14:32:08 GMTSUN · MAY 10
USDEN
HowPhones
SEC 01METHODOLOGY · v4.2— PUBLISHED 12 JUN 2026 · 47 DIMENSIONS · 6 CATEGORIES
← Back to the lab

Forty-seven points.
One score.

Every HowPhones score is the weighted sum of forty-seven scored dimensions across six categories. The weights, the floors, the ceilings, and the formulas are all published below — versioned since 2023, audited annually by an independent test lab.

See all 47 dimensionsOr read the version log →
47
DISPLAY15%
BATTERY20%
PERFORMANCE15%
CAMERA25%
AUDIO8%
SOFTWARE17%
SEC 02/CONCEPTUAL-FRAME

How the score works.

3 PRINCIPLES
PRINCIPLE 01

Six categories, weighted.

Camera carries 25% of the total score. Battery 20%. Software 17%. Display 15%. Performance 15%. Audio 8%. The weights were chosen in 2023 after a reader survey of what actually matters; they're revisited every twelve months.

CAMERA WEIGHT
25%
PRINCIPLE 02

Per-dimension scoring.

Inside each category, every measurement gets a 0–10 score. Two phones with the same overall mark can fail or excel in completely different places. Click any row in the table below to see the exact formula, anchor points, and the scoring window.

DIMENSIONS
47
PRINCIPLE 03

Floor & ceiling.

A 10 is calibrated against the best phone we've measured in that dimension over the last three years; a 0 is the worst. New devices that exceed the current 10 reset the ceiling, and the previous record-holder's score is adjusted. Public log.

CEILING WINDOW
3 YR
SEC 03

Six categories.

WEIGHTS REVISED ANNUALLY · LAST AUG 2025
Display8 DIMENSIONS
WEIGHT
15%

Peak brightness, sustained brightness, color accuracy, gamut coverage, refresh-rate stability, touch latency, outdoor visibility, HDR.

SAMPLE DIMENSIONS
  • · Peak brightness
  • · Color accuracy ΔE
  • · Touch latency
  • · Outdoor visibility
Battery8 DIMENSIONS
WEIGHT
20%

Web, video, gaming, mixed-use endurance; wired and wireless charging; standby drain; thermal behavior during charging.

SAMPLE DIMENSIONS
  • · Web browsing
  • · Gaming endurance
  • · Wired charging speed
  • · Charging temperature
Performance7 DIMENSIONS
WEIGHT
15%

Geekbench single/multi, 3DMark Solar Bay, AnTuTu 11, sustained 30-minute stress, app launch speed, UI fluidity.

SAMPLE DIMENSIONS
  • · Geekbench 6 multi
  • · 3DMark Solar Bay
  • · Sustained stress
  • · UI fluidity
Camera12 DIMENSIONS
WEIGHT
25%

Main, ultrawide, telephoto across daylight, low-light, night. Front camera. Video 4K stability, color science, audio capture, Pro/Log support.

SAMPLE DIMENSIONS
  • · Main daylight
  • · Main night mode
  • · Telephoto 10x+
  • · Video 4K stabilization
Audio5 DIMENSIONS
WEIGHT
8%

Speaker SPL, frequency response, distortion in an anechoic chamber. Vibration motor latency and haptic intensity range.

SAMPLE DIMENSIONS
  • · Speaker SPL
  • · Speaker frequency response
  • · Haptic intensity range
Software7 DIMENSIONS
WEIGHT
17%

OS and security update commitment in years; AI feature depth; on-device privacy posture; bloatware; update timeliness; OS quality.

SAMPLE DIMENSIONS
  • · OS update commitment
  • · Security update commitment
  • · AI features depth
  • · Update timeliness
SEC 04

All 47 dimensions.

