Sigma are pleased to inform you that Sigma dp0 Quattro’s special site: Inside stories of development is up on Sigma SGV website.
From the interview of engineers, you can get an insight on how Sigma dp0 Quattro was developed.
It was March 2014 when the development of the dp0 Quattro started “once again.” As a matter of fact, there was a project to develop the dp featuring an ultra-wide angle 14mm lens (equivalent to 21mm in 35mm format), and it was before the dp2 Quattro, the first model of the Quattro generation, made its first appearance. It was SIGMA’s response to the customers’ request for “the dp with an ultra-wide angle lens,” but the project disbanded before long. The time wasn’t ripe yet.
When Yamaki, the President of the company, casually expressed his interest in the old project that seemed to have been once abandoned, it received the attention in the company again. While the old project member welcomed it, they were also worried. They wanted to respond to the customers’ request. Meanwhile, developing a product involves enormous cost. Even though the focal length is traditional and common as an ultra-wide angle lens, this one is not an interchangeable lens but a lens permanently fixed to the camera, meaning only one focal length can be used. Will users accept it? It was natural for them to be careful in commoditization and the project became controversial. Will we make it or not? After many days of discussion, Yamaki said the following. “As you say, it may not sell. But, let’s make it. SIGMA is a manufacturer who tries. If we stop trying, we’re not SIGMA anymore.” Then, one member made a decision. It was Tomoki Kohno, a lens designer / the Unit Acting Manager of the Product Development Unit 2. “If it won’t sell, let’s do our very best and make a historic 21mm.” It was a “So What attitude” rather than a decision.
Code name “0”
One thing Kohno had decided before he started designing the lens was “to make the distortion rate lower than 0.5%.” Distortion is a phenomenon where straight lines of subjects (such as architectures and horizontal lines) are rendered as curbed lines and this tendency becomes more common and difficult to correct with wider lenses. Yet, if the distortion rate is lower than 0.5%, one can say that the lens captures straight things straight with confidence. And, it became Kohno’s primary goal of the lens design.
At the time, the Product Planning Unit was examining the specification of the new dp Quattro with the ultra-wide angle lens. Because there was no major change in the body, they focused on the lens. The focal length has already been decided. Then, what about the maximum aperture? In order to set the sales price within a certain range, they couldn’t ignore the production cost. The maximum aperture doesn’t always explain the quality of a lens, and it’s difficult to produce an ultra-wide angle lens with a lower F-number. So, they concluded that the maximum F-number is between 3.5 and 4.5. And, this camera got its code name “0.”
Product Development Unit 1
3.5 or 4
Manabu Tominaga, the Chief Clerk of the Product Development Unit 1 (responsible for entire camera design including lenses), was appointed to be the project leader of the “0” at the young age of 33. Tominaga’s job is to draw and finalize a mental picture of the camera and make it real. Upon receiving the decision made by the Product Planning Unit, Tominaga went to Kohno and said, “Let’s make it 3.5.” It was natural for Tominaga to hope for a faster lens even a little. However, Kohno was thinking of F4. 3.5 and 4. The difference between them was just 0.5. But, Kohno, who had been involved in lens design for many years, knew it was a big difference. Of course, they could make a F3.5 lens and it would appeal to the market better. But, the 0.5 will sacrifice something, such as the goal: “the distortion rate lower than 0.5%.” Kohno wanted to perfect fundamental performance of the lens more than to make it faster. In this regard, Kohno often used the expression “nature.” What Kohno wanted to make was a lens with a good nature.
However, Tominaga didn’t withdraw and insisted on F3.5. He had someone bring the plan that was still at the sketching stage, and while pointing at the extra space in the lens-barrel, he said that there was some space to take a larger lens within the lens-barrel. In Kohno’s words, Tominaga “was ever so persistent.” Tominaga had his hope that he could never negotiate as a project leader. If the spec order is “between F3.5 and F4.5,” he wanted to overcome the highest hurdle. The two sides didn’t reach an agreement on this issue for a long time, but Tominaga finally understood Kohno’s decision and the battle over 0.5 was settled. It’s decided. The lens is F4.
To realize the ultimate 21mm
There were three points to be cleared to design the lens.
