Ultimate-Resolution Cameras

There’s more to image quality than just pixel count; color reproduction, dynamic range and low-light/high-ISO performance quickly come to mind. But the more pixels an image sensor contains, the more detail its images can reveal, and the larger you can blow them up before the pixels themselves become visible to the eye (or the more you can crop into an image). This doesn’t mean that more pixels are always better, but it does mean when you need to make huge prints with great detail, more pixels definitely help.

Medium Format Vs. DSLR: Advantages, Disadvantages And Key Features

Today’s highest-resolution, single-shot devices are the 80-megapixel digital backs from Leaf and Phase One. These use 53.7×40.4mm CCD sensors to deliver images measuring 10320×7752 pixels (or very close to that). There are also medium-format sensors delivering 60, 50, 40, 37.5, 31 and 22 megapixels, but currently the 80-megapixel units are the "kings," dwarfing in both area and pixel count the best of today’s DSLR sensors, the 35.9×24.0mm, 36.3-megapixel "full-frame" unit in the Nikon D800 and D800E, which deliver images measuring 7360×4912 pixels.

Besides the advantage in pixel count, the much larger area of the medium-format sensors gives them an advantage in image noise. A larger sensor area can collect more light (photons) than a smaller one when you use the same shutter speed and aperture. And the more light, the higher the signal-to-noise (S/N) ratio. That’s because photonic noise (the noise carried by light itself and the main source of noise in most digital images with current cameras) increases as the square root of the signal: If you have 4 photons of signal, you’ll get 2 photons of noise, for a photonic S/N ratio of 2:1. If you have 100 photons of signal, noise will be 10 photons, for a 10:1 S/N ratio. If you have 10,000 photons of signal, you get 100 photons of noise, for a photonic S/N ratio of 100:1. A better S/N ratio means fewer fine details and steps of dynamic range are lost to noise, and that means better image quality, including higher resolution for a given pixel count (because less detail is lost to noise). Of course, there are other sources of image noise, but from a photonic noise standpoint, bigger is better.

Nearly all medium-format digital cameras are 16-bit devices, while DSLRs are 14- or 12-bit. This means the medium-format cameras can provide a smoother range of tones (especially important in portrait work, but almost always a good thing), theoretically, 65,536 tones or color shades per channel versus 16,384 for 14-bit and 4,096 for 12-bit cameras. (JPEGs from any camera are 8-bit—just 256 tones from black to white—one of many reasons to shoot RAW.) Of course, some of these steps are lost to noise (if the difference from one step to the next is less than the noise level, you won’t be able to differentiate those steps), but still, 16-bit medium-format images can reproduce a smoother range of tones and colors than 14- and 12-bit DSLR images.

There’s also dynamic range. Most of today’s medium-format digital cameras and backs can deliver more than 12 stops of dynamic range without resorting to HDR techniques. While much medium-format work is in the studio, where the photographer can fully control the light, the wide dynamic range is a big benefit with a wide range of subject matter and in outdoor situations where lighting is harsh. Medium-format sensors are CCD units, and they and the onboard processing are optimized for optimum image quality in good light (i.e., low ISO settings). DSLR sensors today are CMOS units, optimized for better performance at higher ISO settings (less light). So low-light specialists will be better off with a full-frame DSLR for the better high-ISO performance. But for ultimate image quality in good light, you can’t beat a medium-format sensor, especially those 80-megapixel units. (Phase One’s Sensor+ technology "bins" pixels—combines data from four pixels into one—to increase the ISO capability four times, at reduced pixel count: The 80-megapixel IQ180 back has a normal ISO range of 35-800 at 80 megapixels and a range of 140-3200 in Sensor+ mode at 20 megapixels. Medium-format companies are also working on medium-format CMOS sensors, which should provide even better low-light performance (as well as DSLR-like live-view and even video capabilities).

Many medium-format backs can be adapted to a wide range of medium-format cameras—Mamiya, Phase One, Contax, Hasselblad V and H, and even view/technical cameras. Thus, when an exciting new sensor (or camera body) becomes available, you need only upgrade that. With a DSLR, the sensor and body are a unit—if you want to upgrade one, you have to upgrade both (this is also the case with the Leica S/S2 and Pentax 645D cameras; they have built-in sensors, not separate backs).

Most medium-format cameras accept a number of leaf-shutter lenses. Leaf shutters offer the advantages of quiet operation with less "recoil," as well as flash sync at all shutter speeds, even at full flash power (focal-plane-shutter DSLRs have maximum flash-sync shutter speeds of 1⁄300 sec. or slower, and those that offer "high-speed sync" do so at much reduced flash power).

