An Engineering Approach to Consistent Images:
Control of Video and Film Image Quality in TV Production

By Fred Benedikt, David F.E. Corley, and Robert Ross

Engineers, video operators, colorists, and producers all share the goal of excellence in television image quality.  This paper explores the reasons why some popular and widely used to standardize video and film quality have met with limited success, and a practicable solution is proposed.  Previous efforts have, by and large, lacked the precision to ensure consistent high-quality images, particularly in regard to tonal range and color.  The system under discussion, although providing production teams with infinite freedom for artistic expression, controls the technical aspects of image quality from camera setup through the post-production process.

Using familiar waveform monitoring techniques, the Ambi/Combi System is comprised of two elements: (1) Optical Signal Generators (OSGs) precision test targets, and {2) a unique portable illuminator.

The system, now in day-to-day use in network engineering and studio operations, can provide similar benefits to the production community.  A major advantage of using this proven engineering tool through all stages of production is that engineers, operators and production personnel can now all speak a common language.  The system benefits television film production and scanner/telecine setup.  Ongoing tests with major film transfer houses are also discussed.

Image inconsistency, particularly scene-to-scene and program-to-program color variations, continues to be a dilemma facing program producers and broadcasters.  Efforts to address the problem have been either technically inadequate, too time consuming, or too expensive.

From the earliest days of motion pictures, test patterns have been used in an effort to control image quality.  Similarly, from the first days of television, signal generators and optical test patterns have been used for the same purpose.  Although there are similarities in the alignment practices of the two media, there are also major differences.  For a number of reasons, film has been more forgiving than the electronic image.  Screened in a darkened theatre, movies have the benefit of human vision adaptability, particularly in regard to color.  Film also has a considerably wider dynamic range than a television system, and, within limits, color/density variations can be corrected in the print stage.

This reality has resulted in two cultures.  Film production has tended to favor the traditional, less rigorous, subjective approach.  Television, on the other hand, having access to signal measuring equipment, has used objective control methods.  To satisfy these requirements, two types of test patterns have evolved: front-lit reflectance test charts, and rear-lit slide transparencies.  Both have their benefits and their deficiencies.

Front-Lit Test Charts

These are relatively simple  to produce and inexpensive but have three major drawbacks that contribute to, and can, under certain conditions, virtually ensure inconsistent results (see Fig. 1, Angles of Illumination):

• Surface flare - varies with camera and lighting angles, also on the type of illuminant (spot or flood)
• Test image inconsistency - occurs from the basic construction materials, surface abrasion and handling
• Contrast ratio - can vary depending on the ambient or background light level behind the camera

The inconsistencies in image quality arising from one or more of the above can be subtle or dramatic, affecting both tonal range and color saturation.  The effects are particularly apparent when attempting to match multiple television cameras using a front-lit chart.  The use of such charts with film cameras can induce similar inconsistencies, which, along with exposure, film stock, and processing variations, are extremely difficult to isolate after the fact.

Rear-lit Slides

These eliminate the flare problems associated with reflectance charts and are considered technically superior, but slides do require light boxes or integrating spheres that are awkward and cumbersome to use; often have poor light evenness; and have color temperature or spectral characteristics that do not match those of the scene.  This has tended to limit the use of rear-lit slides to camera manufacturers, maintenance shops, and other specific engineering applications.

A critical point to consider with any test pattern is accuracy and repeatability.  Using imprecise test materials not only limits long-term repeatability and consistency, but also the ability to conform to a particular industry standard, i.e. SMPTE or EBU.

A Practical Solution: The OSG

Rear-lit test targets, which produce relevant video signals, already existed.  The challenge was to design a lightweight, portable and sturdy illuminator that would provide even illumination having spectral characteristics that matched those of various real-life shooting situations.  With input from production teams and video engineers, a number of designs were prototyped and field tested.  This work resulted in the Ambi illuminator (see Fig. 2, Ambi Senior lit from below).  While originally designed for use with ambient light, a dedicated light source matching that of the key light of the set is now typically used to illuminate the test target.

The following are Ambi's basic components:

• A neutralized light diffuser
• A test target holder for OSGs
• An adjustable mirror

Unlike light boxes and integrating spheres, Ambi's light source is typically 1.5m (5 ft) or more from the diffuser, which tends to produce inherently even illumination.  A companion stand incorporating a standard 5/8" lamppost at its base enables any appropriate light source to be used (see Fig. 3, Ambi Senior on stand).  Although the Ambi Senior Illuminator will accept standard 8 x 10" test slides, it was specifically designed to provide optimum performance from Combi OSGs with an active scanned area of 9 x 6.75".

In response to the increasing interest in wide format television, a 16:9 version of the Ambi illuminator was introduced at the 1997 SMPTE Technical Conference in New York (see Fig. 4, AmbiWide 16:9/4:3).

This wide format illuminator uses the same vertical components as Ambi Senior.  The overall dimension of OSGs used in Ambi Wide is 13.5 x 8", with an active scanned area of 12 x 6.75".  By simply inserting two opaque panels, 3:4 ratio 8 x 10", OSGs can also be used in Ambi Wide.

In addition to its traditional use on a stand, by rotating the mounting yoke and adjusting the mirror, these illuminators may be hung from a lighting grid.  Similarly, with the mirror set to accept light from above, and the mounting yoke tilted back to act as a third foot, Ambi may be used on a news anchor's desk, maintenance shop bench, or any flat surface.

OSG Background

Although the design of the illuminator is important, the heart of the system is the test target, whose precision is vital.  Combi OSGs are the latest in a series of test materials produced by DSC Laboratories.  DSC's first test targets were produced for the CBC in 1966.  These were 35mm and 3 1/4 x 4" gray scale slides, designed for color telecine alignment.  To eliminate the Callier effect of black-and-white film (the silver grains transmit more red and green than blue light), these test patterns were produced on color film stock.  Because the sensitivity of the different dye layers in color film is nonlinear, each step of the gray scale had to be exposed separately.  Generating accurate neutral gray scale steps required the design of unique computer-controlled printing equipment.

Standard 8 x 10" formats for use with studio cameras followed later.  Unlike today's virtually unbreakable OSGs, these early camera setup slides were bound in glass and made as two separate patterns.  One, a crossed gray scale, was used for setting camera transfer characteristics.  The other, a six-step color bar, was used to adjust a camera's matrix for optimum color reproduction.

Combi Test Targets

In 1989, an improved color bar was combined with the gray scale and became the first CBC/DSC Combi-1 OSG.  The luminance and color data provided by this pattern are all that is required for complete and convenient video level alignment (see Figs. 5 and 6, Combi-1 OSG and Combi-1 color calibration sheet).

In the early 1990s, Combi-1 became part of CBC's policy for evaluation and acceptance of video cameras to SMPTE colorimetry.  A calibrated EBU version is also available.

Combi-1 is produced in both 9- and 11-step EIA (Electronic Industries Association) crossed gray scale versions, each having six midsaturation vector colors designed to produce red, green, blue combinations of 40 and 80 IRE (Institute of Radio Engineers) units.  Individual calibration sheets accompany every OSG.

The signals generated are expressed in terms of component RGB levels, composite NTSC levels, and International Commission on Illumination (CIE) L*U*V color error.  For most practical applications, the data sheets provided are unnecessary.  The 100 IRE (white) and cross-over steps have a tight +/-1 IRE tolerance and the optical color bars have minimal color error with a root mean squared (RMS) of <3 CIE L*U*V units. For comparison, well-matched cameras typically have a spread of -10 CIE L*U*V units.

Combi-1W is a 16:9 version of the color bar/gray scale pattern (see Fig. 7, Combi-1W (Wide) 16:9/4:3 ratios).  In addition to the standard (upper) Combi-1 color bar (40/80 IRE spread), this version includes a lower color bar with a 50/70 IRE spread.  With vectorscope correctly adjusted for Combi-1, the upper colors are designed to fall in the center of the boxes and the lower colors on the vectors at half amplitude.  Deviation of the signals from a straight line would indicate non-linearity in color reproduction.  Combi-1W also has four flesh tone patches:light Caucasian, mid-Caucasian, mid-Oriental, and mid-Black.  There are in addition two "Superwhite  patches to set clipping.

As Combi-1 combines several functions to facilitate color and gray scale alignment, Combi-2 (see Fig. 8, Combi-2W - 16:9/4:3 version) combines geometry, resolution and other unique signals for both digital and analog equipment.  Patterns include multibursts up to 800 LPPH (10MHz NTSC), linear ramps, zone plates back focus, smearing/streaking, plus vertical and horizontal hyperbolic resolution wedges.  Both Combi-1W and Combi-2W include 4:3 framing arrows.

OSG Definition

Both Combi test targets are true calibrated OSGs, designed to furnish precise and repeatable electrical signals which are easily interpreted on standard waveform and vectorscope equipment.  This concept has not changed since 1966 and has proven its value over the years.

The manufacturing process and materials used in making test targets are very important.  Typically, test charts and transparencies are photographically screen printed, producing an image consisting of thousands of dots.  This dots can "beat" with a charge-coupled device camera's sensor pixels, creating aliasing and moire effects.  Combi series OSGs are produced on ultrafine grain continuous tone materials, which eliminate these problems.

OSGs - Exploring the New Potential

Well-proven in video camera alignment and matching, the Ambi/Combi System's greatest potential is yet to be fully utilized.

Current users have all the essential components to use the OSG concept as a production reference standard.  Consider the following logical extension of the system (see Fig. 9, Ambi/Combi on the set):

1. Move Ambi into the set, as you would a grey card or chart.
2. Illuminate the pattern with the same type of lighting as the scene.
3. Record a few video or film frames at the beginning or end of a recording session.

These steps capture all the information needed to link the original real image with the electronic representation of the image.

Different organizations have different reproduction objectives; to conform to an international standard; achieve accurate color reproduction; and define a repeatable custom look.  Once a reproduction goal is established, it is imperative that the reference standard be reliable and consistent.  The precision of Combi test targets, in conjunction with and Ambi illuminator, ensures achievement of any image reproduction goal, quickly and effectively.

Using an OSG as a production reference provides a midpoint setting from which creative magic can be performed.  However, only calibrated and precise targets, lit correctly and consistently, can provide long-term repeatability.  This applies equally to both video and film production.  Ambi/Combi also enables film-to-tape, or tape-to-tape, color correction and image matching.

There are a number of reasons that OSGs work where front-lit charts fail.  In addition to the three shortcomings mentioned in Fig. 1, there is the all-important artistic component.  Lighting directors and crews go to considerable lengths to attain a certain "look" for a scene or production.  This often involves color gels and/or mixes of different types of light sources.

There are three possible scenarios to this situation.

1. The scene is shot without any test reference, resulting in the long painstaking process of trying to reproduce the original desired "look" in post-production.
2. A gray card or other front-lit test chart is shot on the set as a reference.  Depending on the type of lighting and camera/lighting angles relative to the chart, this can create numerous problems.  With many variables to track, this practice has the serious potential of changing, or neutralizing the artistic component of the lighting either during production or in post-production; an even worse scenario than scenario 1.
3. A few frames of the Combi-1 OSG are recorded in the Ambi illuminator before or after a scene; this provides a solid reference for the colorist, enabling the scene to be reproduced as the director intended - consistently, time after time.

Ambi/Combi in Film Production

Equally as effective for film as for video, the similar method described below enables a cinematographer to control image quality and its reproduction in the television system, to any desired reproduction goal.

The primary object in establishing an image quality control system for motion pictures is the determination of normal or correct exposure.  Experience has shown that exposing the average well-lit scene on the optimum portion of a film's characteristic curve will result in the most pleasing reproduction of the original scene.  This will ensure that maximum scene detail is available for post-production manipulation.  Placement on the film curve is established by evaluating the scene's light level using an incident light meter, set to the film manufacturer's suggested exposure index.  The Kodak laboratory aim density (LAD) system establishes the optimum position on the curve using an 18% gray patch.  Correct placement on the film curve can be measured by reading the RGB, negative density values produced by shooting an 18% Combi OSG.

Once the exposure level has been determined for the set using an incident meter, the light level of the Combi-1 is matched to the scene level using a spot meter.  A few frames of the Combi-1 pattern are recorded at the head or tail of each camera roll or scene change.

By reading and noting the black-and-white steps of the Combi-1's gray scale using the spot meter, the cinematographer now has control over highlight and shadow information in the final transfer.  Any scene element having a spot meter reading greater than the white step, or less than the black, will be out of the normal video range.  Because the OSG is shot using the same type of illumination as the set, the issue of color temperature is addressed immediately.  Once the telecine/scanner is properly aligned using Combi-1's grayscale and color bar data as recorded on the film, image reproduction is standardized and repeatable.

To extend the process further, the cinematographer can establish a personal reference standard by carefully shooting the Combi-1 OSG on a roll of film, which is then processed normally.  A section of this exposed and processed original negative can be supplied to the transfer house to be used as a telecine reference and comparison standard.  This enables variations in film stocks, lab processing, color temperature, and exposure to be monitored.  DSC also supplies a negative alignment film produced and processed under controlled conditions.

The Ambi/Combi OSG combination also facilitates control of scenes that are not average.  By setting up a standard alignment, a cinematographer can deviate a set amount to produce a consistent, repeatable special "look."  In conjunction with the spot meter, scene lighting levels and contrast may be set to a desired level within the confines of the video system.  Color balance may be similarly controlled enabling consistent image reproduction even when transferred at different facilities.

