Features Getting Creative with Painter Profile: Kevin Kubota Hybrid and Digital Capture Columns Insight/On the Cover Digital Photography Departments Industry News New Products Problems & Solutions Archives

Rangefinder Magazine
April 2003

Hybrid and Digital Capture: by Dr. Glenn Rand
From Silver Halide to Pixels

Very long exposures create digital noise, thus the choice of film.

It is not an issue “if” digital capture will become the major professional photographic capture process… that will come to pass. Today we are transitioning between silver halide and pixels. While new and the future, digital capture is not necessarily the “end all.” The existence of hybrid imaging (scanning film to acquire digital images) admits this point.

As we move from silver halide-based photography to an electronic enabled future, let’s step back to look at the intersection of these processes. In today’s photographic workplace neither digital capture nor hybrid is superior in all situations. Each maintains its unique applications and each has certain advantages. Though digital capture may be our future, hybrid applications are a choice, not a cop-out.

Sensor Size Vs. Film: The comparison of resolution between digitally captured and film-captured images is first. A 35mm negative (ISO 400 color) has well over 15 million locations of image information. A 6x6cm film image has over 30 million locations. If we equate each color location on film as one of three primary colors used to create a digital image, then we can say that a 35mm film image is equivalent to a 15 MB or greater file.

	Another very long exposure shot well before dawn dictated film use over digital.

While today there are cameras with 14–18 megapixels, these are the exception. Most professional 35mm style cameras are 4–6 megapixels. Scanning backs can gather more than 100 million locations with long exposures and area array backs can be over 20 megapixels. Depending on the equipment, the relation to film of sensors is at or below the number of film’s sensitivity locations.

With interpolation to create color, a 5 million pixel (5 megapixel) sensor will generate a 15 MB file in raw 24-bit color and that, it can be argued, is equal to a 35mm film frame. But there are three factors that disagree with this conclusion. First involves the interpolation of color. Each color is not captured at each site, but the light is color filtered and not recorded as three colors at each site. However, in the file its storage requirement is three times the light captured. While the Foveon X3 chip does not have this problem, it has a smaller total number of imaging locations. For this reason the real detail value of a digital image is not its color file size but its pixel count.

The second issue is the fill factor. The fill factor is the percentage of the sensor’s surface that is covered by active receiver. Sensors are manufactured with a sizeable amount of the surface dedicated to the electronics needed to make the sensor function. These non-active areas provide no information for image formation. Today the fill factor for many professional sensors is about 60%. The detail from areas not covered by active areas must be ignored, combined or interpolated.

The last of the three comparisons of resolution is sampling rate. Sampling is the amount of detail that is captured. Abstractly, a minimum two pixels are required to record a detail.

This panorama was shot on large format, using swings and tilts with a super-wide lens at a very small aperture for depth of field control; all of which dictated the image to shot on film and not digitally.

When these three issues comparing pixels to image locations of film are taken together it becomes clear that the issue is not simple. With the exception of some of the newer high pixel count sensors available in 35mm style cameras, today film still provides better detail.

It can be suggested that the use of large format scanning back can gain film resolution. While the file sizes from these backs is greater than 35mm film, these backs are only suitable for studio/still-life imaging.

Lenses: Several factors including the effects of modulation transfer, pixel size and optic requirements of the micro lenses on the sensor limit lens effectiveness. Primarily, the angle of incidence of the light striking the sensor can be critical. These factors can limit wide-angle lens use or with large format cameras can limit the use of camera movements.

Color Interpolation: Color is important in most imaging today, and herein exists another problem for digital capture. To capture color, sensors use multiple sensors, movement or matrix of filters.

Most common is the matrix with a filter over each pixel arranged in alternating red/green or blue/green rows, columns or diagonals. Color is not captured in its hue at each pixel but as a value of red, green or blue. Image color is interpolated by using color information of adjacent sites. Accuracy of color is limited to the sampling combined with the interpolation algorithm. Potentials exist for inaccuracy and artifacts such as moiré patterns.

Since most scanners use either tricolor imaging or trilinear sensors, the color information is recorded as combined hue at each site. In this way the color make-up of the image is representative of the color in the film or print being scanned.

While scanning backs use this type of capture, they require long exposure times to make their images.

	This image of four elevator doors was created with a view camera, using swings and tilt, a small f-stop and a long exposure. Film exposure was indicated.

