The digital SLR first surfaced in various forms in the mid 1980s, in forms as prototypical as the 120MP sensor demonstrated by Canon is now – using tube, then CCD technology and still video capture. The CCD sensor for single, still shot capture would come into its own with the Kodak DCS and related systems, though these models would continue to be a compromise of five-figure prices and quality barely better than a single-frame capture from a Laserdisc – but in the early days $23,000 to capture 700 pixel wide video frames was not unheard of, with devices like the Fuji ES-1 holding a small specialist market.

A decade after the components were in place of CCD, Bayer sensor and microprocessors, film was still very much the dominating force behind photography – and Kodak’s own pioneering work (including a 1.4Mp sensor as early as 1986) clashed with the realities of what was on sale at the time; near VGA resolution field capture cameras. Nikon’s own lead at the time was lost in part due to their production of the first print-quality system – the QV-1000C – a camera which unlike the Canon models needed no credits to justify the poor results in pioneering newspaper appearances.
Ironically perhaps, given current developments, still video “field capture” devices were the heart of digital camera development throughout the 1980s. From backs on various DSLRs to dedicated systems from pioneers like Mega-Vision, floppy discs and still-frame capture were standards and the process of capturing a frame from a video feed advanced almost as quickly as the stills camera itself. Early prototypes started to appear around the end of the 1980s that moved from the video-based format to a pure data model with the availability of a new standard for flash memory – PCMCIA – spurring the smaller models on. Kodak’s first true digital SLR, the DCS-100, appeared in 1991 costing $30,000 and sporting an ISO 100 1.3Mp sensor. Meanwhile, Sony announced the first use of microlenses on the CCD to increase sensitivity…
The 1990s remained dominated by film – with digital cameras split into three primary formats and realistically, only one affordable one until the late 1990s. The high-end studio systems moved towards scanning backs, providing unheard of filesizes and resolutions with incredible quality. The low end, typified by the Canon Ion, Apple QuickTake and Dycam, refined the still video capture process (and my first digital cameras were an Ion and a Tamron Fotovix, the latter being nothing more than a video camera and a video digitiser in the Macintosh IIx I used – though it was still good enough for playing with Photoshop!). The mid-range of DSLR/DSLR-alike systems would take some time to settle, and mostly consisted of Kodak CCD setups shoehorned into a variety of Canon and Nikon bodies. The search for colour accuracy and maximum resolution was typified by devices like Kodak’s DCS-200 “Wheelcam”, which used three exposures to assemble one colour image assisted by a rotating filter disc. This idea had been used previously for colour TV capture and is still being used today as an alternative to Multi-shot backs by Mega-Vision; Leaf and Phase One started out using this system on studio cameras, and tellingly the original colour discs are not the same density throughout – the red includes a 0.6ND filter, the green a 0.8ND filter.
1999-2000 saw the release of cameras like Nikon’s D1 and Canon’s EOS D30 and the announcement of many others, as Kodak’s hold on the digital SLR market weakened following investment in in-house production to access a wider market. So it is at this point that we can fairly consider the wider “digital photography” era to have begun – the accessible professional system, as opposed to low resolution gadgets and compacts or slow, high resolution scanning systems.
Canon’s first offering shipped with 4Mp on an APS-C sensor with 1.6x crop. Nikon’s offering was a 10.8Mp sensor with a 2.7Mp CFA and 2.7Mp output file; the D1 carried a Sony sensor that would set the pattern for professional DSLRs for some time.
And, whilst the speed of progress for these first DSLR seems to have been rapid and relentless – that first sensor in the D1, 10.8 million sensels at a pixel pitch of 5.9µm, would underpin half a decade’s worth of cameras from Sony and Nikon. 2000 also saw Fuji announce the FinePix S1 Pro at 3.1Mp interpolated to 6.2Mp.
