Using a Bus to Route Effects in Ardour

When using computer audio effects, one must always bear in mind that your poor computer has to calculate the effect output everytime you play back your tracks. So when you go too crazy with effects, your computer will probably seize up, roll-over and die! To save your computer, it is important to use buses to route repeated effects in Ardour.

Okay, maybe that sounds like complete gibberish... but please bear with me and you'll soon understand what I mean. Let's take an example of a multichannel recording of a drumbeat in Ardour (i.e. one channel for snare, one for the kick drum, etc.). For each drum, you would want to add certain effects like an EQ, reverb, and compressor. One way to do this is to add the effects to each channel like in the figure below.


Since you would want the same sounding reverb on each drum, your poor computer has to calculate the same thing multiple times. This is unnecessary because you can create a bus and route all the drum channels to it and then only put one reverb on the bus. This method achieves the same sound whilst your computer does less. You can do this by following the steps below:

Firstly, you need to create a new bus. Go to Session, and click on Add Track/Bus. Configure the window that appears to look like the one shown below. i.e. select Busses and Stereo.


You now need to route all your drum channels to the bus (I have renamed mine to be DrumBus). You do this by right-clicking on the black rectangle below the volume fader (this is called the "post effects" for the track). A menu should appear and you must select New Send... Configure yours to look like the one shown below. You do this by firstly adding an output (click the Add Output button) and then selecting where you want the output to go. In this case you choose DrumBus/in1 and DrumBus/in2 from the Ardour tab on the right of the window.


Since you have rerouted the output of the each drum track, it no longer needs to be connected to the Master Bus. So disconnect it by clicking on the Output button (just above the Comments button) at the bottm of each drum channel and select Disconnect. Your drums should all be sounding through your Drum bus. This means that if you adjust the DrumBus volume, all the drums' volumes should change. It also means that any effects you use on the drum bus will be applied over the whole drum kit. The figure below shows how the bus greatly reduces the amount of reverb effects used.


Note that I have left the EQs and Compressors in each channel and not on the bus. This is done because the EQs and Compressors are set differently on each drum and thus cannot be applied to the whole kit via the bus. I hope you found this tip useful, and that your computer will also appreciate it :). A special thanks goes out to Jakob Lund for suggesting this method of adding multiple effects.

Other posts you might find interesting:
Audio Recording in Ubuntu Studio - Part 3: Adding Effects to the Ardour Drumtrack
My Favourite Linux Audio Effects
An Overview of Compression
Panning for Gold : How to Pan Tracks
Mastering Tips: A Picture Says a Thousand Words

Sound Mixing Tips: EQ vs Volume

Here's a tip for live mixing and mastering, but can probably be used in any area of audio recording: If you can't hear an instrument, it's probably because of its EQ - not its volume. Before you push up that volume fader, try boosting the instrument's defining frequencies.

Here's an example. I was mixing for our church, and the band was playing one of those numbers that relies heavily on the keyboard strings. The mix was fine otherwise, but you couldn't hear the strings. The desk I was using had two mid-frequency pots, so I boosted the 250hz (low frequency) one until I could hear the strings. This brought out the full presence of the strings.

Increasing the volume wouldn't have had the same effect. Each instrument has a frequency that defines it - especially in a mix. For strings on a keyboard, I've found that frequency at the low end. Not surprisingly, it's the high end on acoustic guitars (usually 7-10khz and up).

Vocals are trickier, a whole subject on their own. I find that each person has their own 'peak' area that clears up their mix. Males are around 3-5khz and females at around 4-7khz.

To find the defining frequency, create a new peak and boost it until you can hear exactly what that frequency you're dealing with. I usually go full ball here. Then sweep it up and down the frequency spectrum until you hear the instrument come out. Then take the peak down to a reasonable level.

Other posts you might find interesting:
Speaker Placement
High There?!
Panning for Gold
Live Sound, Monitors and Pepper Spray
Audio Recording in Ubuntu Studio - Part 1: Plan your Project

High There?!

Do your recordings lack definition? Are the drums hard to hear with everything else playing? Is your mix a giant mess? Are you feeling depressed?

Well, I may have an answer to help all of the above (except the last one). First, let me ask you: do you use your high-end frequencies (10 k - 20 kHz)? If not - you definitely want to pay attention. If you do - you can add your comments and help the poor sods I mentioned in the previous sentence.

The average person's hearing ranges from 20 - 20000 Hz. So, when you see the faders on your EQ going up to 20 kHz - don't ignore the last ones! This is why: the frequencies above 10 kHz give definition and add a much needed 'sparkle' to certain instruments. Also, if you use the whole frequency spectrum, you can create a feeling of space or bunching. Another way of doing this is by using panning, but these two techniques should be used hand-in-hand. For further information on panning see my post: Panning for Gold.

