User Guide
Everything you need to get started with Auto QC — from uploading your first video to reading detailed quality reports.
Getting Started
Auto QC automatically checks your videos for quality issues — so you can catch problems before delivery, not after.
What is Auto QC
Auto QC is an automated video quality control tool. Upload a video, choose your checks, and receive a detailed report highlighting any issues — from frozen frames and audio clipping to on-screen text errors.
How it works
Drag and drop your video file, or click to browse. Supported formats include MP4, MOV, MXF, and more.
Choose which quality checks to run, select your target platform, and enable any AI text detection you need.
Get a full quality report with an interactive timeline and an exportable PDF.
Who it's for
Auto QC is built for post-production teams, content distributors, and QC professionals who need reliable, repeatable quality checks at scale.
Dashboard
Your central hub for managing all video analyses — track progress, review results, and quickly access recent reports.
Analysis List
The analysis list shows all videos you have submitted, along with their current status. Click any row to open the full quality report. Use the search bar to filter by filename or date.
Analysis is currently running.
Report is ready to view.
Analysis could not be completed.
Stats Overview
Four cards at the top of the dashboard summarise your activity. These update automatically as you submit and complete analyses.
All videos submitted to your account.
Percentage of analyses with no critical issues.
The check that flags most frequently across your videos.
A summary of submissions in the last 7 days.
Upload & Settings
Submit a video for analysis in three steps: upload your file, configure the quality checks, then confirm and start.
File Requirements
Files that exceed these limits will be rejected during upload.
The Three-Step Flow
Drag your file onto the upload area or click Browse. A progress bar shows upload status. You can upload multiple files at once; each will be configured independently.
For each file, choose an analysis preset and customise individual checks. Select a target platform if you want compliance validation.
Review the file details and your chosen settings, then click Start Analysis. You'll be taken to the dashboard where you can track progress.
Analysis Presets
Choose a preset to quickly configure the checks that match your workflow.
Core video checks only (brightness, freeze frames, aspect ratio, borders). Fast and lightweight. Best for quick review.
Full video and audio checks. Recommended for most workflows. Covers all common quality issues.
Everything in Standard, plus AI text and subtitle detection. Use this for content with on-screen text, captions, or hard-coded subtitles.
Pick exactly which checks to run. Useful when you need a specific subset of checks for a particular workflow.
Platform Compliance
When you select a target platform, Auto QC validates your video against that platform's specifications. Platform selection also automatically adjusts other checks — for example, selecting TikTok enables vertical aspect ratio validation.
Analysis Results
The results page gives you three views of your quality report: a summary, an interactive timeline, and a PDF export.
General Report
The INFO tab gives you a complete overview of your video and the analysis findings.
Timeline Annotation
The TIMELINE tab provides a frame-accurate, interactive view of every issue in your video.
- A chronological list of every issue detected, with the exact timecode where it occurs.
- Click any issue to jump the video player to that moment.
- Use the filter bar to show only specific issue types — for example, only audio issues.
- Use search to find issues by type or keyword.
- Each issue card shows: type, severity (Warning or Critical), timecode, duration, and a short description.
- Mark any issue as Fixed (you have addressed it) or Skipped (reviewed and chosen to ignore). These statuses are saved and appear in your PDF export.
Export & PDF
The EXPORT tab lets you generate a PDF report tailored to your needs.
One-page summary with issue count.
Full report with issue list and key metadata.
Complete report including frame-accurate screenshots for each issue.
Plain issue list, ideal for sharing with editors.
Version Control
Upload revised videos and re-run quality checks without losing your previous dismiss states. Auto QC automatically re-applies dismissed errors to the new version, matching timecodes even when the edit shifted the timeline.
Uploading a new version
From the dashboard, open the action menu on any completed analysis and select “Upload New Version”. You can also access this from the Upload New Version button in the analysis result header. Settings from the previous version — detector toggles, platform compliance preset, PlanD configuration — are inherited automatically.
Open the action menu on a completed analysis → "Upload New Version"
Select the revised video file
Auto QC runs a full quality check with inherited settings
Dismissed errors are re-applied — only genuinely new issues appear
Error state inheritance
Dismissed errors carry over to the new version. Errors marked as corrected do not — those require confirmation that the fix actually landed in V2. Auto QC uses a three-layer matching system to decide how each error is re-applied.
