Two hours in the dark, no light to write by, phones off and in many press screenings collected at the door. The film keeps moving whether or not anyone is ready.
Somewhere in the second act there is a specific observation worth making: a cut that undercuts a set piece, a line of dialogue that contradicts the character’s arc, the moment the geography of a fight becomes unreadable. It is precise while the film is running.
Four hours later it has become the film’s action is uneven in the middle. That sentence is true, useless, and interchangeable with the same sentence about a hundred other films.
What is the best way to take notes on a film?
Record a spoken account immediately after leaving the auditorium and before reading or hearing anyone else’s opinion, then convert it to text. Writing during the film is impractical and, in most press and public screenings, prohibited. Writing afterwards at a keyboard is possible but happens too late, because the useful material is gone within a couple of hours and, more importantly, has usually been overwritten by other people’s reactions by the time anyone sits down. Speech is the only capture method available in the ten minutes that matter, which are the ten minutes between the credits and the first conversation about it.
The impression changes overnight, and not randomly
The reason to record immediately is not simply that memory fades. It is that memory is revised in a specific direction by other people.
The effect has been measured directly. In a study published in Applied Cognitive Psychology, Gabbert, Memon and Allan showed pairs of participants different versions of the same event, each containing details the other had not seen, and then allowed some pairs to discuss it before recall. Seventy-one per cent of those who discussed the event went on to report details they had never actually witnessed, having acquired them from the other person.
The research concerns eyewitnesses rather than film reviewers, and the mechanism transfers cleanly. Anyone who watches a film, then reads three reviews, then writes their own has run exactly that experiment on themselves. Details from the reviews get incorporated. The judgement converges on the emerging consensus. And crucially, the writer experiences the converged version as their own opinion, because there is no felt difference between a memory and an absorbed one.
This is the reason a great deal of film writing published in the days after a release says roughly the same thing. Not because everyone independently reached the same conclusion, but because most of them wrote after reading the others.
A recording made in the car park is the only version of the response that predates all of that.
What gets lost is the specific, not the general
The general judgement survives. Whether the film worked, roughly how it compared to the previous one, whether the lead was well cast: none of that needs recording, because it is still available a week later.
Four things go quickly.
Sequence. Which scene came before which, and therefore whether a reveal landed or was undercut by something earlier. Structural criticism depends entirely on order, and order is what memory reorganises first.
The shot. What the camera actually did during a sequence. Whether the fight was shot in wide takes or cut every eight frames, whether the blocking was legible, whether the coverage cheated the geography. These are the observations that separate a review of an action film from a summary of one, and they are gone by morning.
The exact line. Dialogue paraphrased is dialogue defanged. A line worth quoting has to be captured close to verbatim or not used.
Names. The stunt coordinator, the second unit director, the composer, the actor in the two scene part who was better than anyone else in the film. These are read once in the credits and lost immediately, and they are frequently the most interesting thing to write about.
The car park recording
The practical routine is short and slightly antisocial, which is why it works.
Leave the auditorium and do not talk to anyone yet. This is the hard part and it is the whole method. One exchange of first impressions with a companion is enough to start the conformity process.
Record two to four minutes of speech, standing outside. Not a review; a debrief.
State the film, the date, the format and the screening type first. Format matters more than it sounds for action cinema, since a judgement about legibility depends on whether it was seen in a large format presentation or on a small screen with the brightness turned down.
Then work through five things.
The single strongest sequence and what specifically made it work: the staging, the cutting rhythm, the sound, whichever it was.
The single weakest, on the same basis.
Any line worth quoting, said out loud as close to verbatim as memory allows, with an explicit note that it is approximate.
Names from the credits, before they are forgotten.
The one sentence that would open the review if it had to be written now. This is the sentence most likely to be replaced by consensus later, so it is the one most worth preserving.
Only then have the conversation about it. The recording is not a substitute for talking about films, which is most of the point of watching them. It is insurance against the conversation retroactively becoming the opinion.
From recording to a working document
Phones record to .m4a by default, which is AAC audio in an MPEG-4 container, and most transcription services accept the file without conversion.
