Near midnight of April 14th, 1912, the R.M.S. Titanic struck an iceberg at 26mph, about 375 miles southeast of Newfoundland, on her maiden voyage. Three hours later, she sank to the ocean floor, two and one half miles below, killing 1,522 people. Seventy nine years later, we descended below on June 30th, '91 to take the first 1570 LF (Large Format) movies of the wreckage, which lies split into two sections, spaced about 600 meters apart. We filmed from advanced Russian submersibles, MIR1 and MIR2, as part of a joint expedition aboard the Russian research vessel Akademik Mystislav Keldysh.
These three man Mir submersibles are well suited for deep water cinematography in the 15 perforation 70mm format originated by Imax Corporation, with film frames ten times the size of conventional 35mm movies. We made 17 dives, averaging about fourteen hours each, to shoot the brightest, clearest movies ever obtained in the deep sea for an Imax production called Titanica, directed by Stephen Low. This paper will explain the technical challenges in filming successfully from submersibles.
A key consideration for deep sea photography is lighting - at the bottom of the ocean it is pitch black. On free swimming submersibles like MIR, battery power is precious, so luminous efficiency is important. 1570 cinematography requires wide fields of view, and hence broad lighting coverage. For these reasons we decided in Dec 90 that we needed to develop a deep sea version of the High Intensity Discharge Lamps commonly known as HMI™, which is a trademark of Osram Corporation.
HMI lights have Several Advantages over conventional Quartz Iodide lamps traditionally used for deep sea photography:
First, HMI's have an appropriate spectrum for sea water penetration - high in blues and greens, with an average "daylight" colour temperature around 5600 degrees K. This means their bluish light is not absorbed as fast as the relatively reddish light from incandescents.
Secondly, the bluer HMI spectrum better matches the sensitivity of the fast Eastman 5296 EXR colour negative film we used.
Thirdly, HMI's are very efficient - they put out less IR and much more visible light - typically 80 to 100 lumens per watt compared to 15 to 30 lumens per watt for incandescents.
Imax Corporation wanted a safe, dc-powered version of an HMI lamp that would operate reliably at Titanic depths, where the pressures approach 6000 psi. An implosion of a lamp or housing at such depths would create extreme shock waves which could jeopardize the safety of the sub. We contracted Chris Nicholson of Deep Sea Systems International to work with lighting experts Mark Olsson of Deep Sea Power and Light, and Richard Mula of HydroImage Inc. to develop & certify these lights. Olsson worked with Imax physicist Kamal Hassan to calculate the illumination levels, lamp spectral type and angular coverage required for our widescreen format. Olsson and Mula used modified HMI 1200W/SE lamps by Osram, which produce 110,000 lumens at the source. To minimize implosion volume these lamps were repackaged into a minimum volume quartz pressure envelope shaped like a test tube..
Olsson and Mula worked with Coors Ceramics and Cinemills Corp on a crash schedule to build insulating glass-ceramic pressure housings for the 45 Kilovolt LPA lamphead ignitors and 1.2 KW electronic ballasts, which converted the sub's 120 Volts dc battery power into the square wave AC required by the lights, for flicker-free filming. The ballast housings were sleeved with plastic for impact resistance and filled with oil for safety.
The lamps were remotely controlled via oil-filled junction boxes from inside the sub by a custom switching panel built by Nicholson. High efficiency, sharp cutoff wet reflectors were designed by Olsson for wide and medium flood and spot lighting. Each sub was equipped with four 1.2 KW HMI's and four 1.0 KW Quartz Iodide lamps which drew almost 10 kwatts at 120 Vdc. With both subs lighting the Titanic, we had a total effective lighting capacity equivalent to using almost 150,000 watts of incandescent light.
Special through-hull connectors by SEACON/Brantner and Associates were certified to 14,000 psi to ALVIN specs by Barry Waldon at Woods Hole Oceanographic Institution, to ensure crew safety. During the Titanic expedition, Chris Nicholson and Jeff Ledda installed the lights, ballasts, switchboxes, penetrators and control panels into the two MIR submersibles, and also replaced lights damaged during filming.