The race was now under starter’s orders. We had 2.7Mp CCD from Nikon at $6000, Canon’s D30 brought 3.1Mp CMOS sensor in at a lower price point, and Fuji’s SuperCCD architecture was already beginning to muddy the resolution definitions by using paired high and low sensitivity sensels to boost dynamic range. And that initial race was – despite the relatively high cost of the cameras – quite impressive. Two years later PMA would be host to a glut of “affordable” DSLRs from consumer to professional quality – and Sigma’s first entry into the market after decades of lenses and high value film cameras would threaten to disrupt the entire model. Outside of the colour-disc/filtered monochrome models (and a couple of expensive, clunky three-CCD field capture derived cameras), everyone was relying on colour filter arrays and interpolation. With a new 3-layer CMOS process over a 3.3Mp spatial resolution, Sigma did away with the blurring AA filter, did away with interpolation, and delivered class-competitive native resolution at a class-beating $1799 price point. In an era of TIFF files and the remaining hangovers of SCSI, the Firewire connected, raw-only SD9 was far from unusual – and it looked like Sigma had a world beating camera on their hands.
Sigma’s SD9 didn’t have it all, despite the stunning gauntlet thrown down. Early models handled by reviewers had high luminance noise (soon rectified in a firmware update, but notably not retrospectively corrected in online reviews, when the PR efforts of many firms saw coverage revised and re-interpreted) and there was the inevitable brand resistance from people already familiar with the compact, CCD-based digital cameras from Nikon, Fuji and Canon. Had Sigma started out with a compact, who knows – the only compact Foveon camera to make it to market during that era was a Polaroid-branded Chinese product, the x530, which was blighted with engineering problems and poor build quality.
Despite the reticence of reviewers to really push the SD9’s strengths in dynamic range and sharpness, it soon gained a strong cult following. Sales figures suggest that the initial production run sold out well within the gap between the SD9 and SD10 – a marked difference from the SD14 – and user feedback saw the real drawbacks of the SD9 far removed from the image quality. The dual battery setup and reliance on AAs was not uncommon when the SD9 would have been in the design stage, but by 2002 Li-Ion battery packs were powering the competition. Whilst Sigma got to work evolving the SD9 into the much improved SD10 the DSLR market grew massively: rather than sitting as a $1799 body with only a handful of DSLRs on the market (the competition at the time was the Nikon D100 & related Fuji S2 Pro, and the Canon D60 – all of which cost more); of these four cameras announced in 2002, two of them used a different sensor technology to traditional Bayer GRGB CFA layouts, whilst the two traditional Bayer models used differing silicon technology – Canon sporting the relatively young CMOS, also used by Sigma. The spread of sensors in 2003 included two full-frame 35mm models announced at the end of 2002, the Canon 1Ds at 11Mp and the Kodak DCS 14n at 14Mp, costing £7000 and £14,000 respectively. Contax’s N-Digital, a 6Mp full-frame system was announced, and went on sale only to be withdrawn rapidly (Contax themselves disappeared from the market in 2005).
If one thing can be drawn from this period it’s the rapid expansion of the DSLR market. From 1999 and the D1 (ignoring the very expensive, specialised and complex solutions from Kodak and others previously), the market had leapt in four years to full-frame, 14Mp systems, cameras touching the $1000 price point, and essentially doubled the resolution buyers could expect making 6Mp the minimum.
That expansion was not necessarily driven by advances in sensor technology though.
The SD10 was announced near the end of 2003, and launched in 2004 to critical acclaim – with the operational improvements, faster computers to deal with the raw-only image files and an overall refinement of the concept, the SD10 put Sigma firmly on the DSLR map – and with a much lower UK price at £1250 with two lenses.
By that time, the DSLR market had seen new entries from Pentax with the *ist D, Nikon and Canon had brought out entry-level models with last generation sensors – the D100 sensor reappeared in the Nikon D70, the Canon 10D’s sensor emerged in the groundbreaking (in price terms) 300D and prices began to creep down. Nikon’s D2H brought a third sensor technology to the market – LBCAST, or Lateral Buried Charge Accumulator and Sensing Transistor array – though this would prove to be a dead end as a definition of sensors, lessons learned undoubtedly shaped future CMOS units.