1）Unusual corner image quality for an ultra-wide angle lens
In order for digital cameras in general to obtain beautiful images, it’s ideal for the sensor to receive the light as straight as possible, or without angling it. And, this condition has a big impact on the color reproduction around the corners and image formation. The Foveon sensor was particularly sensitive to this issue in terms of structure. There are many lenses out there, and they follow some basic patterns of lens construction. While the majority of the basic patterns are classic, many ultra-wide angle lenses are designed to angle the light before it hits the image sensing plane. There are two reasons for this. One reason is that it was easier to design. The other reason is that films were more tolerant of angled light. In other words, when films were used predominantly, angled light was taken for granted and there was nothing wrong with it. But now it’s a different story.
A lens by nature cannot straighten the passing light completely. Then, to what extent can the light be angled? What the design team did first was to study the admissive value of the angle of light hitting the sensor. Every day, they kept changing the angle of light by 0.1 degree and examined the data. As it turned out, the lens-barrel became the longest possible to maintain the balance with the body. In principle, it was long enough for the original light entered from the front element to be bent little by little while traveling through many lenses. And, by the time the light goes out from the rear element, it becomes reasonably straight. Thus, they could get the result they wanted.
The long lens-barrel also worked to minimize distortion. The extra room of the lens-barrel allowed not only more freedom in the lens design but also trying many ideas they wanted to incorporate. Besides, the advanced lens materials and processing technology gave a boost to the better design. Finally, the “0” adopted the lens construction of “11 elements in 8 groups” including SIGMA’s exclusive four FLD (“F” Low Dispersion) glass elements, two SLD (Special Low Dispersion) glass elements, one wide double-sided aspheric lens, and one single-sided aspheric lens. According to Kohno, “it was absolutely impossible 10 years ago.” After the continuous process of trial and error, they finally achieved the distortion rate lower than 0.5%.
The autofocus has improved as well. The models from the dp1 to the dp3 move the front lens element to focus, but this design increases the mass of the moving part. Heavy parts need more power to move and stop. And, it also affects the accuracy and speed of focusing. The “0” adopted inner focusing that moves inner lenses to shift focus. In the case of “0,” it moves only one lens and it works to reduce the weight of the moving parts, realizing more accurate and quicker movement. And, it was also made possible by having the ample space in the lens-barrel. As for the stepping motor to move the lens, it can move in increments of 15/1000mm. The algorithm for ranging was updated, too. Even an ultra-wide angle lens with greater depth of field has its “real” focal point. And, only by finding it, the high resolution of the Foveon sensor can show its full potential.
Kohno’s “lens with a good nature” was completed in this way and the “0” was ready to appear. At the same time, the code name “0” got its official name “dp0 Quattro” and they finally entered the stage of production.
What Foveon sensor brought
There are many important points to design a lens for the dp and they all work to bring out the full potential of the Foveon sensor. SIGMA chose to design a lens exclusively for the body and fix it to the body because they believe that it’s the only way to draw out the Foveon sensor’s maximum performance.
When they match the body and the lens at the production site, they adjust the alignment of the sensor of every camera by checking their image outputs. And then, they are ready to be shipped. The parts always have very small “dispersion,” no matter how strict the quality control is. In the case of interchangeable-lens cameras that cannot fix a lens permanently to the body, this dispersion has to be ignored as an error. In other words, the error is so small that it can be ignored. However, the dp0 is a fixed lens camera. If the error is ignored, it won’t make sense to be a fixed lens camera. The Foveon sensor doesn’t allow such error. In fact, while conventional sensors allow the dispersion of sensor alignment up to 0.06mm, the dp0’s dispersion of sensor alignment is further narrowed down to be within 0.012mm. This process can never be omitted no matter how complex and time-consuming it is.
There is one more thing. We have so-called “MTF curves.” Simply put, they indicate how accurately a lens reproduces the contrast and it is used to evaluate lens performance. In fact, SIGMA’s MTF measurement instrument has the Foveon sensor. This means that lenses designed for the Foveon sensor can only be measured with the Foveon sensor and the lenses for the dp are specialized for the Foveon sensor to this extent. In addition, SIGMA’s interchangeable lenses for each mount/manufacturer are designed with this instrument. Because it detects even a tiny fault that can’t be detected by conventional measurement instruments, the lens design naturally becomes more delicate. As a result, these lenses have overqualified performance for conventional non-Foveon sensors. The technology that was developed to design the Foveon cameras including the dp and the sd series worked to increase the basic performance of all SIGMA lenses.
On July 10, 2015, the SIGMA dp0 Quattro was launched. Even after the design is complete, engineers never feel relieved until the product goes into the world. But, on this day, Kohno, Tominaga, and all other people in the dp0 project finally heaved a sigh of relief. Nevertheless, they keep working to improve, prepare for the next products, and push forward with basic research. SIGMA will keep trying.