SORTED BY
CategoryDimensionWhat we measureInstrumentWeight
01Display
Peak brightness
Ceiling: 4,000 nits (current record). Linear from 800 nits floor.
Max sustained HDR nits, APL 10%, 5-min holdKonica-Minolta CS-2003%
02Display
Sustained brightness
Floor 60% of peak, ceiling 95% of peak.
Brightness retained after 10-min full-screen whiteKonica-Minolta CS-200 + timer2%
03Display
Color accuracy ΔE
ΔE ≤ 1 scores 10. ΔE > 4 scores 0. Inverse-linear.
Average ΔE across 24-patch ColorChecker, sRGBKonica-Minolta CS-2002%
04Display
Gamut coverage P3
100% P3 = 10. Below 90% = 0.
Volume coverage of DCI-P3 in 1976 u′v′ CIEKonica-Minolta CS-2001%
05Display
Refresh-rate stability
Subjective + frame-time variance. Calibrated quarterly.
LTPO transitions, judder events per 60s scrollHigh-speed cam @ 1000fps2%
06Display
Touch latency
≤30ms = 10, ≥80ms = 0. Linear in between.
Finger-to-pixel latency, 120Hz mode, native appHigh-speed cam @ 1000fps2%
07Display
Outdoor visibility
Ceiling: contrast > 4:1 in sun. Floor: contrast < 1.5:1.
Contrast ratio at 50,000 lux ambient (direct sun)Lux chamber + CS-2002%
08Display
HDR performance
Verified against reference master.
Dolby Vision compliance, tone-mapping fidelityCS-200 + HDR-10 test pattern1%
09Battery
Web browsing
Ceiling 24h, floor 8h. Linear normalization.
Loop at 200 nits, fixed routine, 5G off, Wi-FiEspec SH-262 + custom rig4%
10Battery
Video playback
Ceiling 26h, floor 10h.
1080p AV1 loop, 200 nits, internal speakerEspec SH-262 + custom rig3%
11Battery
Gaming endurance
Ceiling 6h, floor 1.5h.
Genshin Impact 60min, max settings, 30°CClimate chamber @ 30°C3%
12Battery
Mixed-use endurance
"Honest workday" — most-cited consumer metric.
6h scripted routine: browse/video/camera/mapsCustom orchestration rig3%
13Battery
Wired charging speed
Ceiling 25 min to 100%, floor 120 min.
Full curve 0→100%, OEM brick, 25°C ambientPowerZ KM003C @ 1Hz3%
14Battery
Wireless charging speed
Ceiling 30 min to 50%, floor 90 min.
0→50% on OEM wireless pad, 25°CPowerZ + Qi/MagSafe receivers1%
15Battery
Standby drain
Ceiling ≤ 2%/24h, floor ≥ 12%/24h.
%/24h drain in idle, all radios on, no notificationsLogged battery API1%
16Battery
Charging temperature
Ceiling ≤ 32°C, floor ≥ 45°C.
Peak surface temp during wired full charge cycleFLIR ETS320 + thermocouple2%
17Performance
Geekbench 6 multi
Ceiling 10,000, floor 4,000.
3-run median, climate chamber 25°CGeekbench 6 stock3%
18Performance
Geekbench 6 single
Ceiling 3,200, floor 1,400.
3-run median, climate chamber 25°CGeekbench 6 stock2%
19Performance
3DMark Solar Bay
Ceiling 9,500, floor 3,500.
3-run median + sustained performance loop3DMark Solar Bay stress2%
20Performance
AnTuTu 11
Ceiling 2.4M, floor 800k.
3-run median, climate chamber 25°CAnTuTu 11 stock2%
21Performance
Sustained performance
Ceiling ≥ 85% of peak, floor ≤ 45%.
% of peak retained after 30-min 3DMark stress @ 25°C3DMark stress + climate chamber3%
22Performance
App launch speed
Ceiling ≤ 600ms avg, floor ≥ 1.5s.
Cold-launch top-20 apps, frame-by-frame medianHigh-speed cam @ 240fps2%
23Performance
UI fluidity
Subjective sub-component + objective frame variance.
Jank events per minute, GFX overhead, scroll varianceGFXTrace + scripted scrolls1%
24Camera
Main daylight
Editor-blind-graded against reference set.
47-frame test, 1,200 lux + outdoor, JPEG + DNGImatest Master + lab panel4%
25Camera
Main low-light
Noise floor + exposure latency + color accuracy.
15-frame test, 10–50 lux, no flashLux chamber + Imatest3%
26Camera
Main night mode
Exposure window + noise + halo control.
8-frame test, 1–10 lux, night-mode enabledLux chamber + Imatest2%
27Camera
Ultrawide daylight
Edge MTF + geometric distortion.
12-frame test, edge sharpness, distortion correctionImatest + lab panel2%
28Camera
Ultrawide distortion
< 1.5% TV distortion scores 10.
Geometric distortion at corners, faces vs landscapeImatest distortion module1%
29Camera
Telephoto 3-5x
Resolved line pairs/mm at sensor.
10-frame test at marketed zoom, detail retentionImatest + lab panel2%
30Camera
Telephoto 10x+
Penalizes obvious hallucination.
Stability, AI-upscale honesty, color drift at long zoomTripod + lab + drift metric2%
31Camera
Front-facing
Skin tone accuracy + dynamic range.
6-frame portrait test, three lighting environmentsImatest + lab panel2%
32Camera
Video 4K stabilization
Frame-to-frame translation residual.
Walk-test, gyroscopic shake analysis, 60fpsCustom shake rig + gyro log2%
33Camera
Video color science
ΔE between lenses on same frame.
Cross-camera color match (main↔UW↔tele) in same sceneImatest + cross-lens compare1%
34Camera
Audio capture
SNR + spectral cleanliness.
On-device mic SNR, wind handling, voice clarityAPx555 + anechoic chamber2%
35Camera
Pro/ProRes/Log capture
Binary checks + bitrate ceilings.
Availability + bitrate + LUT support + frame ratesCapture inspection2%
36Audio
Speaker SPL
Ceiling ≥ 85 dB, floor ≤ 72 dB.
Max SPL at 30cm, pink noise, no clippingAudio Precision APx5552%
37Audio
Speaker frequency response
Lower max deviation scores higher.
Deviation from target curve 80Hz–18kHzAPx555 + anechoic chamber2%
38Audio
Speaker distortion
Ceiling THD < 1%, floor THD > 5%.
THD+N at 80% volume, midrangeAPx5551%
39Audio
Vibration motor latency
Ceiling ≤ 8ms, floor ≥ 40ms.
Time from API call to first detected vibrationAccelerometer @ 1kHz1%
40Audio
Haptic intensity range
8+ distinguishable steps = 10. <3 = 0.
Distinct intensity steps the motor can deliverAccelerometer + scripted API2%
41Software
OS update commitment
≥7 years = 10. ≤2 years = 0.
Manufacturer commitment in years from launchManufacturer ledger4%
42Software
Security update commitment
≥7 years = 10. ≤3 years = 0.
Manufacturer commitment in years from launchManufacturer ledger3%
43Software
AI features depth
Editorial panel + objective feature inventory.
Count + breadth of useful on-device AI featuresManual evaluation panel3%
44Software
AI on-device privacy
100% on-device = 10. 100% cloud-only = 0.
% of AI features that run without cloudNetwork capture during use2%
45Software
Bloatware
Ceiling: zero. Each removable counts half.
Pre-installed third-party apps that can’t be removedOut-of-box app inventory1%
46Software
Update timeliness
≤7 days = 10. ≥120 days = 0.
Days from OS release to device availabilityPublic release dates2%
47Software
OS quality
Editorial logbook + crash report counts.
Critical bugs / regressions / crashes during 30 daysDaily-driver journal2%
SEC 05/CHANGELOG-INTEGRITY