Many newer medium-format digital cameras provide touch-screen operation (only a few DSLRs do, none of them high-end models) for quicker and easier shooting. This, plus versatile tethered operation, and the superb image quality combine to make medium-format digital the ultimate choice for many pro and fine-art photographers.

Doing The Math

The more pixels an image contains, the larger you can blow it up before the pixels themselves become visible to the eye. Of course, factors like viewing distance come into play—the farther you are from a print when you view it, the less likely you are to see the individual pixels making up the image. Viewed at close range, those giant billboard images aren’t very sharp—and they’re pretty pixelated.

For the print publishing industry (magazines like this one), you generally want 300 pixels per inch. Thus, simply dividing the image’s resolution in pixels by 300 tells you how big you can publish it in inches: an 80-megapixel image from a medium-format camera measures 10328×7760 pixels, meaning you can run it 34.4×25.9 inches at 300 ppi. For a 36.3-megapixel image from a Nikon D800, pixel measurements are 7360×4912 so, dividing those figures by 300, you can run it 24.5×16.4 inches at 300 ppi. Images from 24-megapixel DSLRs measure 6000×4000 pixels, so they can be published at 20×13.3 inches at 300 ppi.

These are measurements of area, of course. Resolution is a linear measurement, so many lines per millimeter or per picture height. An 80-megapixel image contains 2.2X as many pixels as a 36.3-megapixel image, but, all other things being equal, it provides only a 1.4X increase in linear resolution. That’s still a goodly increase, but not the more than double you may expect from the more-than-double pixel count. Similarly, going from a 24.3-megapixel DSLR (6000×4000-pixel image) to a 36.3-megapixel DSLR (7360×4912-pixel image) produces just a 1.22X increase in linear resolution (lines per mm or per picture height). But in each case, file sizes increase with the total number of pixels, not the linear resolution: Double the pixel count, and you double the file size—and the space required to store the image and the processing power needed to process it. Note that amount and type of compression (none, lossless, lossy) also will affect file size.

That said, medium-format digital cameras and backs are expensive, starting at around $10,000 and going above $40,000. The 80-megapixel units run from around $33,000 to $44,000 (street prices, in some ca
ses including the camera body, in others just the back). They’re worth it to photographers who have the budget (or clients) to pay the entry fee, and the need for huge prints and the ultimate image quality in good light. Some photographers (pro and amateur alike) rent medium-format cameras when they’re the right tool for the task rather than buy them outright. And reconditioned, used medium-format digital cameras traded in for new models can be relative bargains.

Phase One IQ260 back

Ultra-Resolution Medium-Format Units

Phase One IQ280. Phase One’s new IQ280 adds some nice features to the original IQ180 80-megapixel back, including wireless reviewing and selecting/tagging images, remote operation and increased dynamic range (now 13 stops). Based around a big, 53.7×40.4mm, 80-megapixel CCD sensor, the back offers ISO settings from 35-800, Phase One’s Sensor+ Mode (which provides a higher ISO 140-3200 range at 20-megapixel resolution with no degradation of pixel-level quality, thanks to clever pixel-binning), and 0.7 fps shooting (0.9 fps in Sensor+ Mode), with a 1 GB buffer. Live view on the 3.2-inch, 290 ppi touch-screen display simplifies operation, as does tethered control via USB 3.0 or FireWire 800, and wireless operation via iPad or iPhone. Focus Mask shows in-focus areas of the image at a glance, there’s an on-screen level, and you can activate a B&W display, if you wish. The back is constructed of aircraft-quality aluminum, with all connectors and ports protected by auto-retracting hatches or rubber covers. Image transfer to memory cards is very fast.

If you can settle for "only" 60 megapixels, the new IQ260 back is the medium-format long-exposure champ, able to make exposures as long as one hour at ISO 140—remarkable considering the heat that a big sensor can develop over such a long period.

If you can settle for "only" 60 megapixels, the new IQ260 back is the medium-format long-exposure champ, able to make exposures as long as one hour at ISO 140—remarkable considering the heat that a big sensor can develop over such a long period.

Leaf Aptus-II 12 80 back

Mamiya Leaf Aptus-II 12 80. The top unit in Mamiya Leaf’s long-running Aptus line, the Aptus-II 12 80-megapixel back features a big 53.7×40.3mm CCD sensor that delivers images measuring 10320×7752 pixels and file sizes as large as 480 MB (16-bit TIFF). Its big, 3.5-inch touch screen provides an intuitive interface, but live view is provided only via tethered computer. ISO range is 50-800, and dynamic range is 12 stops. The Aptus-II R version lets you rotate the sensor internally, so there’s no need to remove and replace the back (exposing the sensor to dust) each time you want to switch between horizontal and vertical formats. Aptus-II backs are available separately or as a DM kit with the Mamiya 645DF camera and 80mm lens. The Aptus-II backs also support a number of other medium-format cameras from Mamiya, Phase One, Hasselblad, Contax, Bronica, Fujifilm and view/technical cameras. The Aptus-II 12 80 is by far the lowest-cost, 80-megapixel digital back.

Mamiya Leaf Credo 80. Featuring essentially the same 80-megapixel sensor as the Aptus-II 12, the Credo 80 is a new design, with live view on the iPhone-like touch-screen display (which includes touch strips outside the image area so you don’t have to put fingerprints in the image display area). With Phase One Capture Pilot, you can use wireless devices such as iPad, iPhone or iPod touch as remote image viewers and operate the camera remotely, or you can work tethered to a computer. The Credo 80 can shoot those big files at 0.7 fps. ISO settings range from 35-800, and dynamic range is 12.5 stops. The unit is weather-sealed, and the battery is internal rather than mounting exposed outside the camera. There’s no R (internal rotating sensor) version like the Aptus-II R, but you can mount the Credo on the camera in horizontal or vertical format. Like the Aptus-II backs, Credo backs are available separately or as a kit with the Mamiya 645DF camera and 80mm lens. And like the Aptus-II backs, the Credos also support a number of other medium-format cameras from Mamiya, Phase One, Hasselblad, Contax, Bronica, Fujifilm and view/technical cameras.

Hasselblad H4D-200MS

Hasselblad H4D-200MS and H5D-200MS. Hasselblad’s highest-megapixel cameras are the H4D-60 and H5D-60 60-megapixel units, but the highest resolution are likely the H4D-200MS and H5D-200MS. These feature a 50-megapixel sensor and can make single 50-megapixel exposures. In MS multishot mode, they can combine four exposures, shifting the sensor one pixel in each direction between them so that every pixel captures all three primary colors, eliminating Bayer-filter-sensor moiré and artifacts in the resulting 50-megapixel image. And they can make six-exposure captures (by adding two shots with the sensor moving in ½-stop increments) to produce 200-megapixel, 1200 MB images.

The new H5D models (not yet available as we write this in March) feature an improved ergonomic design with better weather sealing and an updated graphic user interface and quicker operation. Hasselblad’s True Focus II system with Focus Confirmation automatically compensates for focus-plane shift when you focus and then recompose a close subject like a portrait; the AF algorithms are also improved. HNCS (Hasselblad Natural Color Solution) yields better skin tones and other colors. You now can have print-ready ¼-resolution JPEGs along with the RAW files. The 3.0-inch LCD monitor is now higher resolution (460,320 dots). The camera comes with an eye-level viewfinder, which can be interchanged with several optional finders, including a waist-level one. Tethered operation is possible in-studio or in the field via supplied Hasselblad Phocus software or the Hasselblad Tethered plug-in for the supplied Adobe Photoshop Lightroom 4. The optional Phocus Mobile App provides wireless connectivity to iPad and iPod touch devices. A wide range of leaf-shutter lenses from 28mm to 300mm are available.

DSLRs offer a greater range of available lens focal lengths, from fisheye to supertelephoto, including superfast lenses for low-light work. And most of today’s DSLRs can do full HD 1080 video, not essential for many still photographers, but handy when you want to expand your ways of presenting subjects with motion and sound.

Sony SLT-A99

Highest-Resolution DSLRs

Medium-format cameras have the advantages in image quality and pro workflow, but full-frame DSLRs have their advantages, too. For one, DSLRs are much better in low-light/high-ISO work. DSLRs are also better for action shooting; they have better action AF systems and can shoot at much faster rates (more than 10 fps, for some). DSLRs offer a greater range of available lens focal lengths, from fisheye to super telephoto, including superfast lenses for low-light work. And most of today’s DSLRs can do full HD 1080 video, not essential for many still photographers, but handy when you want to expand your ways of presenting subjects with motion and sound. DSLRs also cost far less than medium-format cameras; full-frame models sta
rt at around $2,000 and top out under $7,000 currently.