At the upper end of the market, Kodak’s DCS Pro models evolved with a Canon-compatible version based on the Sigma SA-9 (it looks remarkably similar to an SD from above, and it’s widely held that Sigma assembled the Kodak DCS SLR Pro/c) and a 12Mp sensor appeared in Nikon’s D2X – a 2004 announcement that would set the template for Nikon’s APS-C sensors until 2010.
in fact, the expansion of the market saw each new introduction migrate into newly created levels of camera. That first 10.2Mp sensor in the D1 was refined with new CFA layers to first fully utilise the resolution (from 2.7, to 5.3, then the full 10.2) before heading into mid range, then entry level cameras; the last camera released with the 10.2Mp CCD was the 2009 D3000, which shared ISO sensitivity and dynamic range with the first full 10.2 implementation in the 2005 D200; meanwhile Sony and Pentax launched models based on that sensor.
The same pattern exists with the 6.1Mp CCD sensor introduced in the D100 (2003, sensor current until the 2009 Nikon D40) – with the ground breaking camera retailing at nearly 10x the cost of the last, closeout prices of the entry-level model.
Canon is unusual in that nearly every generation of their cameras has seen small, incremental and tailored changes to the sensor architecture until recently – when their new APS-C 18Mp sensor has appeared in three different tiers of camera. Almost inevitably, the late 2009 (and still current) 7D suffers by comparison with the entry-level 550D whilst the almost apologetic gap-filler 60D sells purely on the strengths of not being an ‘entry level’ DSLR – there’s so little between the 550D and 60D as imaging devices.
Whilst Sigma’s pattern of development has seen sensors valid for about four years (though no hard-and-fast rules can be drawn at this early stage), the multi-tiered strategy allows Nikon and the other manufacturers to gradually increment their product ranges with a minimal genuine advances. Sigma’s perceived slow pace of development is actually a series of leapfrog moves; if the sensors were otherwise identical Bayer tech, Sigma would have been shifting 9.9Mp when 6.1Mp was pretty high, 14.1Mp when 10Mp was very current – and now, approaching 2011, Sigma has announced the ground-breaking 46Mp sensor of the SD1.
With four years, roughly, for each model (SD9/10 – 2002-2006/7, SD14/15 2006/7-2010/11) Sigma’s sensor lifespan compares remarkably well as a “current” technology with the CMOS 12Mp (2004-2010 and counting), the CCD 10.2 (albeit with CFA upgrades, 1999-2009 – 2005-2009 as a 10.2Mp unit) and CCD 6.1Mp (2002-2009). Even Fuji’s unusual SuperCCD sensor – seemingly a blip on the horizon now development appears to have ceased – managed 2002-2009, with the initial S2 Pro release being refined to the S5 Pro and IS Pro. As with Sigma, Fuji’s release schedule focused on the middle ground, affordable professional bodies, leaving no entry level and really pushing high-end imaging without the more specialised features like high-speed capture or extreme body engineering – had Fuji produced, 18 months after the S2 Pro, an S3 (D100) equivalent in a D80 body would they have had more success in the marketplace?
The biggest changes in the market have been live view (which requires CMOS technology, and was mooted as one of the benefits of the Foveon chip initially) and video, first seen on Nikon’s D90 of 2008 using the Sony-manufactured 12Mp CMOS sensor; since that point 18 months Nikon has released four more APS-C video capable models, two of which use the same sensor.
With the DP1x/DP2s and SD15 models suggesting the X3 14.1Mp sensor has not been allowed to stagnate since first release, it’s possible that the SD15 will finally find a niche in the marketplace for another 3 years as an entry level counter to the high-end, high quality SD1. Sigma will never effectively market the facts of their development, as people buy so heavily on numbers and only see the specifications as compared – nowhere do reviews and comparisons make it clear that even if the output file is 4.5Mp, the sensor and processors are, were, handling 14.1Mp of data – but throughout the development of the Foveon-based cameras, the cutting edge development and high-end work that has gone into the system is apparent once you look beyond that output size.
As the other manufacturers release a new generation of cameras, a new era of sensors has dawned – and Sigma is right there, with a sensor delivering 3x the base number of pixels available in the latest models. What does the future really hold for the SA-mount, Foveon, and Sigma’s camera range?
Whilst the “big two” players in digital SLRs have followed almost identical paths – introducing a high-end body initially at a high price, followed up with a medium range and then entry level offering (Nikon currently have a very interesting split, as the “consumer” D7000 is almost entirely better than the mid-range “small pro/prosumer” body) – with the technology filtering down in a predictable cycle through the product range, it’s not the only path. Pentax’s initial offerings were “mid-range” at a near consumer price point, Sony and Samsung occupied very similar territory (though Sony’s product range is now overwhelming).
