The on-stage guide to Ableton Live

Live 'points' to pieces of files on disk, which is one reason why it's so efficient — if you're playing back a two-bar Clip from a file that's hundreds of megabytes, Live only cares about the two bars. You have two ways to play back this piece of audio: either stream the Clip from disk, or designate it as a RAM Clip. If you choose the latter, Live loads the portion of the file referenced by the Clip into RAM.

As with so many aspects of computers, there's a trade-off involved. Loading too many Clips into RAM reduces the amount of RAM available to your operating system and the program, which can lead to instability issues. But stream too many Clips from your hard disk, and at some point, it won't be able to pull all those files without 'stuttering' (especially if you're using a laptop with a 5,400 RPM drive).

If your Clips consist mostly of short loops (a few megabytes), and you have a Gigabyte of RAM, I recommend converting all Clips into RAM Clips. You used to have to convert them one at a time, but in more recent versions of Live, you can either ctrl-click the various Clips or draw a marquee around them. Select a Clip or Clips, and look at the Clip Overview; the RAM switch field is a toggle. If it has a diagonal line where the upper part is grey and the lower part is yellow, that means some of the selected Clips are RAM Clips and some aren't. Click on the RAM Clip switch, and all of them will be converted to RAM Clips, as indicated by the background field turning yellow. If you click again, the background will turn grey, and the selected Clip(s) will be set to stream from disk.

Streaming is useful for long Clips. For example, in some of my sets, there are somewhat 'ambient' Clips that I want to be able to bring in at any time during the course of the set, and I don't want them to repeat in a way that the audience can notice. These Clips can last 10 minutes or more, and take up over 100MB. I set these up to stream from disk, with the rest as RAM Clips.

However, there's a way to 'cheat' and have all your Clips be RAM Clips, without taking up any system memory at all or stressing your hard drive. I know this sounds too good to be true, but keep reading

Assuming you have a computer with a USB 2.0 port, buy a 1GB (or 2GB if you're feeling extravagant!) USB 2.0 memory stick and transfer your Ableton Live project to it using the 'Collect All and Save' function (found under the File menu). Load the project from the memory stick, select all the clips as described above, and turn off the RAM switch so its background turns grey. Live will think the memory stick is a hard drive, and 'stream' the samples from it. In reality, though, they're being streamed from RAM — just not the precious system RAM in your computer. (Remember to make sure these Clips aren't selected as RAM Clips, or they'll be loaded in your computer's RAM, negating the advantage of this approach.)

There's one caution: if you're using a USB 1.1 stick, this technique may still work, but it will take quite a while for the project to be saved. The total amount of time depends on the size of the samples used in your project. Don't think your system has crashed; be patient. It will save.

The way to obtain the absolute best audio quality is to use loops that are cut to the project tempo. In other words, if the project tempo is 120BPM, then you would want a one-bar loop to last exactly 2.00 seconds. If it deviates from that value, then Live will have to apply stretching (there are multiple stretch options), which has the potential to alter the tone negatively. If there's no difference between Live's tempo and the loop's native tempo, note that you still have to select Warp mode for the loop to repeat; this means you need to choose a stretch algorithm, even if there's no actual need to apply stretching. However, in this case, if you choose any algorithm except for the 'Complex' stretch algorithm, no DSP is applied to the loop, and there is no change in the loop's audio quality.

For very small differences, consider using the Re-Pitch algorithm to stretch, particularly for percussion loops. The DSP it applies is very straightforward and adds little (if any) noticeable coloration to the sound.

To create loops at the Live project tempo as preparation for the performance, I generally import them in Cakewalk Sonar or Sony Acid, and use Acidization techniques to make them stretch as seamlessly as possible to a particular tempo. Once the stretching is perfected, I export the file as a WAV file at the desired tempo. Then when I bring it into Live, stretching is not needed.

Another possibility is to bring a loop into any program with a DSP-based time-stretching algorithm (for example, in Cubase and Sonar, you can ctrl-click on a clip and stretch it to the desired length to match a particular tempo). If the amount of stretching that needs to be done isn't too great, this can give acceptable sonic results. Export the stretched file, then bring it into Live.