Dynamic Range: The largest disadvantage of hybrid imaging is film’s short dynamic range. While the human perceptual system has a range of one million to one or 20 stops of dynamic range, color film records over only a 64:1 range or about six f-stops. Digital capture can acquire about 11 stops or about 2000:1.

The dynamic ranges of the two capture methods are seen in the highlights and shadows as well as the contrast when accurately exposed. While the film will have a solid black and a clear white, the amount of detail captured and available in both highlights and shadows is greater in the digitally captured image.

Spectral Sensitivity & Color Gamut: Silver halide requires more spectral energy to accomplish exposure than a sensor and this leads to a differential in spectral energies effective for exposure. Digital capture devices are far more sensitive to infrared energy than film. At the same time sensors are noticeably inferior to film in the blue and violet spectral areas.

Human vision has the ability to see about 10–13 million colors. Digital imaging with 24-bit color defines over 16 million colors. Because of the way the color is captured through filters, these are different areas of the generalized color space than vision. This is also true of film. With reduced dynamic range of film, a different color space is defined. The output color gamut of both will be subjected to the color management of the digital system.

Electronics: Today’s digital cameras and backs are complex computers and these have the requirements of other computers. First, the issue of power is obvious to anyone who has used a digital camera for an extended time. Electronic requirements of these cameras mean that batteries or power packs need to be changed regularly.

Digital silver print from 8x10 negative, 1981.

Particularly in adverse weather conditions, the battery power or operating system can be compromised. This is true for both heat and cold.

Further, certain types of digital capture systems need connection to computers to maximize the effect of the system. While there are portable computers, this only compounds the problems. Using cameras with removable media cannot eliminate this issue. These media do not remove their dependency on microcircuitry that is both fragile and affected by adverse environmental conditions.

Noise makes using digital capture less satisfactory with some imaging applications. Noise happens when stray electronic information affects the sensor sites. It is exacerbated by heat and long exposures. Noise shows up more in dark areas, making evening and night photography problematic with digital capture.

Once the image is made it must be stored and archived. A six megapixel capture at 24-bit color creates an 18 MB raw file and larger files with higher bit depth. Today the choice of removable media is wide but with an unsettled standard. With large files, the write speed or how fast images can be moved to the media, also becomes important.

Regardless of the type of capture, beyond how much can be saved on the removable media or scanned into RAM, there is the issue of archiving. It would be nice to say that if an image is committed to a CD or DVD that would end the concern. These media are far superior to the removable media for permanence but the usefulness of the media is determined by technologies that will replace CD and DVD. On the other hand, film images have a long useful life with careful storage and scanning when needed.

Workflow: For many photographers the time that separates the act of taking an image and its finished state is very important. With hybrid imaging there are at least two additional steps compared to digital capture. For many photographers the issue of speed from start to finish of the imaging process presents a big positive for digital capture.

Comparison: Comparisons are best split into technology and process issues. In technology the first concern is resolution. Resolution favors hybrid digital in all areas except 35mm based with low to moderate resolution. Dynamic range is better with digital capture—far better. Color has no effect in the comparison since they cover different color spaces. Film-based systems are more upgradeable. Finally, in today’s market, costs favor hybrid approaches. When looking at just the technological side, it can be argued that the advantage falls with hybrid photography.

	Image shot on film with 4x5 for perspective control.

On the process side the point is easily made for digital capture. Digital capture is far less bother with film processing and scanning required for hybrid imaging. For the above reason workflow is simpler for digital capture. The last part favoring digital capture is the environmental issues. The only part of the process group that favors hybrid is its flexibility. For prints there is an advantage for digital with color and for silver halide for black-and-white. This leaves the comparison the reverse of technology for digital capture.

The last area of comparison is that of application. This is the crucial issue. Different applications need different approaches and some are better suited for each. Large format, perspective controlled with depth of focus issues, super high resolution, long or low-light imaging and multiple exposure are all instances when a hybrid approach to digital photography is an advantage. When the speed of workflow with moderate resolution, table top/still-life and general photographic usage are considered then digital capture has the edge.

Glenn Rand has worked as an artist and professionally in photography for more than 35 years. He started experimenting with computer-assisted photography in 1980. He has published and lectured extensively about photography and digital imaging. Presently Dr. Rand teaches in the graduate program at Brooks Institute of Photography in Santa Barbara, California.