Aspect ratio, Loudness
Type match only — no timecode needed
Black frames, Freeze frames, Audio clips, Blur...
Audio alignment → timecode translation → ±0.5s window
Typos, Safe zone violations
Text similarity + aligned timecode
Timeline alignment
When an edit shifts the timeline — cutting two seconds from the head, for instance — every QC timestamp in V1 is now offset from V2. Auto QC uses audio cross-correlation to find the exact offset before matching errors.
The method works after re-encoding because it operates on the RMS energy envelope rather than raw PCM samples. Energy is preserved across codec changes; typical alignment accuracy is within 0.02 seconds.
FFmpeg extracts 8kHz mono PCM from both versions
A 50Hz RMS energy envelope is computed (20ms windows)
The envelope is z-score normalised, then a sliding-window Pearson correlation is run
Forward and backward offsets up to 60 seconds are searched
V1 timecodes are translated into the V2 timeline before error matching
When alignment confidence is below 0.65 — for example, if the audio changed significantly or the video has no audio — Auto QC falls back to inheriting global errors only (aspect ratio, loudness) and shows a manual review prompt in the analysis result header.
Version dashboard
The dashboard always shows the latest version of each video. Previous versions are archived and accessible from the analysis result page. A version badge on each card makes clear which revision you're looking at.
Version badge on each dashboard card — always reflects the latest version.
Accessible from the analysis result page — switch between all past versions.
Older versions are archived. The dashboard only shows the latest.
Video Detectors
Auto QC runs up to seven video quality checks. Each can be individually enabled, disabled, and tuned in the analysis settings.
Bad Edit Detection
Flags cuts where an unusually brief shot — typically lasting only a handful of frames — appears between two larger scene changes. These micro-clips often indicate an unintended edit, a dropped frame cluster, or a rendering artefact.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Sensitivity | Medium | Low / Medium / High | How aggressively short transitions are flagged. Higher sensitivity catches more potential issues but may increase false positives on fast-cut content. |
| Minimum Gap Duration | 5 frames | 1 – 15 frames | The maximum length of the in-between shot that triggers a flag. Shots shorter than this threshold will be flagged. |
Sensitivity
How aggressively short transitions are flagged. Higher sensitivity catches more potential issues but may increase false positives on fast-cut content.
Minimum Gap Duration
The maximum length of the in-between shot that triggers a flag. Shots shorter than this threshold will be flagged.
Each flagged edit appears on the Timeline as a Warning or Critical card, showing the exact timecode, the duration of the anomalous shot, and the severity level.
Freeze Frame Detection
Detects sequences where the video image is completely or near-completely static for longer than expected — a common sign of a technical fault, signal dropout, or export error.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Minimum Freeze Duration | 2 seconds | 0.5s – 10s | How long the video must be static before it is flagged. Short holds or slow-motion shots may need a higher threshold to avoid false positives. |
| Sensitivity | Medium | Low / Medium / High | Controls tolerance for minor pixel variation. At Low, only completely static images are flagged. At High, near-static images with small noise patterns are also caught. |
Minimum Freeze Duration
How long the video must be static before it is flagged. Short holds or slow-motion shots may need a higher threshold to avoid false positives.
Sensitivity
Controls tolerance for minor pixel variation. At Low, only completely static images are flagged. At High, near-static images with small noise patterns are also caught.
Freeze events appear on the Timeline with their start timecode and duration. Very long freezes are marked Critical; shorter ones are marked Warning.
Blur Detection
Identifies frames where the image sharpness falls below an acceptable level, which may indicate out-of-focus recording, excessive motion blur, or compression degradation.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Sensitivity | Medium | Low / Medium / High | The threshold at which blur is considered a reportable issue. Increase sensitivity for content that requires high sharpness (e.g. text-heavy graphics). Decrease it for intentionally stylised or shallow-depth footage. |
Sensitivity
The threshold at which blur is considered a reportable issue. Increase sensitivity for content that requires high sharpness (e.g. text-heavy graphics). Decrease it for intentionally stylised or shallow-depth footage.
Blurred frames are marked on the Timeline by timecode. Each card indicates the severity of the blur and the affected duration.
