A reel collection rarely arrives as one clean, consistent source. A digitization provider may deliver dozens of ProRes files with different exposure levels, splice artifacts, gate weave, and scan borders. A useful batch restoration workflow example must therefore do more than apply the same filters to every file. It needs to separate what can be standardized from what still requires judgment.
For 8 mm, Super 8, 9.5 mm, and 16 mm scans, the goal is repeatable restoration without flattening the character of the original film. That means building a controlled preset, validating it on representative shots, organizing exceptions, and exporting files that match the delivery requirement.
Start with a representative sample, not the whole batch
The fastest way to waste processing time is to queue 40 reels before checking how the restoration settings behave on difficult footage. Select three to five clips that represent the collection: a normally exposed outdoor scene, a low-light interior, a shot with visible dust, a scene with camera movement, and a reel containing splices or severe instability.
Inspect each sample at 100% view. Look for grain structure, scratches, dirt, flicker, color cast, frame jitter, and scan-edge movement. Also identify characteristics that must be preserved. Fine grain in well-exposed reversal film is not automatically a defect. Likewise, handheld camera movement should not be treated as gate weave.
This initial review establishes the limits of automation. If the material comes from the same stock, scanner, and capture session, a single preset may cover most reels. If the archive combines Kodachrome, faded Ektachrome, black-and-white reversal, and several scanning systems, divide it into separate batches. Consistency inside each batch matters more than the total number of files.
A batch restoration workflow example: 24 Super 8 reels
Consider a practical job: 24 Super 8 reels scanned to 10-bit ProRes 422 HQ. The collection contains family footage from the 1970s and 1980s, captured with two cameras. Most reels have moderate dust, slight magenta fading, and periodic gate weave. Several include cement splices that cause a brief exposure jump and frame displacement.
The operator first creates three groups: stable daylight reels, faded daylight reels, and mixed-light reels. This is not unnecessary administration. A color correction that restores a faded outdoor reel can make an indoor tungsten scene look unnatural.
For each group, the workflow uses the same processing order:
- Crop or mask unwanted scan borders before motion-based analysis.
- Apply stabilization only where the frame movement is mechanical.
- Remove transient dirt and scratches while protecting real detail.
- Reduce grain conservatively after dirt cleanup.
- Correct color and tonal balance.
- Treat splice interruptions separately when required.
- Encode a preservation master and a viewing derivative.
The order is deliberate. Stabilization tools such as Perfo Lock or motion analysis based on MVTools2 need a clean definition of the usable frame. Dirt removal before grain reduction helps prevent dust from being softened into persistent image texture. Color correction near the end makes it easier to judge whether cleanup has removed detail or altered the image character.
Build the preset around the film’s defects
A batch preset should contain only the operations that are justified across the group. In this Super 8 example, moderate temporal dirt removal can be applied to all daylight reels because dust appears throughout the scans. A tool such as RemoveDirtMC can suppress transient spots effectively, but aggressive settings may erase fine textures, especially in foliage, fabric, and faces moving through the frame.
Use preview frames with both still detail and movement. A setting that looks excellent on a static portrait may leave motion trails on a child running across a lawn. If dirt is heavy but only present on a few reels, create an exception preset instead of increasing cleanup strength for the entire collection.
Grain reduction deserves the same restraint. Grain is part of the photographic record, particularly on small-gauge film. A modest temporal pass can make compression and delivery more efficient while retaining texture. Stronger reduction is appropriate when scanner noise or severe grain overwhelms the image, but it should be checked against faces, hair, and textured backgrounds. The cleanest image is not necessarily the most faithful image.
For color, begin with neutral references where available: white clothing, gray pavement, clouds, or known film leader. GamMac-style gamma and color adjustments can restore faded contrast without clipping highlights. Avoid forcing every reel toward the same modern saturation level. A 1970s indoor scene may legitimately remain warm and low contrast after correction.
Stabilize gate weave without stabilizing the camera
Mechanical image movement is common in small-gauge scans, but stabilization is one of the easiest operations to overapply. Perforation-based stabilization is especially valuable when the scanner captures a stable perforation reference. It can correct frame drift while preserving deliberate pans, tilts, and handheld movement.
When perforation information is unavailable or unreliable, motion-based stabilization can help, but test it carefully. A wide pan across a detailed landscape may be interpreted differently from a close-up with limited background detail. Review the start and end of each clip, as well as transitions around splices. A stable frame that visibly warps or jumps at a splice is not an improvement.
For reels with repeated splice damage, use a dedicated splice cleanup step rather than asking the global stabilizer to solve it. Splice Cleanup is designed for brief disruptions in alignment and exposure. It is more targeted, and it reduces the chance that normal footage will be affected by settings intended for a handful of damaged frames.
Queue the work with clear naming and exception handling
Once the presets are approved, batch processing becomes an operational task rather than a guessing exercise. Keep source files read-only and use a folder structure that distinguishes originals, restoration masters, review exports, and delivery files. The source filename should remain traceable through every output.
A practical naming convention might retain the reel identifier and add a process suffix, such as `R12_Daylight_Restored_Master`. Avoid names like `final_final_2`. They create uncertainty when a client requests a revision six months later.
Before launching the full queue, process one complete reel from each group. Check the rendered file, not only the preview. Confirm frame rate, duration, audio sync, pixel aspect ratio, color range, and edge cropping. Sound film requires special attention because a video restoration setting should never introduce a timing change that drifts away from the optical or magnetic soundtrack.
Exceptions should be recorded as they appear. If Reel 17 needs stronger scratch treatment, duplicate the base preset, modify only the required parameter, and label the exception. This protects the consistency of the main batch while making the deviation explainable and reproducible.
Choose export formats by purpose
Restoration and delivery are different requirements. A preservation-oriented master should minimize further loss and remain suitable for future re-encoding. FFV1 is a strong archival choice where the recipient supports it, while ProRes can be practical for postproduction interchange. For high-quality distribution, x264 or x265 can provide efficient delivery files when bitrate, chroma subsampling, and playback compatibility are selected deliberately.
For example, a small archive may retain a 10-bit 4:2:2 ProRes restoration master, create an H.264 viewing copy for family review, and generate a higher-efficiency H.265 version for long-term access storage. A postproduction client may instead require ProRes 422 HQ or 4:4:4 output to preserve a grading pipeline. There is no universal best codec. The correct choice depends on whether the file is a master, an edit intermediate, an access copy, or a web delivery.
Do not discard the restoration preset after export. The settings are part of the project record, alongside notes about source condition, crop decisions, frame rate, and exceptions. With AviSynth+ processing presented through a visual workflow, that record provides technical repeatability without forcing the operator to maintain manual scripts.
Review for artifacts, not just improvements
Quality control should focus on what the workflow may have introduced: temporal smearing, ghosting, clipped channels, unstable borders, pulsing exposure, audio drift, or encoder artifacts. Watch at normal speed first, then inspect suspicious frames. Restoration defects often reveal themselves in motion, not in a single paused image.
A useful final check compares the restored output against the source in the same representative scenes used to build the preset. The question is not whether every imperfection disappeared. It is whether distracting defects have been reduced while the film still looks like film, retains meaningful detail, and remains technically dependable for its next use.
The best batch workflow is one you can rerun confidently when another box of reels arrives: controlled enough to protect the images, flexible enough to respect the exceptions, and documented well enough that every decision remains clear.