The DSLR market is certainly reaching saturation point for “tiers”, if not users. Sigma’s offerings have always been priced to sit with the mid-range in RRP; initially trying to outgun the top-end in image quality whilst letting the camera bodies sit as “good enough” in tech, and “capable” in ability. The SD15 is the first Sigma DSLR to have a lower initial price point, though many users clearly feel it should be lower still given the age of the sensor. The SD1’s announcement at Photokina suggests that the professional market is once again in Sigma’s sights, leaving the SD15 to offer an entry-level system (with the move to SD cards, shared architecture with the DP-series and in-house development hinting at cost reductions overall).
At this stage, with the SD1 months away, it seems careless to speculate. However, the way it looks to be developing is this:
Sigma’s SD1 will, if rumours are to be believed, sell around the same price as a mid-range/small pro body from almost any other manufacturer. The new magnesium body, custom functions and overall quality certainly point at a physical product capable of competing in that bracket. Sigma’s firmware and UI compares well with other cameras outside of the Nikon/Canon duopoly; Leica’s menu structure on even the £19,000 S2 system is similarly functional. Whilst some reviewers may focus on pretty graphics and menus, photographers generally want the functions to work and don’t mind what colour the writing is or how well designed the icons are.
The 46Mp sensor will not be a one-off, for a single camera. If we delve into the realms of future-gazing, then I’d expect Sigma’s product plan to see the DP1/DP2 models updated to the 46Mp sensor by 2014 (still some distance away, so don’t go holding back on those purchases!) – and given the success of the Leica X1 and the appearance of Fuji’s large-sensor compact, the DP series has proven not only to be prescient, it’s also been very good value. Particularly with regard to Leica’s model, a 46Mp DP with a more substantial body and pro-focused design could be a high-value product occupying a very different space. Fixed focal lengths have so far been the norm here.
The SD15 body – with its SD card and new shutter design – may also see a 2013/14 announcement of a 46Mp model at a lower price. If that’s the case, then Sigma may discontinue the 4.5Mp sensor altogether (as was the case with the 3.3Mp) or the SD15 may continue as an inexpensive introduction to Foveon/Sigma cameras.
If the rumoured interchangeable lens DP model surfaces, that may serve as the entry-level.
Sigma’s next step upwards is full-frame. The SD1 should have a bigger sibling to consolidate the product range, and full-frame is something every Sigma user has been crying out for since the 5D/D700 showed it could be done at an affordable level – yet the current Foveon sensor indicates smaller sensels have been more successful for them. Few people would want more than 15Mp of genuine, uninterpolated resolution on a 35mm frame – many would claim that the glass simply couldn’t cope, though that never stopped film from retaining a pleasing quality despite some pretty unsophisticated lenses inhabiting the market!
Of more interest from that perspective is where the high end is going generally. Canon have yet to address the disparity between the 5D Mk II and the high-end 1Ds Mk III’s resolution and capabilities, Nikon’s D3x seems a little orphaned as the rest of the range gets video and resolution increases seem to be creeping in from the lower-end upwards. Pentax may have the answer in the 645D – blending medium-format sensor and sensor area with typical DSLR capabilties at a price point not too far from the heights previously reserved for Nikon and Canon’s top of the range bodies. If Pentax can move quickly and get enough of the cameras (and more importantly, lenses) into the hands of working professionals, the vacuum left by the absence of new tech in 35mm, and perceived inaccessibility of MF system cameras could belong to this new breed of camera.
A lot hinges on the success of the SD1, a new era for Sigma and potentially an opportunity to split the product line and cater for a wider range of users. With the current emphasis remaining on pure photography and the opportunity to compete on output file size whilst (assuming all other aspects of the familiar Foveon sensor remain appropriately unchanged!) delivering unparalleled sharpness and detail, pure photographers may be prepared to trade video capability and value-added features for that ultimate goal. With the 70-200 OS, 85mm and 50mm taking Sigma’s lenses into the high-end, maybe the SD1 will follow suit.