Cymbals' frequencies are typically between 2 k and 20 kHz. Thus, if your drum track has the highs cutoff, they will lack definition and will not stand out. Other instruments which require a 'sparkle' factor are acoustic guitars, pianos and vocals. Vocals sound best when boosted in the low-frequencies (100 - 200 Hz) and high-frequencies (10 k - 20 kHz). Boosting the mid-frequencies (800 - 5000 Hz) can actually make the vocals more muddy and this is a common mistake.

To conclude: we have learned that the high-frequencies are important - don't forget them or reduce them on your EQ. It is for this very reason that condenser mics are better than dynamic ones. Condenser microphones have a better high-end response and thus have more DEFINITION.

Panning for Gold

Hey guys, sorry I've been away a while. To ease my guilt, I'm writing a post about panning with an excessively cheesy title.

Panning is exceptionally important for allowing each instrument in a mix to be heard without detracting from the song itself. It can also be used for funky effects, which can make one dizzy, or for highlighting certain sounds. It is easiest to understand panning visually. Please meet my visual aid buddies in the figure below: Burt, George, and Jim. They will be dancing around the stereo plane to illustrate my point.


Refer to the figure below. If all my visual aid buddies are placed in the center of the "stereo plane" (i.e. equal left and right volumes), they are very difficult to see. Luckily each of my buddies looks a bit different. If they didn't look different, one could be hiding behind the other. This is similar to a piano sounding different to a guitar. When they are played with the same pan setting, you can still differentiate between them but when two guitars are played with the same pan setting, it will sound either like a mess or as if one guitar has a chorus effect.

In the figure below, two of the buddies are panned close together, near the center. They are both quite visible but are blending together well. The green guy looks as if he is a super-geek or smells really badly because he is standing out. This is exactly the same idea for panning different instruments... the guys in the center crowd are gonna have the bigger punch, but unless they are slightly spaced out, they will just be a mess. The guys who choose to stand out will have less punch but more definition. However, I warn you that if the nerd stands out too far he can sound really bad. Also, if you space everything out too much and don't have an "in-crowd", you will have no punch and basically land up with a bunch of bad-smelling nerds.

Yes, that's right... the secret of panning is merely peer pressure. I'm sorry if this explanation is really childish. To make up for it, here are some real tips:

• Start with your rhythm instruments... i.e. drums and bass.
  
• Mute everything else and setup your drum's panning similar to as if you were sitting behind the kit... i.e. your snare and bass close to center, cymbals on far left and right, and toms going from left to right.
  
• Next unmute the bass and slowly move the panning fader from far left to far right. You should hear where the punchiest point is. There is no solid rule - you simply have to find it.
• Next unmute the guitar and repeat the same process as the above step.
• Continue this method through all the instruments and then move on to the vocals.
• The lead vocals should always be near the center and background vocals shouldn't be panned too far to the left or right unless you are trying to be arty.
• Make slight adjustments until you are happy.
• A word to the wise: be very careful when panning anything too far to the right or left, no one really likes nerds :)

Panning is useful for both live and recording mixes so don't neglect it if you have the facility. You're the sound-guy... be creative.

Other posts you might find interesting:
Monitor/Speaker Placement
Sound Mixing Tips: Bringing an Instrument Out of the Mud
What Gives a Guitar its Tone?
Sound Mixing Tips: High There?!
What Exactly is Mastering For?

An Example of Loud Mastering

Okay, so you may have read my post called: To Limit or not to Limit. Well, I´m afraid I´m going to have to admit I was slightly wrong in that one. However, it still has some good points about limiting too harshly. Let me explain how I saw the light...

Below is a screen shot of a song I recorded. Take a listen Ain´t no Rock Old.mp3 (This is the same as the one posted in To Limit or not to Limit)

For this song the limiter in JAMin was bypassed. This allows the peaks their maximum amplitude, which is good because no sound quality is lost. The problem is that the track is far quieter than any commercial song. Trust me, this is a big problem - no one is going to take you seriously with a quiet track.

Below is a picture of the same track mastered loudly using the limiter to reduce the higher peaks. Take a listen to this track: Ain´t no Rock New.mp3

So in this track, as you can hear, it is much louder and competes with professional recording volumes. The cost: sound quality... As you can see the peaks have been hard limited to bring the big blob of sound closer to 0 dB. This means, your song can become less punchy. This is because the peaks are mostly from the bass drum and other percussive sounds. However, you regain punch in the fact that it will be played louder.

It´s a real toss-up and mastering professionals will debate this till the cows come home. I will make your decision easier, by saying that if you do not go for the louder mix, your songs are just not going to stand up to professional tracks.

Alrighty, with that all said, how do you get the louder mix using JAMin. This is where I apologise, you MUST use the limiter. Take a look at the screenshot of my limiter setup below.

