Style DNA

Understand the StyleJSON schema — the structured style fingerprint that Clipzy extracts from template videos and applies to your footage.

Style DNA is a structured JSON object that describes every measurable stylistic property of a video: its color grade, edit rhythm, audio treatment, motion intensity, transitions, and beat structure. Clipzy extracts Style DNA from a template video and can apply it to any footage you supply.

Think of Style DNA as a style manifesto — a complete, portable description of how a video looks and feels, independent of its content.

How customers use Style DNA

• Apply to your footage — submit a template video alongside your own clips, and the engine replicates the template’s Style DNA on your footage.
• Inspect and store — retrieve the Style DNA JSON from any completed job and store it in your own system for later use or auditing.
• Compare styles — diff two Style DNA objects to understand how two editors approach the same type of content differently.

The extraction_confidence field (0–1) tells you how reliably the engine analyzed the template. Higher-quality source videos with consistent style produce scores closer to 1.0.

Full example

{
  "version": "1.0",
  "color_grade": {
    "temperature": 10,
    "tint": 5,
    "saturation": 20,
    "contrast": 15,
    "highlights": 10,
    "shadows": -5
  },
  "aspect_ratio": "16:9",
  "frame_rate": 30,
  "audio_style": {
    "normalization_level": 0.85,
    "compression_ratio": 4.0,
    "bass_boost": 20,
    "enhance_speech": true
  },
  "cut_frequency": 3.5,
  "motion_intensity": 65,
  "zoom_usage": 30,
  "primary_transition": {
    "type": "fade",
    "duration_ms": 300,
    "easing": "ease-in-out"
  },
  "detected_beats": [
    { "timestamp_ms": 0, "confidence": 0.95, "frequency": "bass" }
  ],
  "tempo_bpm": 120.0,
  "music_genre": "pop",
  "created_at": "2024-04-05T10:30:00Z",
  "source_duration_seconds": 60.0,
  "extraction_confidence": 0.92
}

Schema reference

Top-level fields

version · string · required

Schema version. Currently always "1.0".

aspect_ratio · string · required

Detected aspect ratio of the template video. One of "16:9", "9:16", or "1:1".

frame_rate · number · required

Detected frame rate in frames per second. Typical values: 24, 30, 60.

cut_frequency · number · required

Average number of cuts per minute in the template video. A higher value means a faster-paced edit style. For example, 3.5 means roughly one cut every 17 seconds.

motion_intensity · number · required

Overall camera and subject motion on a scale of 0–100. 0 is completely static; 100 is near-constant high-speed movement.

zoom_usage · number · required

How frequently zoom effects appear, on a scale of 0–100. 0 means no zooms detected; 100 means zooms are present in nearly every shot.

use_subtitles · boolean · required

Whether the template video uses subtitles or captions.

tempo_bpm · number

Detected music tempo in beats per minute. null if no music was detected.

music_genre · string

Detected music genre (e.g., "pop", "hip-hop", "cinematic"). null if no music was detected or the genre could not be classified.

created_at · string · required

ISO 8601 timestamp of when the Style DNA was extracted.

source_duration_seconds · number · required

Duration of the template video in seconds.

extraction_confidence · number · required

Overall quality score for the extraction on a scale of 0–1. Scores above 0.85 indicate a reliable extraction. Scores below 0.6 suggest the template video may be too short, inconsistent, or low-quality for accurate style transfer.

---

color_grade

The color grade describes the visual tone of the template video. All fields use a scale of -100 to 100, where 0 is neutral.

color_grade · object · required

Hide properties

temperature · number · required

Warm/cool color bias. Positive values push toward warm (orange/yellow); negative values push toward cool (blue).

tint · number · required

Green/magenta color bias. Positive values push toward magenta; negative values push toward green.

saturation · number · required

Color intensity. Positive values increase vividness; negative values desaturate toward monochrome.

contrast · number · required

Difference between light and dark areas. Positive values increase contrast; negative values flatten the image.

highlights · number · required

Brightness of the brightest areas. Positive values recover blown highlights; negative values darken them further.

shadows · number · required

Brightness of the darkest areas. Positive values lift shadows; negative values crush blacks.

---

audio_style

The audio style describes how the template video’s audio was treated and mixed.

audio_style · object · required

Hide properties

normalization_level · number · required

Target loudness level on a scale of 0–1. 1.0 is fully normalized to a consistent broadcast-standard level.

compression_ratio · number · required

Dynamic range compression ratio applied to the audio. A value of 4.0 means a 4:1 ratio — loud sounds are reduced to one-quarter of their original dynamic range.

bass_boost · number · required

Low-frequency enhancement on a scale of 0–100. 0 means no bass boost applied.

enhance_speech · boolean · required

Whether the template uses speech enhancement processing to improve vocal clarity.

---

primary_transition and secondary_transition

Transitions describe how the template moves between shots.

primary_transition · object · required

The most commonly used transition style in the template video.

Hide properties

type · string · required

Transition type. Common values: "cut", "fade", "wipe", "dissolve", "zoom".

duration_ms · number · required

Duration of the transition in milliseconds.

easing · string · required

Animation easing curve. One of "linear", "ease-in", "ease-out", or "ease-in-out".

secondary_transition · object

The second most common transition style, if the template uses more than one. null if only one transition type was detected.

Show properties

type · string · required

Transition type.

duration_ms · number · required

Duration of the transition in milliseconds.

easing · string · required

Animation easing curve.

---

detected_beats

An array of beat events detected in the template’s audio track. The engine uses these to align cuts and transitions in the applied output.

detected_beats · object[] · required

Hide item properties

timestamp_ms · number · required

Position of the beat in the template video, in milliseconds from the start.

confidence · number · required

How confident the engine is that this is a true beat, on a scale of 0–1. Values above 0.8 are high-confidence.

frequency · string · required

The frequency band of the beat. One of "bass", "mid", or "treble". Bass beats are most commonly used to trigger cuts.