Tools / AI Tools / Cinematography Analyzer
PIXELOKAY / CINEMATIC INTELLIGENCE ✦ CREATIVE STUDIO

Cinematography Analyzer

Inspect framing, tonal contrast, color and texture. Add creative context, compare frames and create production-ready shot notes—without uploading your footage.

Composition & thirdsLighting & colorFrame-by-frame reportsOptional on-device AI
ANALYSIS WORKSPACE / 01

Cinematography Analyzer Smart Studio

Import a frame, inspect the evidence, and add your own artistic interpretation.

01

Frame viewer

No source loaded

Ready
◫
Drop images or a video here

Up to 12 PNG/JPEG/WebP images at once, or one video

Private by default · 220 MB per file
0 / 0
No frames

Click the frame to position your focal point. The orange dot is an edge-based texture centroid, not a detected subject.

02

Import & shot sequence

Import multiple images or collect up to 12 editable frames.

Captured frames will appear here.

03

Measured visual signals

Pixel statistics and computational guides, not scene-semantic facts.

Brightness—of 255
Contrast—luminance σ
Balance—left/right
Warm / cool—red − blue
Luminance histogram64 sampled bins
MeasurementValueInterpretation limits
Add a frame to view measurements.
Shot-to-shot comparisonDifferences, not rankings
Add at least two frames to compare tonal and color continuity.
Sequence continuity

Add another frame to review adjacent-shot changes.

Review cues

Not a quality score

Analyze a frame to see context-sensitive review prompts.

Choose a local image or video to begin.

Make deliberate choices, shot by shot.

◎

Composition and focus

Review the rule of thirds, visual-weight distribution and an editable focal anchor. Interpret guides in the context of your scene.

◑

Color and lighting

Compare luminance, highlights, shadows, color bias and brightness asymmetry without pretending to recover physical lights.

▤

Storyboards and reports

Build a sequence from still images or captured video frames, and export editable scene notes with measurable evidence.

Frequently asked questions

Can this tell the camera angle or lighting direction automatically?

Not reliably from a single arbitrary frame. The tool measures tonal gradients, edge texture and composition proxies. Camera angle, physical light direction, depth and storytelling intent are explicitly editable human interpretations.

Is there real AI?

Yes. Model Setup offers an explicit opt-in download of a free MediaPipe object detector. Detect subjects then uses that loaded model. Automated color and composition metrics work offline without the model. Live model integration requires internet access and browser verification.

Is my footage uploaded?

No footage or images are uploaded by the application. On-demand AI inference uses locally processed frames but fetches third-party model/runtime assets and MediaPipe may collect runtime metrics.

Are these measurements suitable for grading and mastering?

Measurements are approximate sRGB pixel statistics, not calibrated HDR, RAW or camera-lighting measurements. There is no depth reconstruction, shot reverse-engineering or story understanding without additional validated models and context.

LOCAL AI SETUP

Set up AI on this device

First, a model needs to be loaded on your device

Cinematography Analyzer uses an on-device MediaPipe object detector. The model and WebAssembly runtime download only after you choose Download & Load. Your images and video stay in your browser.

  1. 1. Check device
  2. 2. Download / cache
  3. 3. Prepare detector
  4. 4. Detect objects

Device & model resources

Measurements may be approximate
Device RAM / model memoryChecking…

Live RAM and model working-memory use unavailable.

Checking MediaPipe CPU/WASM support…

Browser website storageChecking…

Site-wide estimate includes all same-origin PixelOkay tools, and is not RAM.

ONE ACTION · ALL RECOGNIZED AI MODEL DOWNLOADS

Free AI model storage across PixelOkay

Scan this website origin for WebLLM and Transformers.js model caches and recognizable model files, including other PixelOkay tools. Only positively identified model artifacts in Cache Storage can be removed here, including those cached by other tools. Inaccessible IndexedDB, OPFS, projects, drafts and other user data remain untouched.

Run a scan to identify safely removable model artifacts. This release does not delete unverified browser databases.

Cache status not yet verified. Refresh Storage & Cache to check the selected model.

Preparing model setup…

Having trouble downloading or loading AI?

Refresh Storage & Cache, then retry or choose the model again. If storage is the problem, inspect recognized model caches before any deletion. Runtime/worker and network errors are not generally solved by clearing caches.

Do not clear all website data: that can remove saved work across PixelOkay. Only safely identified model assets are considered here.

Technical details and diagnostics

MediaPipe Tasks Vision is pinned to 0.10.22 and EfficientDet Lite0 INT8 uses CPU/WASM. The model file is reusable only with its exact URL and verified TFLite bytes. Cross-runtime WebLLM/Transformers.js weights cannot be used as an object detector. WebGPU is not required.

CROSS-TOOL MODEL CACHE CLEANUP

Delete recognized model downloads?

Other PixelOkay AI tools may need to re-download model files. Close active AI tabs first. Unknown caches and IndexedDB are untouched; never clear general website data without exporting your work.

Using Cinematography Analyzer

Review a sequence of stills or captured video frames with pixel-derived color, contrast and clipping measurements, then add your own focal and story annotations. Optional object proposals are a separate model-assisted step.

  1. Import still images or capture selected frames from a video; keep the sequence to the supported twelve compressed stills.
  2. Compare the measured image signals, add focal anchors and director notes, and request optional object proposals only if you choose model setup.
  3. Export annotated images, a contact sheet, comparison reports or a project containing frame JPEGs and notes.

Example: Compare a warm and a cool setup from the same scene, mark the intended focal point in each frame, and export a contact sheet with your notes for the crew.

Limits and privacy: Metrics describe compressed image pixels; they cannot determine camera angle, scene depth, incident lighting or artistic quality. Projects can embed up to twelve frame JPEGs and notes rather than the original video. Optional MediaPipe and EfficientDet files download from external hosts after an explicit action and may disclose the visitor IP; real model inference remains unverified.

Related tools