Sigma SD 15 vs EOS 1000d: una sfida tra sensori Bayer e Foveon “oldgen“ , a Crespi d’Adda Sigma SD 15 vs EOS 1000d: una sfida tra sensori Bayer e Foveon “oldgen“ , a Crespi d’Adda Sigma SD 15 vs EOS 1000d: una sfida tra sensori Bayer e Foveon “oldgen“ , a Crespi d’Adda Sigma SD 15 vs EOS 1000d: una sfida tra sensori Bayer e Foveon “oldgen“ , a Crespi d’Adda
Medium article

Sigma SD 15 vs EOS 1000d: a challenge between “oldgen” sensor Bayer and Foveon, in Crespi d’Adda.

A few weeks ago I bought an used Sigma SD15, in order to touch the “legendary sharpness” of the Foveon sensor. The benchmark reference could only be my current EOS 1000d, a good SLR that by age and megapixel lends itself particularly to the comparison. The place of challenge: Crespi d’Adda.

First the theory: Bayer sensor and Foveon sensor, two comparing philosophies.

The Bayer sensor
Digital photography as we know it today, from the general public is made by the sensor known as Bayer, used in almost all the ‘digital’ for 20 years now. The Bayer sensor is named after Bryce Bayer, a researcher at Kodak, the company produced it first.

The CCD or CMOS sensor Bayer is composed of a grid of pixels, organized in groups of 4 pixels each sensitive to a particular primary RGB color: more precisely 25% of the pixels are blue, 25% are red, 50% is green , this proportion because the human eye is more sensitive to the green. In practice, the recorded image is made up of a sort of 4-point mosaic: to define 1 pixel-output we need 4 diode pixels: we do not notice it because “demosaicing” is done in camera along with software anti-aliasing processing. The end result is a normal bitmap file made up of colored pixels that we all know and display on a computer.

The Foveon sensor
As few people know, some cameras in circulation are equipped with a conceptually different sensor: the Foveon sensor.
This type of sensor is very similar to a television screen, observed in close proximity: each pixel-output is also a pixel-diode, ie every single pixel can record any of the RGB colors from 0 to 255.
It is not necessary to reorganize the pixels in groups of 4, or to apply anti-aliasing filters.

The name Foveon derives from “Fovea Centralis” an important anatomical part of the eye. Foveon inc is the name of the only company that produces this type of sensors, the result of research by Caltech, physicist and computer engineer, and Professor Carver Mead.

The Foveon Sensor is structurally very different from the Bayer and closely simulates the operation of color films that we all used before the advent of digital photography: three different layers of vertically stacked pixel-diode record color information. The result is that every pixel-output present in the sensor records 100% of the color.

The two SLRs

EOS 1000d, the baseline.
The Canon EOS 1000D has been my trusted reflex since 2009. Launched in June 2008 is the first entry-level of Canon ‘
As often happens for the first model of a product line, Canon was generous, considering the price: 10 MP CMOS sensor, continuous shooting speed of 3 fps, 7-point AF system, ISO sensitivity from 100 to 1600 and speed of shot from 30 seconds to 1/4000 s.
In the test I match the Canon 28–80 USM IV lens, a lens actually full frame, very fast in focusing and quite clear.
As a flaw, it has a tendency to be washed out, and a tendency towards horizontal ABS.

The 1000d has just enough to do serious photography.

This reflex won the 2008–2009 EISA Award for Best product: entry level camera. Do not underestimate it, shooting in RAW can produce excellent photos.

Sigma SD15
The Sigma SD 15, shown in preview in 2008, then released since June 2010, is the fourth digital SLR produced by Sigma.
At its core is the second generation of Foveon X3 sensor, a 4.7MP sensor announced by Sigma as a 15MP.
(Edit)The lens used is the Sigma 17–50 DC OS HSM, a so good zoom with nice colours and good sharpen over 20mm. It matches the 28–80.
The SLR has a professional body of medium size and good ergonomics, 5-point AF autofocus arranged in a cross, ISO sensitivity from 50 to 3200, shutter speed from 2 minutes to 1/4000 s.

What is striking about the SD15 are both ergonomics and the silence of the shutter. The lack of vibration at the time of shooting suggests that it has the curtains lined…

A comparison between different sensors, difficult to compare.
Comparing two normal ‘Bayer’ cameras is simple: just look at the megapixel. Taking a test between Bayer and Foveon requires that SLRs are of the same age and with similar effective resolutions.
On the net, various forums and photography sites try to compare the resolving power of the two technologies:

Sigma declares a yield of 300%, 4.7MP Foveon = 15MP Bayer

Many speak of a yield of around 200%, while a detailed post ( talks about a yield of around 12MP.

My thesis is that the yield something over 200%, in practice,
4.7MP Foveon = 10MP Bayer. This is due to technical limitations of the second generation Foveon X3

Furthermore, the 2 sensors are both APS-C, of very reduced size:
– 20.7 x 13.8 mm the Foveon X3 (x1.7 conversion factor from 35mm)
– 22.2 x 14.8 mm for Canon CMOS (x1.6 35mm conversion factor)
This helps above all using the optics.

The Location: the Village of Crespi d’Adda, an Unesco site

(Photo Canon, RAW not edited)

To be able to make a photographic comparison I needed a meaningful location that would condense different types of situations: panoramic points, long perspectives with houses, architectures with fine details.

A place with tranquility to get around with 2 SLRs under the arm without attracting… The Village of Crespi d’Adda is ideal, and is also a UNESCO site, so the choice.
To find out more about the village visit the related wikipedia entry:

The shots in comparison

In taking pictures I followed the tourist route starting from the panoramic point. Compatibly with the lens differences, I have tried to shoot with the same ISO, aperture and exposure times.
The files below are the RAW converted to jpg from Camera Raw, without applying any adjustments (not even the automatic sharpness, left at 0).

Below, the EOS 1000d shots on the left, and SD15 on the right

PHOTO 1 — View of the village from the hill
PHOTO 2 — Panel with diorama and view around
PHOTO 3 — Flowers along the stairway
PHOTO 4 — Architecture. The Church of the Worker Village
PHOTO 5 — The entrance to the plant and the chimney
PHOTO 6 — Secondary facades in perspective
PHOTO 7 — Zoom in on the dome of the church

Magnification with the lent…

In terms of sharpness, in the full size photos you notice the greater clarity in the details of the images of the SD15. Going to zoom with the practical and merciless luma of Camera Raw, the differences becames clearer.
The X3F files (right) were not interpolated to get to 10MP output resolution, but left at 4.7MP.

PHOTO 1 — In both the details of the kiosk and in the shingles and antennas, precision is higher in SD15
PHOTO 2 — Detail of the metal panel, in backlight. Slightly sharper, but in the shadows the SD15 seems to reduce the advantage.
PHOTOS 3 — The details of the pistils in the flowers seem burned to the left, well engraved on the right
PHOTO 4 — Detail the lamp post
PHOTO 5 — In this case the EOS 1000d is better, in the shaded areas the extra MPs help to draw (partially) the striped textures. But in the illuminated areas things are reversed.
PHOTO 6 — Noticeably better detail of the leaves, in addition to rather marked halos in 1000d. The leaves on the right, although backlit, remain clear.
PHOTO 7 — In shaded areas similar levels of detail, but the backlit crosses are the perfect example of how in Foveon defines the details even with a single pixel.

The verdict

The SD15, looking at the sharpness surprised me for how it managed to capture details with a grid of pixels output from only 4.7MP.
The reference 1000d, however, has behaved well as a definition in the gray areas, in those cases exceeding the SD15 (photo 5 details of the windows).

In other words, I believe that in Foveon’s sensor it is a definition that is greatly enhanced by a certain overexposure. The exemplary case is the fine details of the antennas and crosses silhouetted against the clear sky: the SD15 can describe them with neat black pixel lines, while the Bayer sensor of the EOS 1000d, despite its megapixels, ends up blurring with halftone pixelsand some cyan / red chromatic aberration. It’s the mosaic!

The sigma SD15 and the second generation Foveon x3 sensor win the sharpness challenge!


Looking globally at the incorrect RAW photos, as if they were JPG shots, the 1000d proved to be more reliable, especially in situations with contrasted exposure. No wrong photos.
The photographs, although with those typical slightly faded tones, are more pleasant and luminous, with the right balance of white.

The photos of the Sigma SD15 al naturale look slightly greenish / yellowish, and a bit ‘underexposed’.
Moreover, perhaps due to the knowledge of the in-depth vehicle, some photos of this have presented strange color spots in the lights!

Here the film burned … old memories! Is the Foveon so similar to the film?


Well, after an afternoon of shots and a couple (!) Of video projection to the Mac ends this challenge with shots of tourist photos, to… Crespi d’Adda.

The proof of the 2 SLRs is of relative interest, in a 2018 where expensive cameras of 100.000 ISO and more are wandering, but it will be useful — sooner or later — to some passionate photographer who browsing looking for comparisons between the SIGMA SD15 with its Sensor Foveon (gen2), and a Canon from 2008–2010.

P.S. I reserve to file and add the final photos to the story AFTER the post-production, which continuing the challenge will be the one on Camera RAW 10.x, the other on Sigma Photo pro 6.x!

Share this

Giovanni Minelli

Art Director & Cross media designer, content curator at minodesign. ADCI member.
He deals with advertising projects, graphic design and web design, with a focus on corporate image, UI and image manipulation. Experimenter of new media but also lover of the 70s graphics and 80s90s adv; seek a contemporary synthesis of these three. It is also dilettans mangaka and ebay seller.