can get deep inside the software mixer easily and you have a real knob and fader to move. Even if it only has 8 fader channels you can control a hundred tracks if you want. There is no pre-level setting here, it's all done in the software mixer, no calibration routine. Inside the host application you will find a set of preferences. That's about it.
|Mackie C4 Pro Plug-In And Virtual Instrument Controller|
|Simply put, Mackie Control Pro Series controllers give your music production software what it needs to feel complete.|
You might notice that when your analog output of your mixer reaches your DAW's software mixer the levels are "off". This is because the analog mixer will reference 0dbVU while the software mixer will reference 0dbFS (FS stands for full scale). These two measures are different animals. Don't confuse them as being the same thing! 0dbVU is an electrical measure of signal strength while 0dbFS is the loudest sound a digital system is capable of. It is normal to send an 0dbVU signal to a software mixer and see it come in at around -18dbFS. That is the optimal level to record at--you don't have to boost the analog level so it gets near 0dbFS in the software mixer. To do so will push the analog circuitry and the converters to the breaking point, and less than good sound will result. Instead, record at what may appear to be a lower level and enjoy the headroom you have for software compressors and processors.
Now we get to the beef. The tricky part of the all digital mix, that is, How the heck do you set levels without trims and a 0dbVU reference? Because each audio track has plugins and eq, which changes the gain structure of the track, there is no useable reference except one. Rule #1 No audio track should ever exceed 0dbfs. That's right, never. In fact, it should not even get close to making the overload indicator light up. If it does, you need to notch down till it is within the 0db rule. You may be recording too hot. Bbbbuttt..I can here you guys sputtering in disbelief, sometimes the level can barely be heard if the peak is under 0dbfs it sounds WIMPY!... Answer to this madness: Your overall recording levels are probably too loud and you ran out of headroom. You do not need to record tracks to make them "as hot as possible". That's why you use software compressors, to limit the range of the peak and bring up the overall perceived volume. So. Which tracks do you think should be the hottest after processing? You guessed it, the loudest, most important ones. That may include vocals, kick and snare, and bass. The rest need to be down significantly lower. They might be peaking at -20 for rhythm guitar, -25 for pads, -30 for background instruments and this is after compression, eq, and plugin channel FX. Essentially, you set your loudest tracks first, then you bring up the rest till they sound right. You will find that by simply recording at the average levels your hardware mixer sends you will have the headroom you need in the digital domain to use as many processors as you want and still be well below 0dbFS, the line you must never cross, ever.
Do NOT automate yet, we save that for near the end. Rule #2: Do automation as late in the game as you can, after you have all the levels and all the plugins tweaked to your satisfaction. You will save lots of time this way as you will only have to automate once.
Note the conservative fader settings in Logic on a recent piece. As you see the main outs show a peak of -2.3 was reached. If I were to commit to the mix at this point I'd have that much headroom in the final file, which would be a good amount for later mastering. Note that none of the channels get anywhere near 0db.
In a software mixer like Logic has, you have the option of using as many busses as one could conceivably want. If you have an 8 output audio interface, here's some cool things you can do with busses. If you haven't used your external FX boxes on the hardware mixer yet, you can connect them direct to the audio interface out 3/4 or 5/6 or 7/8. You can use the software bus and assign to it the audio interface out. Route what you want, take a send off the vocals and run it out through the bus, through out 7/8 to your lovely Lexicon prized reverb and bring it back in audio input 5/6 which you can assign to another input object. No Lexicon? Ok use the Nanoverb! You might be surprised, even the lowly NanoNano sounds "right up there" with some software reverbs. And the absolute cool, to die for, thing about this approach is that the sends to the Nano are all fully automatable as are the returns coming in the input object. So, go crazy, slap a stutter LFO plugin on your nanoverb when it's distorting like a buzz saw in a closet--oh wild and wicked tweaks are possible and they don't ruin the mix, or stick some old analog delay in sample loop mode with the significant other screaming "take out the trash, dear, take out the trash dear" and just fade it in/out when you are really bored.
Uh, class, that was a joke. (muffled mirthless laughter)
With busses you can also use them as internal groups. Say you have 5 vocal tracks. Instead of routing them all to the master out, you route them to, say, bus 1-2 and route bus 1-2 to the master out. On bus 1-2 you have one fader that controls all the vocals. Use bus 3/4 for all the drums. Use Bus 5/6 for all the soft synths. Use bus 7/8 to go out to analog (and back as I described above). Use bus 9/10 for all you input objects still streaming audio from the hardware mixer. You can put compressors and limiters and effects on any or all of these, and I highly recommend a limiter on each, if your CPU has room. Same rule as before. Nothing goes above 0db Ever! Got that? Never! Not a single peak indicator should flash. As we prepare for the final mix, we do tiny adjustments to these buses, the most critcal adjustments you will make. Here you go back and forth between the plugins and level faders for each bus, getting each exactly right, adjusting the compression and limiting.
