Sigma 12-24 F4 Art on the H

This is probably the last test and review of a wide angle solution for the H. The new Sigma 12-24mm F4 DG HSM | Art is a completely redesigned lens which claimed to have the largest piece of glass-molded aspherical element ever used on a lens, with the promise of near distortion free and excellent optical performance throughout the frame. This is a full frame lens, and when used on both the sd Quattro and Quattro H, it will produce an angle of view similar to 18-36mm (on APS-C) and approximately 16-31mm (on APS-H) respectively. The aperture remains constant at F4.0 regardless of the sensor size it is being used on.

The dp0 Quattro has been my reference ultra-wide angle solution for more than two years, and it remains one of the best solution without much compromise at a reasonable cost. It is perfectly matched to the Foveon Quattro sensor (1.5x APS-C) in a small form factor. It is the only dp Quattro with the largest aperture of F4.0 which is not a deal breaker given the fact that it uses ultra wide angle which already has the advantage of deeper depth of field at any given aperture compare to a full frame camera system. If you have read my other reviews, there isn’t an ideal ultra wide solution for the sd Quattro H….not until the arrival of this lens.

I did some comparison tests based on real world images, between the dp0 Quattro, sd Q (simulated by turning on DC Crop Mode) and sd QH (the “H”). The results are rather interesting, not something I would expect. Let’s have a look at the first test image:

Test #01 (dp0 Quattro vs sd Quattro + 12-24 Art @ 14mm, both at F5.6):

dp0Q, 14mm, F5.6, ISO100, Daylight WB
sdQ + 12-24 @ 14mm, F5.6, ISO100, Daylight WB

The following can be observed from this test:

  • There is a slight warm cast on the dp0Q image although both are set to Daylight WB. It can also be attributed to the color rendition of the 12-24 Art lens which renders it slightly cooler. Which is more accurate? Neither, because both cameras aren’t profiled, which is not important in this comparison, nonetheless interesting to note.
  • Center resolution is neck to neck, there’s very little to separate the two.
  • The corners are softer on the dp0Q.
  • CA is more pronounced on edges of high contrast on the dp0Q
Pronounced chromatic aberration on the dp0Q. Much less on the sd Q + 12-24 Art lens.
Optical performance at right bottom corner. dp0Q (left), sd Q (center), H (right)
Top right corner. dp0Q (left), sd Q (center), H (right). Again, the CA is very pronounced on the dp0Q but is almost completely absent from the 12-24 Art.

Test 02, Resolution (dp0Q vs H):

In this test, I had to do match the exposure of both images and down size H image to match the dp0Q resolution (5424 x 3616).

dp0Q, Test 02
sd Quattro H + 12-24 Art, downsized to match dp0Q resolution.

The resolving power at the center of the frame is excellent on both cameras although the 30% resolution advantage on the H is pretty obvious.

Left : dp0Q, Right : H + 12-24 Art. The leaves on the trees are better defined on the H image.

The optical performance at the corners are way better on the H + 12-24 Art:

Test 02: Bottom left corner of the frame.
Test 02: Bottom right corner of the frame.

It is obvious that the Sigma 12-24 F4 Art lens performs very well on the H. The optical performance is very consistent across the frame, with negligible issue of CA compare to the already excellent dp0Q. Both exhibits near zero distortion in any real world images. The one major advantage the dp0Q has over the elder sibling is the ability to use screw-in filters and at 58mm, the cost of filters are way cheaper and can be easily purchased. The 12-24 Art requires a special filter adapter, and as far as I know only NiSi has a ready solution at the moment, which holds 150mm panel filters which are costly.

For those who are using the sd Q body (APS-C), the better choice is the 8-16mm which is nearly less than half the price of the 12-24 Art without making any major compromises in terms of optical quality and performance although usage of filters is virtually impossible.

To summarize my findings and test results on ultra-wide angle solutions, kindly refer to the chart below:

*The tests on the H should have been stopped down by a further 0.5-stop in Aperture value to match the depth of field of the APS-C sensor. My bad….