AI summary by TwinMind

What is Happening with iPhone Camera?

Marques Brownlee

13m 39s

iPhone Camera Performance Issues

  • iPhone consistently loses in blind smartphone camera tests despite premium reputation
  • Finishes in middle of pack in scientific test with 20+ million votes
  • Yet received "best overall smartphone camera system" award for 4th consecutive year
  • Growing perception that iPhone camera has taken steps backward recently

Evolution of Smartphone Camera Technology

  • Traditional cameras: simple sensor exposed to light, direct representation of light hitting sensor
  • Modern smartphones: computational photography with multiple rapid exposures
  • Current process includes:
    • Sampling environment multiple times in rapid succession
    • Different exposure speeds
    • Tone mapping and noise reduction
    • HDR processing
    • Software merging all data into "best looking image"
  • Software now makes bigger difference than hardware/sensor quality

Google Pixel Camera Development Case Study

  • Google used IMX 363 sensor starting with Pixel 3
  • Perfect software tuning created instant success
  • Same sensor/software combo used across: Pixel 3, 3a, 4, 4A, 5, 5A, 6A
  • Pixel 6A won December scientific blind test with 4-year-old sensor
  • Success due to nailing basic photography fundamentals better than competitors

Pixel 6 Transition Problems

  • Google upgraded to massive 50MP sensor from 12MP
  • New sensor caused over-processing issues:
    • Over-sharpening, "crunchy" appearance
    • HDR-heavy processing ("hdria" look)
    • Running algorithms at maximum intensity unnecessarily
  • Larger sensor gathered more light but software still processed as if light-starved
  • Result: over-processed photos, arguably worse than older Pixel models
  • Required full year of software adjustments to fix

Pixel 7 Recovery

  • Same 50MP sensor as Pixel 6
  • Software finally dialed in properly
  • Pixel 7 finished right behind Pixel 6A in blind test

iPhone 14 Pro Similar Transition Issues

  • Used small 12MP sensor for years
  • iPhone 13 Pro: slightly larger sensor
  • iPhone 14 Pro: first dramatic jump to 48MP sensor
  • Experiencing similar over-processing issues as Pixel 6:
    • Inconsistent photo quality
    • Over-processed appearance
    • Nothing extreme but noticeable
  • Prediction: iPhone 15 Pro will introduce new software improvements with marketing buzzword

Blind Test Methodology and Human Subject Impact

  • Test photos designed with multiple judging factors:
    • Two different skin tones
    • Two different colored shirts
    • Texture details for sharpness testing
    • Sky for dynamic range/HDR testing
    • Short and long range depth of field falloff
  • Human subjects chosen because most important photos are of people
  • Using creator's face revealed smartphone face processing differences

iPhone Face Processing Characteristics

  • iPhone automatically identifies and evenly lights all faces
  • Attempts to remove natural shadows and lighting variations
  • Example: Low light test showed iPhone completely removed natural shadow from half of face
  • Creates unrealistic lighting that doesn't match actual light source direction
  • Sometimes produces "weird" results due to aggressive face lighting

Skin Tone Representation Differences

Google Real Tone Technology

  • Google trains camera software on diverse skin tone datasets
  • Aims for accurate representation across all skin tones
  • Addresses historical issue of film cameras being calibrated for lighter skin tones
  • Significant factor in Pixel cameras performing well in blind tests

Apple's Approach

  • Focuses on evenly lighting faces across the board
  • Doesn't account for different white balance/exposure needs for various skin tones
  • Less sophisticated skin tone representation compared to Google's Real Tone

Regional Variations

  • Phones sold in China often brighten faces universally (regional preference)
  • Different manufacturers make different software choices for face processing

Camera Comparison Complexity

  • Modern smartphone photo comparisons extremely difficult due to software differences
  • Users may prefer different phones for different subjects:
    • One phone better for landscapes
    • Different phone better for personal skin tone
    • Another better for pet photography
  • All due to varying software processing approaches

Scientific Test vs Overall Camera System Award Explanation

Scientific Test Limitations

  • Only tested small compressed images (postage stamp size)
  • Focused on exposure, colors, general factors
  • Did NOT test:
    • Sharpness and detail (lost in compression)
    • Autofocus speed and reliability
    • Camera app launch speed
    • Shot-to-shot reliability
    • Video quality and microphone quality
    • File formats and video HDR

iPhone 14 Pro Award Justification

  • Despite middle-of-pack performance in blind test
  • Superior performance in untested categories
  • Overall camera system includes video, speed, reliability factors
  • Award based on comprehensive evaluation beyond single photo quality test

Future Predictions

  • Apple will continue improving iPhone camera software
  • iPhone 15 Pro expected to introduce new camera software features
  • Ongoing software updates will improve current iPhone 14 Pro performance
  • Modern smartphone photography is "partly photons but primarily processing"

TwinMind

Get summaries like this for your own meetings

TwinMind takes notes in the background during meetings, lectures, and calls, then writes the summary for you.

Free plan to get started

No bot joins your calls

Phone, desktop & Chrome apps

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome