The short version: Double-tracking is the biggest free upgrade available to a home recording, and it is also the most expensive thing you can do with a short session. A second take costs roughly two and a half times the length of the part. Spend it on rhythm parts, riffs, and anything repeatable, where the two takes fuse into one wide guitar. Do not spend it on leads, bends, or heavily modulated parts, where they will not fuse and you have burned the session for a flam. When the time runs out, a doubler block is a real answer with one real limitation.
The advice everywhere is "double-track your rhythm guitars." It is good advice. It is also written by people who are not looking at the clock.
If your session is a fixed block that ends when it ends, a second take is not free. It is the most expensive decision in the whole recording, because it is the only one that costs you time proportional to the length of the part. So the question is not whether double-tracking is good. It is which parts are worth it when you only get to double one thing.
Here is how I decide.
The Decision Table
| Part | Double it? | Why |
|---|---|---|
| Strummed chord rhythm | Yes | The most reliable fusing part there is. Biggest payoff per minute. |
| Palm-muted eighth-note riff | Yes | Repeatable, transient-heavy, locks tight. |
| Single-note riff, moderate tempo | Yes | Works if you can play it twice without thinking. |
| Arpeggios | Maybe | Fuses well, but the note ends are exposed. Costs more takes. |
| Sustained chords, whole notes | Maybe | Sounds huge when it works. Pitch drift shows badly. |
| Lead line with bends and vibrato | No | Vibrato will not match. You get a wobble, not width. |
| Anything with chorus, flanger, or phaser | No | The modulation is at a different point in its cycle each take. |
| Anything drenched in delay or reverb | No, not wet | Double it dry, add time effects after. |
| Fast, technically hard passages | No | The second take costs four takes. Use a doubler. |
| A part you have not learned yet | No | You are not doubling. You are practicing on tape. |
The pattern is that fusion needs agreement, and agreement is easiest where the part is rhythmically simple and the notes are short. Two takes of a Courtney Barnett-style strummed part will lock without you trying. Two takes of a slow bend will never agree on where the pitch arrives, and your ear hears the disagreement as a mistake rather than as width.
What Double-Tracking Is Not
This trips up more home recordings than anything else in this article, so it goes before the technique.
Copying the track, pasting it to a second lane, nudging it 20 milliseconds, and panning the two apart is not double-tracking. It is a short delay wearing a costume.
The reason two real takes sound wide is that they are genuinely independent. The micro-timing is different, the pitch wanders a couple of cents in different directions, the pick hits at slightly different angles. There is no fixed relationship between them, so there is no fixed interference pattern.
A copy has a fixed relationship, which means the moment those two channels sum to mono you get comb filtering — a stack of evenly spaced notches straight through the guitar. A phone speaker does that. A club PA does that. A bar TV does that. The part sounds enormous in your headphones and hollow everywhere else, and you will not find out until it is too late to re-record. There is a full breakdown of which stereo tricks survive a mono fold-down if you want the mechanism, but the short version is that true double-tracking is on the survivors list and copy-and-nudge is not.
Same problem, slightly disguised: reamping one DI through two different amp models. The tone differs, so it feels like two guitars. But it is one performance, so the transients are perfectly aligned and you have the same correlation problem with extra steps.
The Cost Side
Nobody publishes the math on this, so here it is roughly.
A second take costs about 2.5x the length of the part. One clean pass, plus a punch-in or two, plus the twenty seconds of getting your hands back on it.
- 40-second rhythm part: about 4 minutes. Do it.
- 2-minute full-song rhythm bed: about 8 minutes. Do it if rhythm is the song.
- 90-second lead: about 15 minutes, and it will not fuse. Do not do it.
In a 20-minute session that math is the whole strategy. You get one, maybe two, doubles. Spend them where the table says the payoff is highest, which is nearly always the rhythm bed.
One thing that makes the cost lower than it looks: record the second take immediately after the first one. Do not listen back in between. Do not go make a coffee. Your hands still have the part, and take two right after take one is dramatically closer than take two after a three-minute listen. If you break the chain you pay for the second take twice.