Turning M4A voice memos to text produces a punctuated, editable document from the recording, which is what makes it usable as the basis for a draft rather than something to be listened to again. Vomo also produces a short summary with the key points for longer recordings and exports to plain text, Word, PDF or Markdown, so the debrief can sit directly above the draft in whatever the review gets written in.
One check worth making: proper nouns. Names of crew members, characters and production companies are the least reliable part of any automatic transcript, and a misspelled stunt coordinator in a published review is the kind of error that gets noticed by exactly the readers who care most.
The archive is what pays off
A single debrief improves a single review, modestly.
The value compounds across a body of work. Anyone reviewing action cinema regularly ends up writing about franchises, about the same directors over a decade, about the drift in how fights are staged as the technology and the insurance requirements change. Those pieces need specifics from films seen years ago, and specifics from years ago exist only if something recorded them.
A searchable file of first impressions makes claims possible that are otherwise unavailable. That the second film in a series had longer takes than the third is either a documented observation or a vague feeling, and the difference is visible immediately to a reader. The same file also produces the occasionally uncomfortable discovery that a film now regarded as a low point was described enthusiastically on the night, which is worth knowing about one’s own judgement.
The Q&A afterwards is the part worth recording properly
Festival screenings, premieres and repertory events frequently end with twenty or thirty minutes of the director, the star or the stunt coordinator answering questions from a moderator and the room. This material is different in kind from the debrief, and it deserves a different treatment.
It is quotable, and quotation demands accuracy. A remark about how a sequence was shot, or why a scene was cut, is the sort of thing that gets repeated across a dozen outlets, and the version that circulates is whichever one got transcribed first. Getting it wrong is a correction; getting it right first is a piece nobody else has.
Three things make a Q&A recording usable rather than merely present. Speakers need separating, since a panel with a moderator and two guests produces answers that are useless if attributed to the wrong person. Time references need to be close together, so that any line going into a piece can be checked against the audio in seconds. And the exchange needs to survive intact, because a Q&A answer is frequently unintelligible without the question that prompted it, and the question is the part that gets dropped.
Two practical notes. Check whether recording is permitted before doing it, since festivals and studios set their own rules and press access is easy to lose. And treat the transcript as a source rather than a script: proper nouns, crew names and film titles are the least reliable part of any automatic transcription and are precisely what a Q&A is full of.
Home screeners change the calculus, and not entirely for the better
An increasing share of review viewing happens on a watermarked link at home rather than in an auditorium, and the conditions are different enough to be worth treating separately.
The obvious gain is that the film can be paused. Notes can be typed. A sequence can be run twice to check whether the geography really was incoherent or whether it only felt that way. For structural criticism this is a genuine advantage and there is no reason to refuse it.
The less obvious loss is that pausing changes the object being reviewed. A film watched in eleven segments across two evenings, with the phone in reach, is not the film an audience will see, and judgements about pacing, tension and whether a set piece sustains itself are exactly the judgements that fall apart under those conditions. A reviewer who paused four times through the third act is not in a position to say the third act drags.
The workable compromise is to watch the first pass straight through, without stopping and without notes, record the debrief immediately as described above, and only then rewatch selected sequences with the pause button. That preserves the reaction of a viewer and adds the precision of an analyst, in that order.
Two things not to record
Worth stating plainly, because the subject invites confusion.
Do not record anything during the film. Capturing audio or video from a screen is piracy, it is a condition of entry that press and public screenings both enforce, and being caught doing it ends access to screenings permanently. The recording described here happens outside, afterwards, and contains one voice: the reviewer’s own.
Do not publish before the embargo. Press screenings frequently come with a date and sometimes a time before which nothing may appear, and a recording made on the night is a private working document until that point. Writing early and publishing early are different decisions, and only the second one breaks anything.
The short version
The observations that make a review worth reading are made during the film and are gone within hours, partly through ordinary forgetting and substantially through absorbing other people’s reactions.
Nothing can be written in a dark auditorium. Something can be said in a car park, and converting it to text costs a couple of minutes.
The alternative is writing tomorrow, from a memory that has already been edited by everyone else who saw it.