Strawberry Fields Forever

Story Section status: Outdated

Recording

Equipment

Recording Timeline

• 24 November 1966 — Studio Two, 7.00pm–2.30am: take 1 — The first song taped for the sessions that became the next album, with George Martin producing, Geoff Emerick engineering and Phil McDonald second.¹ One take in seven and a half hours, and it is not a sketch: the recording sheet for the first session shows a fully-realised performance built on Paul’s Mellotron “flute” and Ringo’s drums.² Everything the song went on to acquire — brass, strings, a second band arrangement, a splice — was added to a record that was already complete on the first night.
• Source conflict — the date of the first session — The session sheets date the opening night 24 November 1966; the ‘A Closer Look’ entry for this song heads its account 25 November 1966.^1,2 A booking that ran to 2.30am crosses midnight, which is the likeliest explanation for a one-day discrepancy of this kind, but the mounted record does not resolve it and both dates are given here rather than one being chosen.
• 28 November 1966 — Studio Two, 7.00pm–1.30am: takes 2–4, mono remixes 1–3 — Three more takes, and then mono remixes 1 to 3 made from take 4 in the same booking.¹ Mixing a song on the night it is recorded is what a studio does when it thinks the song is finished. Those three mixes were never used for anything, and the take they came from was superseded the following afternoon; but they mark the point at which the record could have stopped, and the whole of the song’s reputation rests on the fact that it did not.
• 29 November 1966 — Studio Two, 2.30–8.00pm: takes 5–6, reduction into take 7, superimpositions — Takes 5 and 6, then take 6 bounced down into take 7 and overdubs laid on top of it.¹ A reduction is what you do when the four tracks are full and there is still something to add, so take 7 is by definition a second-generation tape carrying more parts than the machine has tracks.² Take 7 is the arrangement the band settled on, and it supplies the first minute of the record everybody knows.
• 8–9 December 1966 — Studio Two: the re-make, takes 9–24 — A complete re-recording, logged on the sheets as a re-make and running to sixteen numbered takes with editing on the same day.¹ Nothing on take 7 was wrong; the song had simply been re-conceived as a heavier, faster piece, and a band that had already spent three sessions on it started again from nothing. That decision is the reason this song has a splice at all.
• 15 December 1966 — Studio Two, 2.30pm–12.00am: take 25, reduction into take 26, mono remixes 5–9 — Take 25 reduced into take 26 with superimpositions onto it, and mono remixes 5 to 9 run from take 26 before the night ended.¹ Four trumpets and three cellos were scored for this version and overdubbed onto it.¹ Take 26 is therefore also a second-generation tape, and the released record splices one bounced-down master to another.
• 22 December 1966 — Studio Two, 7.00–11.30pm: the varispeed convergence and the mono master — Martin and Emerick worked in the control room on the problem that the two versions were a semitone apart, speeding the first up and slowing the second down until they matched.^1,2 The mono master came out of the same booking: remixes 10 and 11 edited together, the edit itself numbered remix 12.¹ The join sits precisely sixty seconds into the released version.¹
• 29 December 1966 — Studio Three control room: the stereo — Stereo mixing from take 26, and then stereo remixes 1 and 2 edited together as a single master.¹ The stereo was built a week after the mono and by the same method, which is why the two share a splice point but not a balance; the mono was the format the record was made for (see Studio Notes).
• February 1967 — issued as a double A-side with Penny Lane — Issued as a single in Britain on Parlophone R 5570 and in the United States on Capitol 5810, some six weeks after the last edit and three months after the first take.⁴ The song had been intended to open the next album; it went out as a single instead and never appeared on a British studio LP (see Releases).

