Free Video Enhancer

An enhanced frame is only as good as its record

Every enhancement changes pixels. A brighter frame, a steadier clip or a sharper plate is a derivative made for viewing, and the moment it is shown to anyone the next question is what was done to produce it and whether the same steps would produce the same picture again. That is the reason this tool exists. It applies filters in one fixed order, records the exact numeric setting of every step, names the interpolation and the color space it worked in, hashes the original and the output, and writes the frames it processed into the report, so the answer to that question is a document rather than a memory.

Keep the original with its collection context and treat the enhanced file as a viewing aid. When a frame matters to a matter, an examiner can tell you on a free first call which steps are defensible and how to preserve the footage before anything is enhanced.

Twelve steps in a fixed order

The order never changes because the order changes the result. Duplicate frame removal runs first so that repeated frames, which many recorders and exports produce, do not weight the temporal filters; the tolerance you set is the mean absolute difference in brightness below which a frame counts as a repeat. Deinterlacing by field follows for footage recorded as two fields, keeping the field you name by line doubling or blending. Rotation and crop square the frame and remove what you do not need before anything is resized. Resize scales the frame by the factor you choose with the interpolation you name, nearest neighbor, bilinear or bicubic, and the report states which. Brightness adds a fixed offset and, when you enable it, equalizes the luma histogram with the strength you set; contrast scales the frame around the midpoint by the factor you set.

Denoise runs a spatial median over the radius you choose and a temporal median across the number of neighboring frames you choose on each side, which is the stationary camera case the guide describes. Frame averaging takes the mean of the frame and a stated number of neighbors on each side, and the report lists the frames that went into every output frame. Unsharp masking and deblurring sharpen with the radius, amount and deconvolution iterations you set. Deblocking smooths the seams between compression blocks of the size you name. Channel isolation and grayscale select the channel or the luma you want to read. Stabilization runs last, estimating the shift of each frame against the reference you choose and moving it back by whole pixels, with the shift recorded frame by frame.

Frame averaging for a stationary scene

The oldest trick in surveillance enhancement is also the most defensible, because it adds nothing that was not in the footage. When the camera and the subject hold still, each frame carries the same scene under different noise, and averaging several frames keeps the scene and divides the noise. Enable frame averaging, set the number of neighbors on each side, and the tool averages that many frames before and after the target, records the exact frames it used and marks the result as the mean of those frames.

The report shows the original and the averaged frame side by side. The averaged frame is a derivative, and if the subject moved during those frames the average blurs it, which is why the number of frames and their timestamps travel with the output.

Reading a plate without claiming what it says

The VideoCleaner user guide walks through a license plate tutorial, and the sequence it teaches is the one this tool supports: crop to the plate, resize with a named interpolation, average the stationary frames, adjust brightness and contrast, then sharpen lightly. Run those steps here and the report records every setting, so a reader who disputes the enhanced plate can reproduce it or challenge a step.

What the tool does not do is read the plate or claim what the result shows. The enhanced frame is a processed image for a person to examine, and the report says exactly how it was processed and nothing about what it depicts. When a plate, a face or a serial number carries the matter, bring the original and the report to an examiner for media authentication, because an enhancement that cannot be reproduced and explained will not survive cross examination.

The original stays untouched and both files are hashed

The tool never writes into the file you selected. The original is read, hashed with SHA 256 and left alone; the enhanced still is written as a new PNG file and the enhanced clip, where your browser can encode one, as a new WebM or MP4 file, and each output is hashed before you save it. The report carries the original hash, the output hashes, the range and frames processed, the decoder and browser version, and the plain statement that the output is a derivative for viewing and not the evidence. Keep the original and its collection context, keep the report with the output, and produce them together.

Your files and reports remain local

The selected clip, file names, hashes, decoded frames, enhanced frames, notes and report data never reach a server. Optional usage measurement stays off unless you allow Analytics. With permission it sends GDF only the tool name and fixed action labels, such as analysis completed or report prepared, never an input, result, filename or visitor identifier. Google and advertising scripts do not load on this page. Use Cookie settings below to change your choice. Program files, the parser and fonts download from this website; nothing about your footage travels with them.

Clear cancels active work and removes the displayed records without deleting your original or any file you saved. Saved reports can contain private file names, hashes and notes; review them before sharing.

Save the report

Download the PDF or HTML report in the Global Digital Forensics four section layout: the executive triage and file identity with the full SHA 256 of the original, the plain English interpretation guide, the structured findings inventory with every filter in order and its exact settings, the output hashes, the frames processed and your notes, and the evidentiary scope advisory with the beta notice. The Excel ready CSV gives one column per field and the JSON is the lossless record of hashes, container and track observations, the settings of every step, the frames each output frame was built from, the encoder used and the browser version. The PDF and HTML reports carry the side by side figure of the original and enhanced frame; no report contains the video itself.

Formats and resource requirements

Each clip is limited to 64 MiB and each decoded frame to 2.1 million pixels, so a 1920 by 1080 frame is the largest the tool will process. Each range is at most 60 seconds and 120 frames, the resize factor runs from one quarter to four times with an output limit of 4.2 million pixels, the temporal radius and the averaging neighbors are limited to twelve frames on each side, and the operation time limit is 120 seconds per stage. Limits stop work with an explanation rather than a shortened result. When the page opens it tests for hardware graphics acceleration and, once a clip is read, for a hardware decoder for its codec; when either is missing the workbench says so, because the work still runs in software but takes longer.

What the report does and does not say

The report records what the tool did to the footage: the filter order, every setting, the interpolation, the color space, the frames processed, the frames averaged, the shifts applied and the hashes of the original and the output. It does not say what the enhanced frame shows, whether the footage is authentic or whether any step improved anything, because those are judgments for a person and, where a matter turns on them, for an examiner. The output is a derivative for viewing and not the evidence. Tell us about a problem with the tool; the link attaches no selected file or report.

Related tools and methods

Video Forensics gives the manual frame workbench with fifteen inspection views, Video Analyzer the guided workflow with optional local models and the Surveillance Video Tamper Check the export protection, stream structure, frame test and clock layers for CCTV exports; reselect your clip there, because inputs are never placed in links. Photo Forensics applies error level and noise analysis to still images. When a matter turns on the footage, expert witness services begin with a full examination.

Versions and methods

Local container parsing uses Mediabunny and browser decoding and encoding are described in WebCodecs. The filter order follows the enhancement workflow in the VideoCleaner user guide and the methods are standard image processing: temporal median and mean filtering, Gaussian and median smoothing, unsharp masking, Richardson Lucy deconvolution, histogram equalization, bicubic interpolation after Keys, and block matching for stabilization, with the Scientific Working Group on Digital Evidence practices for video enhancement as the frame for the record the tool keeps. These references describe the methods; they do not certify this tool. Version 1.0.0 beta. All free tools.