Versioned. Published. Audited.

SINCE 2023
VERSION CHANGES
v4.2CURRENT
12 JUN 2026

AI features re-weighted to reflect post-launch on-device shift.

  • ·AI features depth weight raised from 2% to 3%
  • ·New dimension: AI on-device privacy (2% weight, replaces "Bloatware lite" sub-score)
  • ·Update timeliness: floor relaxed from 90 to 120 days post 2026 carrier policy changes
  • ·Camera audio capture: new sub-component for wind-noise handling
v4.1
08 JAN 2026

Sustained-performance test extended from 20 to 30 minutes.

  • ·Sustained-performance stress test duration: 20min → 30min
  • ·Mixed-use endurance: scripted routine refreshed for 2026 app usage data
  • ·Charging temperature: now scored separately from charging speed
v4.0
20 AUG 2025

Major revision — three-year ceiling window introduced.

  • ·Ceiling now floats over a rolling three-year window (was static)
  • ·Camera weight raised from 22% to 25% — Battery dropped from 22% to 20%
  • ·Telephoto split into 3-5x and 10x+ as separate dimensions
  • ·Removed "Speaker max volume" — folded into Speaker SPL
v3.5
15 FEB 2025

Last v3 revision before the v4 architecture change.

  • ·Front-camera scoring: now uses 6-frame test (was 3)
  • ·Wireless charging: floor relaxed for non-MagSafe phones
  • ·OS quality: added "critical regressions" tracking
INTEGRITY COMMITMENTS
PUBLISHED METHODOLOGY
Every methodology change is dated and explained.

Versioned in public. Older versions remain accessible. When a scoring weight changes, every prior score is recomputed under the new method so cross-version comparisons stay valid. Nothing is silently re-graded.

OPEN DATASET
Raw lab measurements available on request.

The CSV behind every published score is available to other journalists and researchers under attribution. Camera test panels and audio waveforms are shareable. We've never been asked to retract a measurement.

THIRD-PARTY AUDIT
Annual audit by an independent test lab.

Once a year, an independent test lab runs our methodology against a sample of devices we've already scored. Discrepancies above 0.3 points are investigated and the audit summary is published in November.

CONFLICT DISCLOSURE
Affiliate revenue does not influence scores.

Affiliate links exist on every recommendation but no editor has ever seen our affiliate revenue per device. Editors are paid a fixed salary; their compensation cannot be linked to which device gets which score.