Brightness Analysis
Checks for frames that are significantly overexposed (too bright) or underexposed (too dark). Both extremes can make content unwatchable and may indicate a grading error or a source file problem.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Black Frame Threshold | 10 / 255 | 2 – 30 | Frames where average brightness falls below this value are flagged as underexposed or blacked out. |
| White Frame Threshold | 245 / 255 | 200 – 254 | Frames where average brightness exceeds this value are flagged as overexposed or blown out. |
| Minimum Duration | 0.5 seconds | 1 frame – 5s | How long a brightness anomaly must persist to be reported. |
Black Frame Threshold
Frames where average brightness falls below this value are flagged as underexposed or blacked out.
White Frame Threshold
Frames where average brightness exceeds this value are flagged as overexposed or blown out.
Minimum Duration
How long a brightness anomaly must persist to be reported.
Brightness issues appear on the Timeline labelled as 'Black Frame', 'White Frame', or 'Overexposure'. Each entry shows the affected timecode range and average brightness value.
Aspect Ratio Check
Verifies that the video's aspect ratio matches your intended output specification throughout the entire duration. Unexpected ratio changes can cause letterboxing, pillarboxing, or cropping on delivery platforms.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Expected Ratio | 16:9 | 16:9 / 9:16 / 4:3 / 1:1 / Custom | The ratio your finished video should conform to. Set to match your delivery requirement. |
| Tolerance | ±1% | 0% – 5% | Allowable deviation from the target ratio before an issue is reported. A small tolerance accommodates minor encoding rounding. |
Expected Ratio
The ratio your finished video should conform to. Set to match your delivery requirement.
Tolerance
Allowable deviation from the target ratio before an issue is reported. A small tolerance accommodates minor encoding rounding.
A Pass or Fail result is shown in the General Report. If the ratio is incorrect, the TIMELINE tab shows which sections are non-conforming and what ratio was detected.
Border Detection
Scans for persistent black bars or coloured borders around the edges of the frame. These often appear when widescreen content is incorrectly letterboxed, when a source clip has embedded borders, or after an incorrect export.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Border Tolerance | 8 pixels | 0 – 40 px | Borders narrower than this value are ignored. Increase the tolerance if your content intentionally includes a thin frame or vignette. |
| Minimum Duration | 1 second | 1 frame – 10s | How long a border must be present before it is reported. |
Border Tolerance
Borders narrower than this value are ignored. Increase the tolerance if your content intentionally includes a thin frame or vignette.
Minimum Duration
How long a border must be present before it is reported.
Detected borders are shown on the Timeline with their location (top / bottom / left / right / all sides), width in pixels, and the timecode range during which they appear.
Platform Compliance
Validates your video's technical specifications against the official requirements of your chosen delivery platform. Catches resolution, frame rate, and format mismatches before you submit.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Target Platform | YouTube | YouTube / TikTok / Instagram / Broadcast TV / Custom | The platform whose spec sheet your video is validated against. Select Custom to define your own resolution, frame rate, and format requirements. |
Target Platform
The platform whose spec sheet your video is validated against. Select Custom to define your own resolution, frame rate, and format requirements.
A compliance table appears in the General Report showing Pass, Warning, or Fail for each specification (resolution, frame rate, codec, bitrate, colour space, etc.). Failed items are also listed in the Timeline as General Error cards.
Audio Detectors
Five audio checks cover common delivery problems — from clipping and silence to loudness compliance and speech intelligibility.
Audio Clipping
Detects moments where the audio signal peaks beyond the maximum level, causing distortion commonly heard as crackling or 'hard clipping'. This is one of the most common technical audio faults in delivered content.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Clipping Threshold | −1.0 dBFS | −6.0 – 0.0 dBFS | Signal peaks above this level are flagged as clipped. Tighter thresholds (e.g. −3 dBFS) are recommended for broadcast delivery. |
| Minimum Duration | 10 ms | 1 ms – 1 s | How long a clipping event must last to be recorded. Very brief peaks (under 10 ms) may be inaudible and can be excluded. |
Clipping Threshold
Signal peaks above this level are flagged as clipped. Tighter thresholds (e.g. −3 dBFS) are recommended for broadcast delivery.
Minimum Duration
How long a clipping event must last to be recorded. Very brief peaks (under 10 ms) may be inaudible and can be excluded.
Clipping events appear on the Timeline with their exact timecode, peak level in dBFS, and duration. Severe or prolonged clipping is marked Critical.