Studio Notes

• Why the two versions would not fit — The obstacle was not tempo alone but pitch: the two recordings were originally a semitone different.¹ A semitone is the largest interval an ear will not forgive; a listener who cannot name the problem will still hear the record change key mid-sentence. Tempo can be edited around by cutting on a downbeat, but pitch cannot be edited around at all. The only lever available in a 1966 control room was the speed of the transport, and speed moves pitch and tempo together, which is what makes the solution here so unusually lucky rather than merely clever.
• Varispeed — the fix, and the reason it worked — Speeding the first version up and slowing the second down brought them together, in Emerick’s phrase, well enough to “weld together”.¹ The reason that manoeuvre closes the gap at all is that the two errors ran in opposite directions: the slower take was also the flatter one, so raising its speed corrected pitch and tempo at once, and the reverse held for take 26. Had both faults pointed the same way no amount of varispeed would have helped. Martin’s own verdict on the arithmetic is in the quote above, and it is not a boast.
• What varispeed cost the record — Nothing on the released master is playing at the speed it was performed at. The album’s varispeed log lists this song’s vocal at 53 cycles for take 7 and 47 for take 26 — the widest span for any single song in the table, and on either side of the 50-cycle standard.² Move a voice that far and the timbre moves with it: the take-7 half of the record is a voice recorded fast and played slow, which thickens it, and the take-26 half is the opposite. The famous half-asleep quality of the vocal is partly performance and partly a tape transport.
• The splice is between two second-generation tapes — Take 7 came from a reduction of take 6, and take 26 from a reduction of take 25.¹ Bouncing is what a four-track machine does when the parts outnumber the tracks — when the tracks are full a reduction mix is made and the freed track is spent immediately.² Both halves of the record therefore carry a generation of tape loss before the edit, and the edit then joins two independently degraded tapes and varispeeds both. That the result sounds as solid as it does is the strongest single argument for how good the Studio Two signal path was in 1966.
• Mellotron Mk II — the instrument that opens the record — The introduction is a Mellotron, played by Paul on the first night.² The Mk II is a tape replay keyboard rather than a synthesiser: each key triggers its own length of magnetic tape carrying a recording of a real player, and the available voices include piano, flute, guitar and organ.² Paul and George encountered one at IBC Studios on 9 August 1965, which is most likely where the group first met the instrument; EMI did not buy machines of its own until February 1968, so the Mellotron heard here was not yet a piece of Abbey Road equipment.² Because every note is a separate tape running at its own slightly imperfect speed, the instrument breathes in a way a flute section does not, and that instability is exactly what makes the intro identifiable in two bars.
• Mellotron on the second version too — It is easy to hear the Mellotron as belonging only to the acoustic conception and the brass and strings as belonging to the re-make, but it was recorded onto the second version as well.² That matters for what the splice actually does. The edit does not move the listener from a keyboard record to an orchestral one; it moves between two versions that share a palette, which is part of why a join between two separate recordings, in two keys, at two speeds, goes unnoticed by most listeners for a lifetime.
• Backwards tape — what is actually reversed — The reversed element on this record is drums and percussion, and it enters at 1:28.² That is inside the take-26 half of the record, seconds after the splice, and it is a rhythmic rather than a melodic reversal — a percussive attack turned into a swell. The backwards performance was not recorded to the four-track machine, as had long been assumed; the recording sheet shows otherwise.²
• Four trumpets and three cellos, and why they are not an orchestra — The score is small and specific.¹ Seven players is a chamber group, not a section, and it is arranged so that the cellos sit under the band rather than beside it — the surviving studio talk about the scoring turns on getting the bottom string of the cello and moving a part up a notch to reach it.² A larger group would have made the second version sound like an arrangement; this one makes it sound like the same band with the room enlarged.
• How the drums were captured — The Studio Two setup reconstructed for these late-1966 sessions runs a D19c as overhead and a KM56 below the snare, with their signals sent to an RS144 premix box before the desk, and a D20 on the bass drum on its own fader.² Premixing the kit down before it reaches the console is the compromise a four-track machine forces: the drum balance is committed at the moment of recording and cannot be revisited later. On a song that was going to be bounced twice and spliced once, that early commitment is why the drums survive the process at all.
• Rehearsing onto tape — The group had begun keeping the four-track running through rehearsal rather than starting the machine only for a take.² It is a small change of habit with a large consequence, and it explains how a first night produces exactly one numbered take that is nevertheless a complete performance: the take number counts what was logged, not what was played, and by the time anyone said “take one” the arrangement had already been found.
• The record was made for mono — The mono master was assembled on 22 December and the stereo not until 29 December, by which time, in Emerick’s recollection of the era’s working practice, the mono record was the one everybody had attended and stereo was made after everyone had gone home.² On a song whose entire identity is a balance between two incompatible recordings, that ordering is not a footnote: the mono is the mix the decisions were made in, and the stereo is a reconstruction of those decisions in a wider frame.
• Finishing the edit with minutes to spare — The engineer’s own account of the splice adds a detail no session sheet records — “I had literally finished the edit just a few moments before he arrived”.³ The most closely studied edit in the catalogue was cut against the clock, by hand, with a razor blade, and heard for the first time by the man who asked for it a few minutes after it existed.

Releases

• 1967 UK mono single — Strawberry Fields Forever / Penny Lane (17 February 1967, Parlophone R 5570) — Dense and centred, the version the record was made for, with the voice well forward, the brass and cellos packed in behind it rather than spread, and the join at one minute almost impossible to catch. This is the reference for everything below.
• 1967 US single — (13 February 1967, Capitol 5810) — The same coupling, out four days before Britain. Mono and stereo pressings circulated; the mono is the December mix, and the American cut is louder and slightly brighter at the top than the British one.
• 1967 US LP — Magical Mystery Tour (Capitol MAL-2835 mono / SMAL-2835 stereo) — The song’s first album home, and an American one; there was no British studio LP to put it on. The stereo here is the 29 December mix: voice to one side, band to the other, and the two backwards-percussion entries and the Mellotron much easier to separate than in mono — a wider picture, and a less concentrated one.
• 1971 stereo remix — German Magical Mystery Tour (Hor Zu SHZE 327, later Apple 1C 062-04 449) — A genuinely different stereo, wider and drier than the 1967 one, with a balance closer to modern practice and a coda handled differently. Remixed at AIR on 26 October 1971 and issued in Germany rather than Britain or America; later pressings of the German album carry it.
• 1987 CD — Magical Mystery Tour (EMI CDP 7 48062 2) — The first digital issue, carrying the 1967 stereo.
• 1996 CD — Anthology 2 (Apple CDP 8 34448 2) — Two pre-splice recordings rather than a remix. Take 1 appears in a mono transfer of the November 1966 tape — the song as it stood on the first night, Mellotron and drums and nothing else — alongside a stereo outtake mixed in 1995 whose verse-three backing vocals, long known from bootlegs, are pulled down.
• 2009 stereo remaster — Magical Mystery Tour (Apple/Capitol 99460-2) — The 1967 stereo remastered rather than remixed, so the wide 1967 placement stands and the splice sits where it always sat; what changes is that the cellos and the backwards percussion are audible against the vocal instead of behind it. The mono of the same programme carries the 22 December mix.