A cinematographer with a firm understanding of the operating philosophy of the Ambi/Combi system has a powerful tool with which to control image quality.  The look that was so important to establish at the beginning of a production may now be reproduced on a daily basis and becomes independent of the post-production facility.  Exposure of the original negative can now be carefully controlled, assuring that the original negative contains adequate density range to enable transfers that satisfy both domestic and foreign markets.

Practical Test Results

Running a series of tests as outlined above proved the practicality of the system.  These were evaluated using one of CBC Toronto's Rank URSA Scanners (see Fig. 10, 16mm Combi-1 negative on a CBC URSA).

Based on the satisfactory performance achieved, five additional tests were shot, including simulated LAD patches, and sent to lab/transfer houses across North America for processing and transfer to tape. 

On return, the "simulated LAD patches" were measured and compared.  Densitometer measurements and subjective evaluation of the five tests on a CBC URSA showed minimal differences in processing between the five labs.  However, there were major differences in the transfers made by the different labs.

These tests confirmed that the reproduction of film materials is still plagued with difficulties brought on by the subjective nature of the work and the lack of industry standards.  These results come as no surprise, because this type of variation is well-known and has been documented over a number of years.  These results come as no surprise to those using multiple transfer houses.

An interesting side note is that one of the participants, a major lab/transfer house, asked to keep a few frames of the Combi-1 negative because it provided an excellent base setting for their scanners.  They reported that even different film stocks reproduced better when the scanner was aligned to the Combi-1 negative test film.  Clearly these tests demonstrate the industry's increasing need for higher standards.

Summary

The Ambi/Combi System was developed in collaboration between DSC Laboratories and the Canadian Broadcasting Corp. and is based on sound engineering and scientific principles.  Embracing international standards, it uses conventional waveform and vectorscope instrumentation.

Combi OSGs are multifunctional rear-lit test targets.  Useful not only as camera and telecine alignment tools, OSGs are also valuable in film and video production as objective reference standards.

The Ambi/Combi System provides both the precision demanded by engineers with the ease of use essential in production.  It performs the important function of enabling image quality to be precisely and consistently quantified.  The signal information is presented in a familiar manner, readily embraced by all working in the television industry.

The Authors

Fred Benedikt holds a B.Sc. Honours Physics from Concordia University.  His experience in broadcast engineering activities is derived from his work in several CBC engineering departments as a technical specialist for television cameras, picture monitors, and video signal distribution.

Benedikt developed the Combi-1 test pattern for the CBC, with the manufacturer DSC Laboratories, to facilitate television camera equipment evaluation, alignment and color matching.  He was a consultant for the Advanced Television Evaluation Laboratory (ATEL) for the viewing, alignment, and operation used to assess the various advanced television proponents.  Presently, his activities and responsibilities deal with the implementation of radio and television studio projects.

Benedikt is a Fellow of SMPTE.  He was a member of several SMPTE Engineering Committees involved with colorimetry, telecine, color picture monitors and viewing conditions, and hybrid television systems.  He is an active member of the Toronto Section Board of Managers and was recently elected Chairman for the Section.

David Corley was educated at Kings Canterbury, and Christ Church Cathedral School, Oxford.  He immigrated to Canada in 1950 and worked in many aspects of motion picture and television production.  Because of his interest in the technical aspects of image quality, Corley and his wife established DSC Laboratories in 1962, to provide precision products and services to the television and AV communities.  His patents include seamless (soft-edged) panorama technology, dichroic additive printer, electronic shutter, ambient light illuminator, and Optical Signal Generators.

Corley has been a SMPTE member since 1966, and a Fellow since 1986.  Active in SMPTE engineering committees, he was honored with the Fuji Gold Medal in 1994.  Currently, as SMPTE Director of Education and Canadian Governor, he is dedicated to the SMPTE Student Chapter Program and providing ongoing education through the Internet for Society members, budding engineers, and technologists.

Bob Ross holds a Bachelor of Technology - Photographic Arts from Ryerson Polytechnical University.  He was with the film operations of the CBC at its engineering headquarters in Montreal and the Toronto Production Centre until his retirement in 1997.  For the last seven years he held the position of supervisor of quality control EFP/film and was responsible for the technical quality of film program materials produced for the CBC by the Toronto Production Centre.  Of particular interest was the interface between film and video and the maximization of quality between during the transition from one medium to the other.

Just prior to his retirement, Ross was nominated for the CBC's Micam Award and was recognized for his work in developing and implementing methods to apply the new generation of HFE film solvents to the CBC's film operation.  He is currently partner in a consulting firm Perfect Pictures Group, and specializes in the issues surrounding image reproduction and quality in film and video.

(Reprinted with the permission of the SMPTE Journal)

Go to DSC Home Page