Now is the time, you tweakophiles (smile, sorry had to use that). Do it with the mouse, do it with the control surface, however you do it, you watch the level meter as you go. If you pop in the red, adjust it till it is not in the red. Remember, if this was an analog mix you would be being very careful, right? Be careful here too. Tiny fader movements make a difference. If you must automate your core kick or bass, be very gradual or you will upset the balance you worked to achieve. However, for effects and wild stuff, get as crazy as you want. Make the panner revolve at 78 rpm, just keep the output under...what was that? I can't HEAR YOU (the class chants..."0db, 0db, 0db!!" Tweak: Yes! Rock On!
Note: Loudness optimization is something that is done with limiters at the mastering phase. You don't do that during the mix. Some engineers recommend having the loudest peak of the mix down -6dbFS, -12dbFS or more to allow the mastering engineer some room for processing. Others will let the mix climb as high as it can without exceeding 0dbFS. Everyone agrees you cannot let the signal exceed 0dbFS even momentarily. I say you are best off leaving a cushion between your highest peak and 0dbFS. Why? Well, number one, you can't trust software meters in most sequencers. They might not catch a superfast transient that breaks through the converter's limit. Number two--you can always raise the level at the mastering phase to make the song as loud as possible, if that is what you want. You can always add loudness; you can never remove damaged audio. Don't fall into the trap of recording the final mix level "as loud as possible". Its a mistake.
Universal Audio UAD2 Solo Flexi DSP Card (Macintosh and Windows)
The UAD2 Solo Flexi combines
Universal Audio's revolutionary new DSP acceleration hardware with a
customizable selection of UA's premier plug-ins. If you need the best
sounding collection of plug-ins for your DAW but want the flexibility to
"pick and mix" and build your plug-in collection gradually, the UAD2
Flexi Family is for you.
If you are planning to master later, this is the point where you bounce to stereo. It is probably better to do a non-realtime bounce here, to give the processors plenty of time to do a good render.
Bounce the Master out to wave, or if you want to go to an external digital recorder instead, route the master out to s/pdif and connect your recorder to the s/pdif out. If you are going to apply mastering processors after the mix is bounced, bounce to a higher resolution file format, such as 24 bit/96khz. Then just set the song in the locators, press bounce, type a file name and go to the refrigerator for the beverage of choice. Because you set the levels right, when you come back, you will have a sparkling wave file with the perfect gain optimization, ready for mastering to cd, mp3 or medium of choice.
You might decide to master while mixing if you are doing a quick and dirty bounce to MP3 or if you just want to work that way. For best results, you should take your final mix to a mastering engineer and let them tweak your final master. But I know may of you want to do it all at once, so these notes show you how.
The master output channel fader (not the signal) sits at 0db and you never touch it, except for fading out at the end of the song. That is it's whole function. It is important to keep it exactly at 0db. Here's why. On your master output channel you want to patch in a limiter. A brickwall limiter like the Waves L1 UltraMaximizer or UAD-1 Precision limiter is perfect here. You tell the brickwall limiter where to set the wall. Typically at -.02db. If the limiter is good, no matter what you do the final signal will not exceed minus .02db. Of course if all your tracks are flashing red, it's going to sound like utter garbage, but it will still be within limits. If you plan to master the final wave file later, then do without the limiter by bypassing it, but use it's level meter so you can see if you are going over 0 db. It is essential that you do not go over 0db no matter what. Your final wave file will be ruined if you do. Now it should make sense why I have said this was a rule from the very beginning. The only way to make sure your overall song has bandwidth to breathe is to make sure no red lights are a flashing. If you did your automation too early, you will probably find you have to go back and lower the volume of ALL your tracks globally. See. You should have waited. Add at least an hour to fix this mistake.
It depends on the song and where you are at with it. I start all my songs in method No.1. When i am halfway through, I decide if I am going to go to digital or analog mix. I have my board patched so I can change over to either mix style just by pressing the buss buttons. Typically, if I just want to get a song out quick and dirty I will use the analog mix method. If the song is complex, I will go all digital. But it really doesn't matter. Work the way you like to work. A good mixologist will be able to make an analog mix sound as tight and digital as they want, and a good plug-in head can make their all digital mix warm and smooth like a reel to reel. Mixing, particularly digital mixing has a large learning curve. I hope I saved you some time and frustration with the above tips.
Mix On! Into the Wilderness of the Audio Uncharted.
Go to the Next Class
Go to the Previous Class
Want to discuss this article?
Join the discussion in the Studio-Central Topic dedicated to this article
"It is not difficult to compose, but what is enormously
hard is to leave the superfluous notes under the table."
Johannes Brahms (1833-1897); German composer.