Pitch a Film Restoration as a Specific Version of a Work

A restoration pitch usually gets asked to be one thing — an announcement — when it needs to be three. What survives? Which version is the project trying to make available? And when the work is finished, what will an audience actually be able to see?

Those are separate answers, and the word "restored" hides all three. A reader can finish a pitch that promises "the complete original, restored at last" without learning whether the archive holds a print or a negative, which cut is being assembled, or whether the result will ever leave the vault. It sounds finished and says nothing anyone can check.

The distinction has institutional backing, though you should take it as one institution's framing rather than a rulebook. In a 2014 post on its Unwritten Record blog, the U.S. National Archives and Records Administration separated preservation from restoration and described preserving The March photochemically before producing a digitally restored exhibition version ("Film Preservation 101: Is Restoration the Same as Preservation?", 19 March 2014). That is a historical explanation of one archive's practice. It tells you nothing about your own material's condition and it is not a workflow to copy. What it does is name two different results — a source protected, a presentation prepared — that a single word tends to merge.

A usable pitch keeps them apart from the first paragraph, and names the version it intends to make. Here is what that looks like in practice.

Start from what someone has actually checked

The material section of a pitch is an inventory, not a diagnosis.

An inventory says what exists, how it is identified, and who established that. "The archive holds one 35mm composite print; its head titles identify it as the 1948 general-release version; the count and completeness of its reels were checked by the archive in March." That is a sentence a reader can act on.

A diagnosis is different, and usually not yours to make. "The print is badly warped and should be scanned at 4K" is a claim about condition and treatment. If the pitch writer is not the archivist, that claim belongs to the archivist — quoted, attributed, and dated. You do not establish it by looking at a thumbnail, a JPEG, or somebody else's old transfer. Those tell you what a previous copy looked like, not what the element in the vault is doing.

Keep confirmed items and hoped-for discoveries visibly separate. "The archive holds a print" and "a negative may survive in a private collection" are both legitimate sentences in a pitch, but the reader has to be able to tell which is which. Use may, reported, not yet confirmed where they apply. A proposal built on an unchecked hope is not a bad proposal; a proposal that disguises an unchecked hope as an inventory is.

Name the version the project is serving

Most films exist in more than one state. There is a release cut and a reissue cut. There is an export version with a different ending, a silent version with a different score, a television version with footage removed or added. Choosing one of those and naming it is not a limitation on ambition. It is the thing that makes the project legible.

Promise "the original" and you take on a claim you probably cannot support. Name the 1948 general-release version and you have given the archivist a testable proposition — and given yourself a defensible one. The evidence that identifies a version is usually concrete: head titles, running order, running time, a distributor card, documentation in the archive's files, a cutting continuity, a censor record. Say which of these supports your target.

When the target is disputed or incompletely established, show that instead of smoothing it over. If two archive staff disagree about whether a surviving print is the first release or an early revision, a pitch that says so is more useful than one that picks a side and calls it recovery. It also protects the project later: a decision that was visible from the start is a decision, while one that surfaces after the money is spent is a dispute.

This is the part of the pitch that most often goes wrong in the other direction, too — the pitch that offers a preferred reconstruction as though it were recovered fact. A version you are assembling is a version you are proposing. Say which it is.

Preservation, assembly, and access are three different promises

A restoration proposal commonly bundles three kinds of work, and readers assume they are one deliverable unless you tell them otherwise.

Preservation is work on the source material: stabilizing it, copying it, protecting it against further loss. The pitch states that this work is planned and who owns it — usually the archive, following its own plan. The pitch generally does not need to specify the method, and should not invent one.

Assembly is work on the presentation: what the audience will encounter when the version is put together. This is where material is inserted, omitted, reordered, or replaced — and it is where the project's editorial choices live.

Access is the copy people will actually watch. Which one, in what form, available how.

These can all belong to the same proposal. They are not interchangeable results. A new access copy does not describe every action taken on the source, and protecting a source does not by itself produce anything an audience can see. The NARA account of The March is useful precisely because it holds two of these apart: a source preserved, and a separate restored exhibition derivative. Treat that as an example of how two results can coexist in one project, not as a required sequence for yours. Definitions of preservation and restoration vary between institutions, and your pitch should state its own deliverables in its own words rather than borrow a vocabulary it cannot define.

An invented case: two copies of The Salt Road

Everything that follows is made up for the exercise. The film, the archive, and the people are fictional, and the numbers are stipulated so the reasoning can be followed.

The Kestrel Archive proposes a project around The Salt Road, a 1948 feature. Two elements matter.

Copy A is a 35mm composite print held by the archive. Its head titles identify it as the general-release version. It is complete except for about five and a half minutes missing at a splice in reel four. As it survives, it runs roughly 78½ minutes. Contemporary press material held by the archive gives the release running time as 84 minutes.

Copy B is a 16mm reduction print of a 1956 reissue, held by a private collector. It carries reissue credits and a slightly different running order. It contains a passage covering the same story beat as A's gap — but edited differently, and about four minutes long in the reissue's arrangement.

Two ways to handle the same gap

Both options are honest. They produce different works.

Retaining the gap gives an audience the general-release version as it actually survives, with the hole in it marked rather than filled. The result is a shorter film than the 1948 release, and it says so. This is the more conservative choice and, when documentation is thin, usually the more defensible one.

Inserting Copy B's passage gives an audience a more continuous film. It is not dishonest — provided the pitch describes it accurately. What can be said is that a 1956 reissue print contains a passage covering the same narrative beat, edited differently, at a smaller gauge. What cannot be said is that this is the sequence the 1948 version contained. The shot order differs. The length differs. The gauge differs. The reissue is evidence that the scene existed and roughly what it showed; it is not a record of how the general-release cut arranged it.

