HD Lens Shootout

BE
Benros Emata
Published Oct 02, 2011 04:00 AM

Conventional wisdom says that lenses make a big difference in image quality but the question remains: "How much of a difference?" With the shift to megapixel cameras increasing demands for higher quality lenses, this question has gained even greater importance.

We wanted to better understand quality differences by doing our own series of shootouts. This is our first report in that series.

Lens Shootout Approach

The most common lens type used with megapixel cameras are varifocal lenses with a relatively wide field of view (e.g., 3 - 8 mm). This is the normal stock or default lens provided with most megapixel cameras today.

We wanted to start with this type, selecting 5 of the more common lens manufacturers. This test include lenses from:

  • Computar
  • Evetar
  • Fujinon
  • Tamron
  • Tuss

We then used these lenses with three commonly used IP cameras:

  • Axis P1344
  • Panasonic WV-SP306
  • Sony CH140

We conducted tests in 3 different scenes:

  • Indoor office for a medium size area
  • Enclosed parking structure for a larger area
  • Indoor darkroom for night time testing

Since lens focusing is a key issue in optimizing cameras, we used the auto back focus features available in all 3 cameras to provide a standardized way to focus all lenses.

Preview

Below is a preview of one of the scenes that shows variations in performance achieved.

Inside the Pro section, we dig into the key differences and how the lens performance compared.

Future Tests

Here are the key findings:

  • Among MP lenses, differences in quality/detail levels more apparent in wider FoVs
  • Loss of detail at fringe more evident at wider FoVs (e.g. ~35ft HFoV)
  • The SD lens on a MP camera produced somewhat worse image quality, more pronounced at wider FoVs
  • The practical differences between worst and best MP lens tended to be modest
  • Fujinon and Tamron lenses tended to outperform other MP lenses throughout test scenarios
  • However, Fujinon and Tamron were not the best in every test but was at least average in all and the best in many
  • Fujinon lens exhibited the least image quality degradation at fringe
  • Evetar performed the poorest especially at night

In light of these findings the following is recommended:

  • Avoid SD lenses with MP cameras. This should be obvious but the tests results do bear out diminished performance (as to be expected).
  • The Fujinon might be preferable if you want to maximize performance. However, differences likely will not be huge.
  • The only lens that really concerns us is the Evetar one. While it is by far the lowest cost, the negative tradeoff in performance is clear. Evetar is increasingly popular with camera manufacturers and the biggest challenge is that you may not know that you are getting such a lens because the manufacturer may remove the original manufacturer's label.
  • Do not presume that all Fujinon and Tamron lenses are better. While both manufacturers do have good overall reputations, we have only tested one lens out of the dozens that each company makes (model numbers of lens tested listed at end of report).

Picture of Indoor Scene

Below is the indoor scene we used. The yellow/orange boxes indicate the areas zoomed into and analyzed. We also had the subjects move to the right side for a fringe test. We had both a near and far subject. For the near subject, there were no material differences in image quality so we skipped an analysis of that. 

Indoor Daytime Moderately Wide FoV (~35ft)

Axis P1344 (Center/Fringe)

In this scenario, all six test lenses are setup with equal lens angles. A subject is positioned at a distance corresponding to a ~35ft HFoV  near the center of the FoV and then moves laterally towards the fringe of the FoV. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

 

Images (both center and fringe) from the SD lens are clearly inferior, less detailed or 'fuzzier' than any of the MP lenses. The SD lens also shows a high degree of detail loss from center to fringe.

Center images from MP lenses exhibit subtle/slight but detectable differences in detail levels. Both Fujinon and Tamron appear to produce the best detail levels among the MP lenses in both the center and fringe images. The Computar MP lens appears to produce the most center vs. fringe detail loss or quality degradation.

Sony CH140 (Center/Fringe)

In this scenario, all six test lenses are setup with equal lens angles. A subject is positioned at a distance corresponding to a ~45ft HFoV  near the center of the FoV and then moves laterally towards the fringe of the FoV. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

Overall the SD lens images (both center and fringe) provide less detail than MP lenses. However, the SD performs roughly similar to the Evetar (worst performing MP lens).

The Fujinon lens looks best compared to the field in both center and fringe images. The Computar MP lens provides relatively strong details at the fringe. Oddly it appears to perform slightly better at the fringe than center in this particular test case.

Parking Garage Artificially Lit Wide FoV (~50ft/~75ft)

Below is a sample picture of the parking garage scene. The subject first went to a mid-way point and then to the far wall as depicted by the orange/yellow boxes displayed.