Silence Detection
Identifies sections of the audio track that are completely silent or near-silent for longer than expected. These gaps may indicate a missing audio segment, an edit error, or an encoding fault.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Silence Threshold | −50 dBFS | −70 – −30 dBFS | Audio below this level is considered silent. Raise the threshold to flag near-silent passages; lower it to ignore low-level ambience. |
| Minimum Duration | 2 seconds | 0.5s – 30s | How long a silent section must be before it is flagged. Short pauses in dialogue should not trigger this check. |
Silence Threshold
Audio below this level is considered silent. Raise the threshold to flag near-silent passages; lower it to ignore low-level ambience.
Minimum Duration
How long a silent section must be before it is flagged. Short pauses in dialogue should not trigger this check.
Silent sections are shown on the Timeline with start timecode and duration. Sections longer than 5 seconds are marked Critical by default.
Background Noise
Measures the level of continuous background noise — room tone, HVAC hum, electrical hiss, traffic, or similar ambient sound. High background noise can reduce perceived audio quality and affect platform acceptance.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Noise Threshold | −40 dBFS | −60 – −20 dBFS | The average background noise level above which an issue is reported. Adjust based on your content type: documentaries may tolerate higher ambient noise than studio productions. |
| Sensitivity | Medium | Low / Medium / High | Controls how aggressively noise is isolated from the signal. Higher sensitivity is better for content with frequent speech pauses. |
Noise Threshold
The average background noise level above which an issue is reported. Adjust based on your content type: documentaries may tolerate higher ambient noise than studio productions.
Sensitivity
Controls how aggressively noise is isolated from the signal. Higher sensitivity is better for content with frequent speech pauses.
A noise level score (Low / Moderate / High) is shown in the General Report. Affected time ranges appear on the Timeline for content with elevated noise.
Loudness Compliance
Checks that the overall loudness of your audio track conforms to the standard used by broadcast networks and streaming platforms. Non-compliant loudness is a common cause of content rejection.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Loudness Standard | EBU R128 (−23 LUFS) | EBU R128 (−23 LUFS) / ATSC A/85 (−24 LUFS) / Custom | The loudness specification to validate against. Use EBU R128 for European broadcast and most streaming platforms. Use ATSC A/85 for North American broadcast. |
| Tolerance | ±1 LU | ±0.5 – ±2 LU | Acceptable deviation from the target integrated loudness. Broadcast deliverables typically require tighter tolerance. |
Loudness Standard
The loudness specification to validate against. Use EBU R128 for European broadcast and most streaming platforms. Use ATSC A/85 for North American broadcast.
Tolerance
Acceptable deviation from the target integrated loudness. Broadcast deliverables typically require tighter tolerance.
A single Pass / Fail result is shown in the General Report, along with the measured integrated loudness (in LUFS), true peak level, and loudness range (LRA). Non-compliant results also appear as a General Error card in the Timeline.
Speech Clarity
Evaluates how intelligible the spoken dialogue or narration is relative to the audio background. Low scores can indicate excessive background noise, music that overpowers speech, or poorly recorded dialogue.
This check runs automatically with no configurable parameters.
A clarity score (Excellent / Good / Acceptable / Poor) is shown in the General Report. Sections where clarity drops below Acceptable are listed on the Timeline with their timecode.
AI Text Detection
Three AI-powered checks scan on-screen text and subtitles for spelling errors, safe zone violations, and extraction. Enable these under the Advanced preset or in Custom settings.
Typo Detection
Reads on-screen text and embedded subtitles frame by frame, then flags words that appear to be misspelled or grammatically incorrect. Supports over 100 languages.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Language | English (en) | 100+ languages supported | The primary language of your on-screen text. Set this to match your subtitle language for the most accurate detection. |
| Confidence Threshold | Medium | Low / Medium / High | How confident the system must be before flagging a word as a potential error. Lower confidence surfaces more candidates; higher confidence reports only clear mistakes. |
Language
The primary language of your on-screen text. Set this to match your subtitle language for the most accurate detection.
Confidence Threshold
How confident the system must be before flagging a word as a potential error. Lower confidence surfaces more candidates; higher confidence reports only clear mistakes.