The arithmetic matters here and it is easy to get wrong. If A runs 78½ minutes and B's version of the missing passage runs about four minutes, inserting it produces a version of roughly 82½ minutes — not 84. The assembled film will still be shorter than the 1948 release, for a different reason than before. A pitch that states this in advance is a pitch that will not be embarrassed by it later. The reconstruction closes a gap; it does not restore a running time.

The credits row is the same kind of judgment in miniature. Copy A's titles are the evidence that identifies the target version. Copy B's titles identify a different release. Putting B's reissue credits onto A's cut would attach a 1956 release identity to a 1948 version — a small decision with a large consequence for what the resulting film claims to be.

One question the table raises and the pitch should not answer: whether Copy B can be used at all. That involves an agreement with its holder and whatever clearances apply. Flag it as an open question for the archive and the rights holders. Rights clearance is a separate piece of work with its own expertise, and a restoration pitch is not the place to resolve it.

Turn the table into a paragraph

The table's value is that it converts into an ordinary project account an archivist can read quickly.

The Kestrel Archive proposes to make available the 1948 general-release version of The Salt Road, using its 35mm composite print (Copy A) as the primary source. Copy A is complete except for approximately five and a half minutes missing at a splice in reel four; as it survives, it runs about 78½ minutes against a documented release running time of 84. The project will either retain that gap, marked and documented, or insert the corresponding passage from a 1956 16mm reissue print (Copy B) as a disclosed reconstruction. In the second case the inserted material will be identified as a different cut at a different gauge, not as a record of the 1948 arrangement, and the resulting version will run about 82½ minutes. Head titles will be taken from Copy A; Copy B's reissue credits will not be used. The archive's preservation plan for Copy A proceeds separately. The access deliverable is a viewing copy of the assembled version with a change record listing every intervention and its source. Open questions: whether any other element holds reel four; what documentation describes the missing passage; whether Copy B is available for use.

That paragraph is checkable. An archivist can confirm or correct the element descriptions, agree or disagree with the target, accept or reject each editorial decision, and answer the open questions — all without having to guess what the project meant. A promise to "restore the complete original" gives that reader nothing to check, and quietly asks them to approve work that the material may not support.

Say what the audience can encounter, and leave the rest to the specialists

The access section should describe a result, not a demand. In the invented case, the surviving material for the 1948 version is a 35mm composite print held by the archive, while the later cut survives only as a collector's 16mm reduction print. Neither is a form most people can see. The project's outcome is a viewing copy of the version named above, with its absences and its interventions disclosed. That is a specific thing to fund, and it does not require a claim about how many people want it.

What the pitch should not specify is everything downstream of purpose. Condition assessment, treatment methods, scanning or photochemical choices, schedules, facility capacity, and cost belong to the specialists who will do and price the work — and in most projects, to the archive that owns the source. The pitch's job is to say what the project is for and what shape its results will take, then hand the feasibility questions to the people who can answer them. If the proposal needs a cost figure, it should come from those people and be attributed to them, not assembled from a rule of thumb.

For that reason, keep the uncertainty where it can be seen. A reconstruction whose limits sit in a footnote reads as a finished work with a minor caveat. The same reconstruction described in the main account — this passage comes from a different cut, this arrangement is not established, this section is absent and stays absent — reads as what it is: a considered proposal with known boundaries. The second version survives contact with the archivist, the reviewer, and the audience. The first one doesn't.

Frequently asked questions

Why is saying 'restored' not enough in a restoration pitch?

The word hides three separate answers: what survives, which version the project is trying to make available, and what an audience will actually be able to see. A pitch can promise 'the complete original, restored at last' without saying whether the archive holds a print or a negative, which cut is being assembled, or whether the result will leave the vault. Those are different results, and readers need them kept apart.

What is the difference between an inventory and a diagnosis in the material section?

An inventory states what exists, how it is identified and who established that. A diagnosis makes a claim about condition and treatment, such as saying a print is warped and should be scanned at 4K. That claim belongs to the archivist unless the pitch writer is the archivist, and it should be quoted, attributed and dated. A thumbnail, JPEG or old transfer tells you what a previous copy looked like, not what the vault element is doing.

How are preservation, assembly and access different promises?

Preservation is work on the source material: stabilizing, copying or protecting it against further loss. Assembly is work on the presentation: what the audience encounters when the version is put together, including inserted, omitted, reordered or replaced material. Access is the copy people will actually watch. They can belong to the same proposal, but they are not interchangeable, and definitions vary between institutions.

In the Salt Road example, what changes if Copy B's passage is inserted?

Retaining the gap presents Copy A as it survives, with the absence marked and documented, and the result is shorter than the 1948 release. Inserting Copy B's passage gives a more continuous film, but B is a 1956 16mm reissue with a different edit, length and gauge, so the inserted material must be labeled as a disclosed reconstruction, not as a record of the 1948 arrangement. The assembled version would run about 82½ minutes, not 84. Head titles should come from Copy A; Copy B's reissue credits should not be used.

How should the access section and uncertainty be handled?

The access section should describe a result, not a demand: a viewing copy of the named version, with its absences and interventions disclosed. Condition assessment, treatment methods, scanning or photochemical choices, schedules, facility capacity and cost belong to the specialists who will do and price the work. Keep uncertainty in the main account rather than in a footnote, so the reconstruction reads as a considered proposal with known boundaries.

Further reading

• A Documentary Uses a Composite Character. What Can the Pitch Still Claim? — Read alongside any reconstruction: it shows why combining sources into one continuous account creates a new claim and what a pitch may still honestly say about it.
• Build an Archive-Led Film Around the Image You Cannot Find — When the pitch depends on material nobody has found, this weighs retaining the absence against designing around it and rewriting claims to match the record.