RESEARCH · 20

Production specs

Camera. Lighting. Audio. Editing pacing. Captions. Music. Export. The build sheet.

In brief
  • Camera: 4K/24, eye-level, medium close-up, 35-50mm. Tripod locked.
  • Audio: lav 6-8" below chin + phone backup. Soft room, HVAC off.
  • Captions: burned-in, Inter/Montserrat 700, white + black outline, orange highlight on 2-3 nouns/beat.
  • Levels: -16 LUFS integrated, -1 dBTP true peak. Export: H.264 1080p 24fps + 1:1 and 9:16 cut-downs.

Camera and framing

  • Camera: Mirrorless or modern phone at 4K/24 or 4K/30. Locked tripod. Eye-level (lens at Waheed's eye height).
  • Framing: Medium close-up. Top of head near the top of frame, chest visible, eyes on the upper-third line.
  • Lens / distance: 35 to 50 mm equivalent, 4 to 6 feet back. Avoid wide-angle face distortion.

Background

  • Real office, intentionally dressed.
  • Orange (#FF6600) and black brand element visible (the brand element on the wall, a print, anything that signals "this is New Limits").
  • 6+ feet separation between Waheed and the background for natural depth blur.
  • No virtual backgrounds. Ever. The edge-detection artifacts are immediately visible and signal amateur production.

Lighting

  • Key: one large soft key at 45 degrees camera-left.
  • Fill: bounce or reflector camera-right.
  • Hair / back: small back light or hair light to separate from background.
  • Face brighter than background by approximately one stop.
  • No windows behind talent (uncontrollable exposure swings).

Audio (non-negotiable)

  • Mic: lav clipped 6 to 8 inches below chin. Rode Wireless GO, DJI Mic, or equivalent.
  • Backup: simultaneous recording on phone as a backup track.
  • Room: soft-furnished. Kill HVAC and fridge for the take. Close doors. Soft-test by clapping once and listening for slap-back.

Wardrobe

  • Solid mid-tone shirt.
  • No fine stripes (moire on camera).
  • No competing logos.
  • No pure white (blows out under key light) or pure black (kills shadow detail).

Editing pacing

  • First jump-cut before 0:08. Industry rule of thumb across talking-head VSLs to lock retention.
  • Average shot length: 3 to 6 seconds in the front half, 6 to 10 seconds in the founder section.
  • Remove every "um" and every breath longer than 0.4 seconds.
  • Silence works once, intentionally: right before the offer reveal in Body Beat 5. Nowhere else.

Captions

  • Burned-in throughout. Not toggleable CC. 78%+ of paid social VSL views are sound-off.
  • Layout: single-line, two-row max, lower-third position.
  • Font: bold sans-serif (Inter or Montserrat, weight 700).
  • Style: white text with black outline.
  • Highlight: brand-orange (#FF6600) on 2 to 3 key nouns per beat. Suggested highlight words: leak, categories, refund, Manager, two weeks.
  • Hook only (0:00 to 0:15): word-by-word reveal style. Steady captions for the rest.

Music

  • No music in the hook (0:00 to 0:20). Voice alone.
  • Low-bed instrumental from 0:20 onward, ducked to -22 to -26 LUFS under voice.
  • Slight tension build under the screen recording section.
  • Drops out for 2 beats before the offer reveal in Body Beat 5.
  • Returns warmer for the guarantee and CTA.
  • Avoid: corporate stock acoustic, EDM, anything with vocals or melodic hooks.

Final audio levels

  • Integrated: -16 LUFS
  • True peak: -1 dBTP

Export

  • Codec: H.264
  • Resolution: 1080p
  • Frame rate: 24 fps
  • Encoding: 2-pass, ~12 Mbps for web
  • Masters: the primary 16:9 master, plus a 1:1 square cut-down and a 9:16 vertical cut-down from the same master for social repurposing.

Sources

  • Wistia and Vidalytics caption-strategy data (78% sound-off statistic).
  • Standard VSL production reference set: Stefan Georgi production notes, Frank Kern delivery archives, industry guidance on jump-cut cadence for talking-head retention.