Axis P1344 (~50ft FoV/~75ft FoV)

In this scenario, all six test lenses are setup with equal lens angles. A subject is positioned at a distance corresponding to a ~50ft HFoV near the center of the FoV and then moves back corresponding to a ~75ft HFoV - still in center. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

Both ~50ft and ~75ft HFoV SD lens images indicate some clear deficiencies of SD in comparison to MP lenses. The wider field of views used in this comparison tend to make separation between lenses more detectable or apparent without having to digitally zoom any further. The Tuss appears particularly less 'sharp' or clear as the other MP lenses. Meanwhile, Tamron looks the most detailed in the ~50ft HFoV image comparison, with Fujinon just slightly behind. The Fujinon appears best in the ~75ft HFoV comparison.

Indoor Low-Light (~1 Lux) Narrow FoV

For our low light tests, we used an enclosed area and an eye chart to discern differences:

Axis P1344 (~5ft HFoV)

In this scenario, all six test lenses are setup with equal lens angles. An eye-chart is positioned at a distance corresponding to a ~5ft HFoV. Each camera/lens combination is auto focused in b/w mode with ~300 lux of ambient light. Lighting levels were subsequently reduced to ~1 lux for testing. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

In this narrower HFoV (~5ft), differences between the lenses are not as distinct. For example, even the SD lens provides detail levels nearly on par with the MP lenses. Overall, differences among MP lenses are quite minimal in this narrow HFoV. The Evetar lens is notably 'darker' than the field and as a result provides less image contrast.

Panasonic WV-SP306 (~10ft HFoV)

In this scenario, all seven test lenses are setup with equal lens angles (this test includes an additional stock lens). An eye-chart is positioned at a distance corresponding to a ~10ft HFoV. Each camera/lens combination is auto focused in b/w mode with ~300 lux of ambient light. Lighting levels were subsequently reduced to ~1 lux for testing. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

Note that differences are considerably more detectable in this wider HFoV (~10ft), compared to the previous ~5ft HFoV. Both the SD lens and Evetar lens provide the least in terms of image quality. However, they present with slightly different problems. The SD lens image appears blurry with less contrast, whereas the Evetar exhibits fairly strong contrast but characters on the eye-chart appear 'blocky' or tend to 'bleed'excessively.

The Fujinon lens provides greatest details by quite a margin over the other MP lenses. Tamron and the stock Panasonic lens are above average, as well.

Sony CH140 (~10ft HFoV)

In this scenario, all seven test lenses are setup with equal lens angles (this test includes an additional stock lens). An eye-chart is positioned at a distance corresponding to a ~10ft HFoV. Each camera/lens combination is auto focused in b/w mode with ~300 lux of ambient light. Lighting levels were subsequently reduced to ~1 lux for testing. The composite below shows each of the camera/lens combinations digitally zoomed to better reveal differences between them. Note that a green box indicates our assessment of the best image quality, a yellow box indicates slightly better than average, and a red box indicates worst or very poor image quality.

Again the SD lens image is clearly less detailed than any of the MP lens images. The Computar MP lens exhibits the least detailed image among the MP lenses. The Tamron appears to produce the most detailed image, edging out both the Fujinon stock and test lenses.

Methodology

Here are the lenses tested:

  • Computar (SD) - F1.0; 3.5 to 10.5mm; 1/3"; CS
  • Computar TG3Z0312FCS-MPIR (MP/IR) - F1.2; 3 to 8mm; 1/3"; CS
  • Evetar M13VD2812IR (MP/IR) - F1.4; 2.8 to 12mm; 1/3"; CS
  • Fujinon YV2.1x2.8SR4A-SA2 (MP/IR) - F1.3; 2.8 to 6mm; 1/3"; CS
  • Tamron M13VG288IR 3(MP/IR) - F1.2; 2.8 to 8mm; 1/3"; CS
  • Tuss VT3Z3280MDIR - (MP/IR) - F1.4; 3.3 to 8mm; 1/3"; CS

Here are the cameras used:

  • Axis P1344 (online $759) - 720p D/N; 1/4" CMOS; 0.05 Lux (BW)
  • Panasonic WV-SP306 (online $550) - 1.3MP D/N; 1/3" MOS;  0.3/0.05 lux (Color/BW)
  • Sony CH140 (online $800) - 720p D/N; 1/3" CMOS; 0.1 Lux (BW)