How Tight Is Tight Enough
Looser than you think.
Attacks landing within about 20 to 30 milliseconds of each other read as one guitar. Past 40 or so you start hearing a flam, and it sounds like a mistake. Under a few milliseconds you have effectively made one take at higher level and given up the width you paid for.
So there is a window, and it is wide. Most players who think they cannot double-track are actually failing at something else.
The Part I Had Wrong
I spent a long time assuming a smeared double was a timing problem at the attacks. Play the downbeats more accurately, get tighter, fix the smear. That is where all the attention goes, because attacks are what you hear yourself playing.
It is not the attacks. It is the ends.
Your ear locks two sounds together at the transient — that is what fusion is, and the window for it is generous. What it will not ignore is two notes stopping at different moments. One take rings for another 200 milliseconds after the other one has been muted, and that tail is suddenly a solo guitar sticking out of a stereo image, completely exposed, drawing the ear straight to it.
The test takes thirty seconds. Take a double that sounds loose, find a chord where it smears, and trim or mute both takes so they stop at the same point. It tightens up immediately and you did not touch a single attack.
Which changes how you play the second take. Stop thinking about hitting the downbeats together and start thinking about releasing together. Muting is the part to match. For strummed parts that means your fretting hand relaxing at the same point in the bar. For palm-muted riffs it barely matters, because the palm is already ending every note for you — which is a decent explanation for why chugging riffs are the easiest thing in the world to double and sustained open chords are the hardest.
The Setup That Makes It Faster
Five things, none of which take long.
- Play to the click or the drum track, not to take one. Have take one audible but quiet in the phones. If it is loud you will chase it, drift toward it, and end up too tight.
- Mark your picking hand position. Playing take two closer to the neck changes the timbre enough that the two takes sound like two different guitars in two different rooms instead of one wide one. Same spot, same angle.
- Do not change anything on the amp. Not the gain, not the tone, not the pickup. Same everything.
- Record dry if the part has time effects. Print the doubles clean, pan them, then send both to one reverb. Two reverbs is two rooms.
- Pan hard and match levels. Hard left and hard right is the standard for a reason. Then check that neither side is louder, because the louder one pulls the whole image over and you will misread it as a performance problem.
If you have a two-input interface and a second guitar you like, tracking with a different guitar on take two is a genuine upgrade rather than a compromise. Different pickups, slightly different scale length, more difference between the two voices. Not necessary. Just nice.
When the Doubler Is the Right Call
Sometimes the answer is that you cannot play it twice, or you have four minutes left. A doubler is a legitimate tool and I would rather use one than ship a thin track.
On an HX Stomp or a Helix, the block is called Double Take, and it goes after the amp and before your stereo blocks. It is built to imitate performance variation rather than to be a plain short delay, which puts it well ahead of a copy-and-nudge. Most DAWs have an equivalent; in GarageBand the fastest version is a duplicate track with a small pitch and timing offset, which is cruder but takes about forty seconds.
The honest limitation is the same one from earlier. Anything that manufactures a second voice from one performance is correlated with that performance, so how it behaves in mono depends entirely on how it was built. The pitch-shift part of a doubler survives mono fine, because a pitch offset never settles into a fixed relationship. The delay part does not. If you know the track is going to a mono PA, lean on the pitch-based settings and keep the delay times short and small.
And the mixing choice that matters more than which doubler you picked: a doubled part does not need to be as loud as you think. It is doing width, not volume. Bring both sides down until the vocal sits back on top, then leave it.
The Twenty-Minute Version
If you take one thing from this, take the running order.
Record the rhythm bed. Record it again immediately, before listening back. Pan hard left and right. Match the levels. Add everything else as single takes, mono, centered.
That is it. Four minutes of extra work on the one part that repays it, and the arrangement has a floor to stand on. A single-tracked lead in the middle of two solid rhythm guitars sounds intentional. A single-tracked everything sounds like a demo.
You are not going to double the whole song. Nobody with a real schedule doubles the whole song. Pick the part that holds the arrangement up and give that one the second take.