Furthermore, it's important to remember that Live offers several stretching algorithms. You can think of the default algorithm, Beats, as cutting digital audio into a series of slices, which are then played further apart or closer together to change tempo. This works extremely well with drum loops that have well-defined transients, but is not ideal for program material. Simply switching over to a different algorithm, though, can produce far better results. Live currently offers five stretching algorithms: Beats, Tones, Texture, Re-Pitch and Complex. Try each one with a loop and choose whichever one sounds better.

And while we're talking about sound quality many Live users use Windows laptops, and the built-in audio on almost all laptops leaves a lot to be desired. Not only is there likely to be hiss, but also there may be crackles from when the hard drive is being accessed, as well as other artifacts. Furthermore, there may not be low-latency drivers for the built-in audio, forcing you to use high-latency protocols (like MME for Windows) or a 'generic' driver, like ASIO4ALL. (This is one area where the Mac laptops offer an advantage; they're inherently designed to use Core Audio, and the hardware quality often seems better as well.)

For the best audio performance on Windows or Mac laptops, use an external audio interface that connects via a card slot, USB, or Firewire. Of these, I've had the least amount of problems with cards and USB; Firewire seems hit-or-miss, particularly with Intel Mac laptops (at least in the past — this kind of thing is always subject to change without notice, as one little driver tweak can make the difference between something working properly or being a major headache). There are several small, reliable interfaces, like the Echo line of card-based interfaces and the CEntrance USB MicPort Pro. I use a PreSonus FireBox because it's compact, seemingly indestructible, compatible with just about everything (including 64-bit Vista) and, most importantly, has MIDI, so I can hook up a control surface to Live and audio inputs so I can feed instruments into Live in real time.

To use external sources with Live, you first need an audio interface with enough inputs to accept whatever signal you want to add. If a preamp is needed, the interface will have to supply any additional application because Live (like most programs) doesn't have the option to raise or lower the volume of external signals prior to hitting the program.

Here's how to set up Live for external audio:

The only real problem with this technique is that if both your hands are busy playing, it's tough to turn off the loop if you want to play through right to the end of the loop. In that case, simply stop playing at the end of the loop, let the loop play through one more time, then stop recording during the final loop iteration. Next, do the following:

You need to load the Pok editing software that comes with the footswitch, assign the footswitches so that they correlate to QWERTY keyboard assignments you made within Live, and then you're ready to go. There are also shift-click and double-click footswitch setups if you need more options than just eight footswitches.

I've found only two limitations: Pok is not cheap (then again, it's sturdy and quite cool — besides, you can use it to punch in and the like with any DAW), and you have to remember to hit the footswitch before you actually want to punch in and punch out. It took me some time to unlearn a lifetime of trying to hit a footswitch exactly on the beat!

However, there is a limitation. A very important part of my act is soloing a track for a measure or so, then having everything else come crashing back in — your basic breakbeat. But Live can't record solo button presses. Ableton's reasoning is that the solo button is a studio diagnostic tool to see what's happening with a specific track, and you don't really want to record that. When I queried Ableton about the possibility of implementing record for the solo button function, they said the solo button functionality is embedded so deeply in the code that there's no simple fix.

One way to get around this and create an accurate audio record of what you've played is to resample the audio from the mixer output, and record it into a track within Live. If the performance is great, then you can simply use this audio. But if not, you can use the audio track as a guide to carve away the unwanted audio in the breakbeat sections, then proceed with any additional editing you want to do in the Arrangement view. For example, if I'm listening to the audio guide track and hear there's a breakbeat, I cut away the tracks in the Arrangement that aren't supposed to be playing and move onto the next breakbeat. While time-consuming, this does create the same end result as if Live had recorded the solo button presses. Here's how to set up the audio track for recording your performance output:

As you play, the audio will be recorded in the track you created. Remember to solo the audio track so the other tracks whose moves were recorded don't play back at the same time.