You can see that I set the boost to the maximum - remember that we are trying to achieve a LOUD mix. Next, and most importantly, the Limit is set to -0.2 dB, this is the value at which the peaks will be cut off (you can play around with this, but only between values of -0.3 dB to 0 dB). I set the Release to round 400 ms, you want quite a long release otherwise the sharp peaks will not be cut. Lastly, we fine tune the limiter by adjusting the Input level. We do this by observing the Output level meter on the right-hand side of the JAMin window. You want your loudest parts of the track (i.e. chorus) to stay between about -2 dB and -0.2 dB, without hearing any distortion in the slightest. Make sure your track does not merely sit on the -0.2 dB mark - this means the limiter input is too high and you are limiting too much!

That´s the gist of it... give it a try and please don´t forget to rate this post.

A Picture says 1000 Words

As audio artists, we tend to rely on our hearing to determine how well a mix is coming out. This may work, if you understand the basics of mixing and mastering, but whilst you´re still learning, seeing if a track looks right can help you more. Below is a picture of one of my earlier recordings...

It looks alright, it has what one expects, a few sharp drum peaks with lower hanging notes. But when you compare this to a professional recording like the pictrue below, taken from AFI´s Love Like Winter, my recording is like taking a pocket knife into a sword fight.

In this recording, the track has been mastered loudly, with a large percentage of the song between -2 to 0 dB. Note however, that there is no clipping of the bulgy melodic parts, only the drum peaks have been hard limited.

So, I encourage you to plug your favourite band´s best song into Audacity and see what it looks like. What remains is how to get your songs looking more like the ones you like. For this and more...tune in next time!

Normalisation...never heard of it!

Normalisation (or normalization for the Americans), is a very useful tool when mastering or mixing down tracks. As I´ve said earlier, mastering aims to achieve the loudest possible mix without losing too much sound quality. Normalisation helps us achieve this by doing the following. Firstly, it finds the highest peak in a track. Secondly, it raises this highest peak to the loudest it can go (normally 0 dB or 0.3 dB/98 %). And lastly, it raises everything else in the track to a point which is relative to the highest peak. Thus, normalisation creates a louder track without affecting the signal-to-noise ratio [1].

This is how one can normalise tracks in Ardour. If you look in the Edit menu of Ardour, you´ll find a dropdown called Region Operations and inside you´ll find a Normalize Region command. The shortcut for this command is `N´. So select the track you wish to normalise, then simply press `N´. Your track has now been normalised. You might even see the track peaks change a little. I recommend you do this on all your recorded tracks to achieve a louder, more professional mix.

[1] Normalisation Wiki Entry: http://en.wikipedia.org/wiki/Audio_normalization

What Exactly is Mastering For?

This one goes out to Post Paint Boy...

If you are new to home recording and your tracks sound unprofessional and quiet, this is most probably due to either bad or no mastering. Mastering is essentially a process which concerns the output levels of your final musical piece. In particular, it's about trying to get your recordings to an appropriate listening volume without sacrificing the sound quality too much. It is also about getting different tracks sounding good on one CD.

You might be thinking, 'Why can't I simply turn up the output volume of my recording to get a louder track?'
Well, the answer is that by doing this, you are bound to cut off some of your peaks and create some distorted parts in your song. Also, without using mastering techniques, you are bound to get inaudible verses and painfully distorted choruses. You want a song, which your listeners can hear with ease from start to finish.

There is a fantastic article about mastering here, this is based on an even cooler article by Rip Rowan which seems to have gone missing. It is most definitely worth a read as it explains the basic idea behind mastering and how to achieve a louder overall mix. Essentially, you can achieve your louder mixes by reducing your peaks with compression. With your louder peaks reduced, your track can be boosted without saturating the output. Thus, the quieter bits become audible whilst the louder bits remain loud.

Other posts you might find interesting:
To Limit or not to Limit
Interfacing JAMin with Ardour
An Overview of Compression
Normalisation...Never heard of it!
An Example of Loud Mastering

To Limit or not To Limit

When mastering, is limiting a good or bad idea? Here are my thoughts...

I recently recorded some songs, they sounded fantastic in Ardour but when I mixed then down I found then slightly unsatisfying. Let me explain why... When mastering with JAMin, a limiter is activated by default. Essentially what this does is it takes all the louder peaks of your music and just chops them off. This can be seen in the figure below.

This made my music sound... well...crap. As you can see in the picture, whenever the music got louder like in the chorus, some of it was chopped off, creating slight distortion. The solution to my problem was to simply bypass the limiter (there is a button for this in JAMin). This yielded an undistorted, more pleasing recording as seen in the figure below.

One more thing...please understand that limiters have their purpose and can help you out in a tight spot, but never limit your music drastically - it is simply a bad idea.