Video summary

Galaxy Watch Ultra - In-depth Scientific Review! (BEST Samsung Smartwatch Yet)

Main summary

Key takeaways

Product Review

Product reviewed

Samsung Galaxy Watch Ultra 2 (often described as effectively similar sensor-wise to the original Ultra; “Ultra” used throughout to refer to the latest model).

Key features tested

  • Heart rate tracking (PPG on wrist) across:
    • Indoor cycling (“spinning”)
    • Running (including concern for cadence lock)
    • Biking (harder conditions; missed/false peaks noted)
    • Weightlifting (highest difficulty: tension causes false/missed peaks)
  • Sleep stage tracking (deep/light/REM; compared against an ECG headband “silver standard”)
  • GPS tracking performance (location overlap/route deviation on cycling and running; compared qualitatively to sports watches)

Heart rate tracking performance (main findings)

Indoor cycling / spinning

  • Correlation: about 0.94 (target mentioned: > 0.95).
  • Overall: “pretty good” with occasional over/underestimation (sometimes too high, sometimes too low).
  • Comparison:
    • Polar Pacer Pro / Coros Pace 3 / Galaxy Watch 8 are close (Galaxy Watch 8 nearly matches Ultra 2; same sensor likely).
    • Apple Watch + some Huawei + Pixel Watch are better (correlation ≥ 0.98 cohort).

Running

  • Correlation: about 0.97.
  • Overall: reasonably happy, mostly close to reference line.
  • Issue suspected: minor cadence lock (device may misread running frequency as heart rate).
  • Individual runs:
    • One run “doesn’t look quite as good” with noise and possible too-high HR (possible cadence lock).
    • Interval run: still “good enough” though not perfect.
  • Comparison:
    • Pixel Watch (Pixel Watch 3/4) generally more reliable on the reviewer’s body.
    • Apple Watch also “always does well” on the reviewer.
    • Note: Pixel Watch had one “weird result” in earlier testing, treated as a possible fluke.

Biking (harder exercise)

  • Correlation mentioned as ~0.90 (borderline).
  • Reviewer observed:
    • Missed peaks and added heart rate at some times.
    • One notably poor segment (entering apartment/indoor transition), but overall still “good enough” for the reviewer.
  • Comparison:
    • Galaxy Watch Ultra 2 is similar to Galaxy Watch 8 (expected due to same sensor).
    • Pixel Watch 3/4 perform better (when zoomed into ≥ 0.95 correlation cohort).

Weightlifting (hardest for wrist HR)

  • Correlation: about 0.86.
  • Pattern:
    • In higher HR ranges, Ultra 2 often reads too low.
    • Many peaks not detected; sometimes false positives.
  • Guidance/recommendation:
    • If you lift seriously, chest straps (ECG) are recommended.
    • Reviewer recommends Polar H10 ECG and says Wahoo chest strap produced similar results to the Polar H10 in their use.
    • Key point: wrist-based PPG struggles especially during upper-body tension; lower-body tends to be better.

Practical takeaway on HR accuracy

  • On most exercises, “good enough” threshold for many users.
  • Reviewer frames Ultra 2 as “second tier” for heart rate vs Apple/Pixel and especially vs chest-strap ECG.

Pros / Cons (heart rate)

Pros

  • Strong performance for spinning (high correlation ~0.94, often close)
  • Running is fairly strong (~0.97 correlation), mostly reliable
  • Galaxy Watch ecosystem value sometimes offsets accuracy concerns (plus pricing/bundles noted)

Cons

  • Biking drops to borderline (~0.90) with missed/added peaks
  • Weightlifting is noticeably weak (~0.86) due to tension artifacts
  • Cadence lock suspected during running (and possibly other activities)

Sleep stage tracking (key findings)

Method

  • Compared against an ECG headband measuring brain-wave-based stages.
  • Focused on: deep sleep, light sleep, REM/RAM sleep
  • Awake detection considered overestimated by the reference setup.

Results vs ECG reference

  • Deep sleep agreement: ~76%
  • Light sleep agreement: ~57% (worst stage; confusion occurs)
  • REM/RAM sleep agreement: ~77%
  • Reviewer notes:
    • Many devices struggle with REM/RAM, and Ultra 2 does pretty okay there.
    • Samsung appears to have improved sleep stage tracking vs older Ultra (2024) based on their trend.