Nikon D800

Nikon D800/D800E. Just as medium-format cameras have advantages over DSLRs in sensor area and pixel count, full-frame DSLRs have such advantages over smaller-format DSLRs (APS-C and Four Thirds). The current king of the hill is Nikon’s 36.3-megapixel D800, which comes in two versions: the D800, with OLPF, and the D800E, with the OLPF effect cancelled. The OLPF (optical low-pass filter, also known as the AA, or anti-aliasing filter) eliminates moiré and other artifacts that result from the demosaicing process with Bayer-filter image sensors, the type used in all DSLRs except Sigma’s Foveon-sensor cameras. The OLPF eliminates moiré and artifacts by slightly blurring the image at the pixel level. When pixel density gets high enough, diffraction serves as an OLPF, and the filter can be eliminated, resulting in even sharper images. Hence, the 36.3-megapixel D800 delivers amazing images, and the D800E even sharper ones, but with the possibility of moiré/artifacts in images of subjects with fine repeating patterns. (Medium-format digital cameras don’t employ low-pass filters.)

Up-Resing Software

If you need to make prints larger than your camera’s pixel count permits, you can increase the pixel count via up-resing software. Adobe Photoshop, Lightroom and Camera Raw all have up-resing capabilities. There are also specific up-resing software products, such as Alien Skin Blow Up 3, Perfect Resize from onOne Software, PhotoZoom from BenVista and Qimage from DDI Software. Free trial versions often are available, so you can try several and see which you prefer.

For more information, check out Photoshop guru Jeff Schewe’s article "The Art of the Up-Res" in Digital Photo Pro Magazine at www.digitalphotopro.com/technique/software-technique/the-art-of-the-up-res.html.

Besides that DSLR-highest, 36.3-megapixel sensor, the D800/D800E feature rugged construction with sealing against dust and moisture, slots for CompactFlash and SD/SDHC/SDXC media, and 4 fps shooting (5 fps in DX crop mode). AF and metering systems are the same ones used in the flagship D4 pro camera. Like the D4, the D800/D800E can autofocus in light levels down to EV -2 and with lens/converter combinations of ƒ/8 and faster. The 200,000-cycle-rated shutter provides speeds from 30 to 1⁄8000 sec., with X-sync up to 1⁄250 sec.

Canon EOS 5D Mark III

In the hierarchy of resolution, after the D800/D800E in the DSLR megapixel scale, you have the 20- to 24-megapixel full-frame cameras: the Nikon D600 and Sony SLT-A99 (24.3 megapixels), Canon EOS 5D Mark III (22.1 megapixels) and Canon EOS 6D (20.2 megapixels)—all capable of excellent image quality, but not providing the resolving power of the D800/D800E. And APS-C DSLRs have reached the 24-megapixel plateau, with Nikon’s D3200, D5200 and new D7100 joining Sony’s SLT-A77 and SLT-A65 at that level. These cameras have the pixel counts to produce huge enlargements, but their smaller sensors mean they will have more photonic noise than full-frame DSLRs, and thus somewhat lower image quality and resolution, although still outstanding.

Sigma SD1 Merrill And The Foveon Sensor

Sigma’s SD1 Merrill DSLR (and DP1, DP2 and DP3 Merrill fixed-lens compact cameras) features the unique Foveon X3 image sensor, which operates on a different concept than the sensors in other DSLRs and medium-format digital cameras.

The photodiodes ("pixels") in an image sensor can’t detect color; they just detect how much light strikes them. To get color information, conventional sensors employ a grid of red, green and blue filters called a Bayer array (named after the Kodak scientist who developed it), so that each pixel receives red, green or blue light.

The missing colors are derived from data from neighboring pixels using complex proprietary algorithms in a process known as demosaicing. This process produces moiré and other artifacts. To minimize those, manufacturers place an optical low-pass filter over the sensor, which slightly blurs the image at the pixel level.

Rather than use the Bayer filter grid, the Foveon sensor takes advantage of the fact that light penetrates silicon to different depths, depending on wavelength. The Foveon sensor stacks three layers of pixels, the top layer recording mainly blue light, the middle layer, green, and the bottom layer, red. Thus, every pixel site receives all three primary colors of light, so there’s demosaicing, no moiré and no need for the blurring low-pass filter. This gives the Foveon sensor higher resolution than conventional Bayer sensors of equal horizontal-by-vertical pixel count. (The Foveon sensor requires processing of its images, of course, but doesn’t produce the Bayer-sensor artifacts.)

Bottom line: The APS-C-format Foveon X3 sensor delivers images measuring 4704×3136 pixels (14.8 megapixels), but they provide much higher resolution than conventional sensors of that horizontal-by-vertical pixel count, and images rivaling those of 20-plus-megapixel Bayer-sensor DSLRs.

Leave a Reply