When you are choosing cameras, the three resolutions you will actually see are 1080p (2MP), 5MP, and 4K (8MP). Each tier serves a different job, and the right answer is almost never "the highest one everywhere." This guide compares all three so you can assign the correct resolution to each camera location.
The quick answer
5MP is the balanced default for most coverage; step down to 1080p indoors and up to 4K only where long-range identification matters.
Detail and identification distance
Under DORI standards with a typical 4mm lens:
- 1080p: identify a face at ~15-25 ft.
- 5MP: identify a face at ~25-45 ft.
- 4K: identify a face at ~50-80 ft.
So 4K is not "a bit sharper" — it roughly doubles 5MP's useful range. That is why 4K earns its keep at entrances, gates, and parking approaches, while 1080p is wasted on a wide driveway.
Low-light performance
Night quality follows sensor size and pixel size, not megapixels. On a small sensor, 1080p and 5MP often beat 4K after dark. Practical rules:
- Dim indoor area → a Starlight 1080p or Starlight 5MP camera.
- Lit outdoor area → 5MP is the sweet spot.
- Large, bright outdoor area needing long-range ID → 4K with at least a 1/1.8" sensor (look for Starlight/ColorVu).
Storage and bandwidth
Relative daily storage (H.265, continuous):
-
1080p: ~10-15 GB/day
-
5MP: ~16-25 GB/day
-
4K: ~30-50 GB/day
4K is about 2.5x the data of 1080p. Use H.265+ compression and motion-only recording (cuts 60-70%) to keep NVR size sane, especially in all-4K deployments.
Mixing tiers by zone (recommended)
A modern NVR mixes resolutions without issue. A sensible plan:
- Entry, gate, perimeter: 4K for identification-grade footage.
- General coverage (rooms, floors, yards): 5MP — the best detail-to-cost ratio.
- Constrained indoor spaces (hallways, small rooms): 1080p to save storage.
Related guides
- Full breakdown of every tier: IP Camera Resolution Guide.
- Pairwise comparisons: 2MP vs 4MP and 4MP vs 1080p.
- Top-tier decision: 4K vs 5MP.
- Planning bandwidth and storage? See the PoE NVR setup guide.