Comparison

  • Close to Galaxy Watch 8 (reinforces “same sensor” idea)
  • Better than Ultra (2024) in their earlier testing
  • Still not as good as top sleep devices:
    • Eight Sleep is called one of the best sleep stage trackers (also an expensive sleep improvement system)
    • WHOOP and Aura described as strong competitors
    • Apple Watch: struggles mainly with deep sleep (detects too little deep sleep), but total sleep score can be similar; Aura/OURing more consistent by stage.

Sleep verdict

  • Galaxy Watch Ultra 2 is described as “second tier” for sleep stage tracking:
    • Better than sports watches
    • Competitive with other wearables but below the best-in-class sleep trackers/improvers.
  • Reviewer speculates it may be moving toward better sleep tracking over time via updates (referencing study-style comparisons and improvements seen on Galaxy Watch 8).

GPS tracking performance (cycling + running)

Cycling (to/from work)

  • Reviewer uses route overlap and deviation as the measure.
  • Findings:
    • GPS acquires quickly (good cold-start behavior)
    • But there is frequent deviation and mixed consistency
    • Sometimes the route looks “split” / suggests potential calibration or satellite consistency issues
  • Comparison:
    • Reviewer says sports watches do better:
      • Mentions Garmin Forerunner 965 and Coros/Suunto devices (Race 2 mentioned)
    • Ultra 2 is described as not terrible, but still struggling compared to sports watches.

Running

  • Mixed results based on interference:
    • Runs near a wall show more deviation (interference clearly impacts performance)
    • One run (winter) looks cleaner overall
    • Reviewer notes Apple Watch Ultra 3 seemed to do better in the interference scenario

GPS verdict

  • Good enough for many people, but not perfect.
  • Clear conclusion: for high GPS demands (long runs, mountains, accuracy-heavy tracking), a dedicated sports watch is recommended.

Overall recommendation / verdict

  • Galaxy Watch Ultra 2 is not a “best in class” sensor watch, but it can be good enough depending on what you need:
    • Best fit (per reviewer): if you want the Samsung ecosystem and have a Samsung phone, and you’re okay with second-tier health metrics (especially vs Pixel/Apple and vs chest-strap ECG).
    • If you lift seriously: wrist HR is a weak point—get an ECG chest strap (external, not watch-compatible per reviewer’s assumption).
    • If you need the best GPS: choose Garmin/Coros/Suunto/Polar sports watches instead.
    • If you care most about sleep stages: top sleep-focused devices (Eight Sleep / Whoop / Aura) outperform; Ultra 2 is “second tier.”

Bottom line: Recommended as a versatile smartwatch in Samsung’s ecosystem, but not the top choice for maximum HR accuracy during strength training or best GPS versus dedicated sports watches, and not the top choice for sleep stage precision versus specialized devices.

Unique points explicitly mentioned (consolidated)

  1. Ultra 2 vs original Ultra: main differences largely memory/software, sensor assumed similar.
  2. Heart rate correlation targets: prefer R > 0.95 (spinning ~0.94).
  3. Indoor cycling: occasional HR over/under.
  4. Running: cadence lock suspected; still strong overall (~0.97).
  5. Biking: harder performance (~0.90), missed/false peaks.
  6. Weightlifting: weak (~0.86), tension causes failures/false positives.
  7. Best HR method: ECG chest strap (Polar H10 cited; Wahoo similar).
  8. Sleep testing method: ECG headband (“silver standard”), focuses on deep/light/REM stages.
  9. Sleep stage agreements: Deep ~76%, Light ~57%, REM/RAM ~77%.
  10. Sleep is “improving” vs 2024 Ultra; close to Galaxy Watch 8.
  11. GPS: route overlap shows mixed deviation, sometimes interference-heavy (walls).
  12. GPS acquisition can be quick but tracking consistency varies.
  13. Sports watches (Garmin/Coros/Suunto/Polar) provide better GPS in reviewer’s testing.
  14. Recommendations depend heavily on ecosystem and priorities (Samsung vs iOS vs Android vs subscription straps vs sleep-only devices).
  15. Battery life tradeoff mentioned: sports watches generally better battery than smartwatches/Pixel watches.

Speakers / contributors

  • Single main speaker: Rob (scientist specializing in biological data analysis), conducting the systematic data-based testing and comparisons.

Original video