Each flagged word appears as a Timeline card with the timecode, the detected text, the suspected error highlighted in context, and a suggested correction. For subtitle content, the full subtitle line is shown with the problematic word underlined.
Safe Zone Violation
Checks whether on-screen text, graphics, or subtitles are positioned within the platform's required safe zone. Content outside the safe zone may be cut off, overlapped by platform UI, or flagged during QC review.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Safe Zone Standard | Action Safe (90%) | Action Safe (90%) / Title Safe (80%) / Platform Specific / Custom | The safe zone boundary to validate against. Action Safe is the default for most delivery; Title Safe is stricter and recommended for subtitle-heavy content. |
| Target Platform | All Platforms | YouTube / TikTok / Instagram / Broadcast / All | Validates against the overlay areas specific to the chosen platform (e.g. TikTok username and caption regions). |
Safe Zone Standard
The safe zone boundary to validate against. Action Safe is the default for most delivery; Title Safe is stricter and recommended for subtitle-heavy content.
Target Platform
Validates against the overlay areas specific to the chosen platform (e.g. TikTok username and caption regions).
Violations appear on the Timeline with their timecode, a description of what element is outside the zone, and its distance beyond the boundary in pixels. A frame screenshot is included in Professional PDF exports.
Subtitle Extraction
Automatically reads and extracts all subtitles and on-screen text from your video into a structured list. Useful for reviewing subtitle timing, content accuracy, and for generating a reference transcript.
| Parameter | Default | Range | Description |
|---|---|---|---|
| Language Filter | All Languages | Any supported language | Filter extracted text to a specific language if your video contains mixed-language content. |
Language Filter
Filter extracted text to a specific language if your video contains mixed-language content.
All recognised text events are listed in the Timeline tab under a dedicated Subtitles section, each with its timecode, duration, and the detected text. The full list can be downloaded as a plain-text file from the Export tab.
Logo & Brand Detection
Two AI-powered checks scan every sampled frame for brand logos and on-screen branded products, flagging potential copyright and endorsement risks before publishing. Enable these under the PlanD & AI Detection settings.
Logo Copyright Risk
Detects brand logos and trademarks visible in the frame that may require licensing clearance. Each detection is traced across consecutive frames to confirm it is a genuine, sustained appearance rather than a flash or artefact.
This check runs automatically with no configurable parameters.
Each detection appears as an Amber timeline marker with the timecode and duration of the logo appearance. A dashed bounding box tracks the logo position in the video player as you scrub through the flagged segment. The error card shows the model confidence score and detected label.
Brand Product Risk
Identifies full branded products in frame — laptops, phones, packaging, and similar items — where the brand is clearly visible and may imply endorsement. Distinct from logo copyright risk, which targets trademark symbols specifically.
This check runs automatically with no configurable parameters.
Flagged segments appear as Orange timeline markers. Each card includes a recommendation to confirm rights or obscure the product before publishing. The bounding box in the video player shows the tracked product region during playback.
Subscription Plans
Auto QC offers three tiers. All plans include all 14 quality detectors — what changes is monthly quota, storage, and advanced features.
Free
$0
10 min / month · 1 GB storage
All 14 detectors, basic PDF report, timeline annotations.
Pro
$29.99/mo
60 min / month · 10 GB storage
All detectors, full PDF report (screenshots + waveforms).
Ultra
$89.99/mo
300 min / month · 50 GB storage
All detectors, full PDF report, priority queue, API access.
How to upgrade
Go to the Pricing page and compare plans.
Click Upgrade to Pro or Upgrade to Ultra.
Complete payment in Stripe Checkout.
Your quota updates the moment payment clears.
Billing & invoices
All billing is handled by Stripe. Auto QC does not store your card details. Every successful payment generates a PDF invoice listed under Settings → Billing — downloadable any time.
To update your payment method, switch plans, or cancel, click Manage Subscription on the Subscription settings page. This opens the Stripe Customer Portal directly. Cancellations take effect at the end of the current billing period.
Quota management
Quota is measured in video minutes analysed per month. Analysing a 5-minute video consumes 5 minutes of your allowance. Quota resets automatically on the first day of each billing cycle.
The Settings → Subscription page shows a live progress bar for both analysis minutes and storage. Check it before uploading a large batch. If you run out of quota mid-cycle, upgrade from the same page — your new allowance activates immediately.