How to record 1 track seperately while listening to another?

I googled around and found what I thought wouldl be the answer to my question on the audacity FAQ. What I want to do is to listen to one track, while recording a new one, but without the first track recording into the new one along with the new audio. When I followed the procedure as detailed in the FAQ (under “multitrack recording”), I could hear the first track, but it was recording into the new track along with the other, which isn’t what I want. So, is there any way to get it to work the way I want it to?

I’m using Audacity 2.0 (installed from the .exe) on Windows XP SP3.

Many thanks,

You can get that multiple recording business if you’re recording from Mix-Out, Stereo Mix, or Wav Out instead of a real device like Mic-In or Line-In. I thought the wiki piece covered that?


I don’t see that it’s even mentioned. I find the terminology very imprecise, ambiguous, and confusing.

In any case, here’s how I do it: with a mixer. What about latency? I don’t get it, what’s the issue?

My computer is connected to my stereo receiver as though it were a CD player, i.e. to a set of “tape monitor” inputs and outputs. The settings:

I. Windows 7

A. Playback device is headphones
B. Recording device is Line In

II. Audacity 2.0

A. Playback Device set to Headphones
B. Recording device set to Line In
C. Overdub (in Edit/Preferences) is On [I question the use of the term “overdub” in this context; see below]

III. Computer and Mixer (Behringer Xenyx 802):

A. Computer headphone jack to Mixer CD/TAPE inputs. On my computer this mutes the line out to the stereo receiver.
B. Mixer Main Out to Stereo Amp Aux inputs jack.
C. Stereo receiver (tape monitor circuits) set to “source”
D. Instrument and/or mic plugged into Mixer inputs.
E. Headphones plugged into Mixer Headphones jack.

The key is the mixer. It mixes the instrument/mic input with computer output (the first track) and routes it to the headphones so you can hear both at the same time, using the first track as something to play to. But the only thing the mixer sends to the computer (via the stereo receiver) is the instrument/mic input, unmixed with the first track, via the line in. Voila, two discrete tracks, the first used as a time cue for the second.

True, the computer is playing and recording at the same time, and so there is latency when both are played back. I fix that manually in Audacity with time shift. Takes ten or fifteen seconds, at most. OTOH I’ve wasted hours just trying to understand what various people mean by their latency fixes, let alone making them work.

My terminology beefs:

  1. Mixing = two or more tracks mixed to produce a new recording.

  2. Overdubbing = one or more tracks playing, while, in real time, another is recorded and mixed with them to produce a new recording.

  3. Multi-tracking = two or more tracks played back in synchronization to sound like one, though they remain discrete sources.

E.g. I record and then synchronize a bunch of computer synthesizer tracks with Audacity; That’s multi-tracking. When I push “play”, the computer mixes them to two stereo tracks and sends them to the headphones and/or speakers. Then I want to play some guitar on top, which I do with the setup above, which produces an additional track. This is still multi-tracking, not overdubbing or mixing. Having recorded the additional, discrete, not over-dubbed, not mixed guitar track . . .

First, I synchronize the guitar track with the computer tracks, then tweak the guitar track with levels, effects, etc. Then I adjust the relative levels of the various tracks. Then I have Audacity export the whole shebang to a recording (actually when you push play a mix is sent to the headphones). That’s mixing. There is no overdubbing. Overdubbing would be if be if I was recording a real-time mixture of computer tracks and live guitar, voice, something. Normally this is precisely what you are trying to avoid, as with the setup above. I suppose it can be used as a creative technique (some sort of echo track) but wonder if it’s worth it; some other method is very likely easier, faster, and more controllable. In any case, IMO the term is used carelessly and causes needless confusion, frustration, and waste of time and money. “Overdubbing” does not mean mixing or multi-tracking. Now that we have digital multi-tracking, over-dubbing is of very limited usefulness.

  1. Monitoring = listening while recording.

A. Listening to a playback is not monitoring.
B. Listening a mix of playbacks, or a playback and a performance mixed together, for purposes of producing a new recording, is monitoring. IOW you are monitoring the mix as it is being recorded, or will be recorded once you get done fiddling.

Monitoring is the most important term, and for some reason the most negligently used. What do you get to monitor, and how? – is the question you must answer before you start recording. It is the most important question in selecting what equipment to buy. For a variety of reasons the mix you want to monitor might be different than the mix you want to record. For one thing, it’s the only way to record a second track using the first track as a guide, metronome, rhythm section, whatever you want to call it and not have the two mixed together, before you want them to be mixed together. The mix that goes to your ears is not the one that is recorded, that’s the entire point.

That’s a particularly narrow definition. The term is more commonly taken in a broader manner:
Some people consider that “Multi-track Recording” means “recording multiple (separate) tracks simultaneously”, which is a narrower definition than your interpretation of the phrase.

Personally I prefer “Multi-channel recording” to mean “recording multiple (separate) tracks simultaneously”, “Multi-track recording” to mean "layering multiple tracks to create a single composition (aka “overdubbing”), and “Overdubbing” to mean “recording a mix of a live input with a pre-recorded (bounced) track”.

I set up the latency correction - it took a couple of minutes, but now the tracks are always in time with each other with an accuracy of a couple of milliseconds, saving me the bother of manually correcting every track. For me it was a couple of minutes well spent.