The HMI lights were carried on hydraulically adjustable swing-out, lighting booms on each side of the MIR subs, to avoid "backscatter" from particulate matter, which produces an effect similar to driving in a snow storm. Getting the lights out almost 10 feet away from the camera lens axis significantly reduces this problem. By using two subs, and some switching and aiming of the lights, we were able to achieve a variety of interesting cross and back lighting effects, impossible to do with a single sub.
HMI's do have a Few Disadvantages: 1 to 3 minute warm-up times, susceptibility to blowouts from hot restrikes, and the added weight and cost of the inverter/ballasts. Additional syntactic foam had to be added to the Mir subs to compensate for the weight of lights and camera. Also, our ceramic housings exhibited some thermal problems which meant the HMI's could only be run for 20 minutes at a time, before needing shutdown for cooling.
However, these are the brightest lights ever used in the deep sea, on any submersible, and dwarf those used on previous Titanic expeditions. We found that we could often see out 50 to 75 feet in clear sea water on the wreck and capture a wide area of acceptable exposure. In contrast, up to 8 kilowatts of QI lights used in early tests did not provide useful exposures much beyond 7 to 10 feet. These new Deep Sea HMI Lights resulted in the largest, clearest motion pictures ever obtained of this mammoth wreck.
Filming from submersibles also presents some unique optical challenges. The central port of the Mir sub is a flat, 7 1/8th inch thick conical acrylic plastic port, designed to withstand 6 kilometre operating depths. Fortunately, it is of wide diameter - 200 mm inside, 480 mm outside - much larger than the ALVIN or NAUTILE submersibles, and hence can better accommodate 1570 fields of view as wide as our 40mm lens with no vignetting. There are also two smaller ports of 120 mm diameter angled off about 30 degrees to each side, which we used primarily for the sub pilot and camera assistant to steer and focus by, respectively.
There are several optical compromises associated with filming through flat ports. The most significant is the loss of angular field of view, which necessitates being further away from the underwater subject, which inevitably affects both illumination and clarity, due to backscatter from silt in the water. This can also be thought of as an increase in magnification of about 33%, which leads to the second problem - focus shifts. Distances are compressed by about 25%, such that a Titanic feature 10 feet away in fact will appear to be only 7 1/2 feet away to the lens. Thirdly, flat ports introduce some optical aberrations, consisting of mostly lateral colour, astigmatism and distortion. Some image quality fall-off was predicted and observed in the corners, but is generally acceptable.
We used a variety of Zeiss prime lenses, but mainly the 40, fast 50 and 30mm fisheye. In the future, the most significant optical improvement for general filming from subs like MIR would be to design faster wide angle rectilinear lenses in the 32 to 40mm focal length, f/1.4 to f/3.5 range, if possible.
Another key technical challenge in filming from the MIR submersibles was devising lightweight, adjustable but rock-steady camera mounts that were easy to work with, both for pilot and camera crew, and yet strong and safe. Imagine the nightmare scenario of our 95 lb IMAX® camera impacting a plastic viewport under 6000 psi pressures. Most of the basic rules and procedures of cinematography had to be addressed. Camera adjustability, mag loading, focus & iris control, changing lenses, and a comfortable operating position for the cameraman are obvious essentials.
In February 91 we built a lightweight wooden mockup of our IW5A camera and made fit checks in a fullscale MIR sub trainer in Moscow. Our goal was to discover if it was feasible to film from the large 200mm central viewport, which is usually used by the pilot to fly the sub. We also needed to determine how closely we could position our camera lenses to the viewports, to obtain maximum fields of view, without vignetting. We resolved many minor details like whether or not our camera would fit through the hatch, and if there was room for three men and a camera with 80 pounds of film inside a 6 1/2 foot ball.!
We obtained the pilot's tentative consent to block his view of the central port and associated controls and displays. The Russians also agreed to relocate the main control stick, and two manipulator joystick controls to accommodate our camera. They decided to remove one sub manipulator arm to save weight, but the other was kept for science experiments and safety reasons. This later proved to be a wise decision. In addition to these changes to the subs, Imax had to design new tapered bases for our cameras to help clear the pilot's control panel. We agreed that the mount had to allow quick removal of the camera in an emergency to give the pilot fast access to his normal viewport and controls.
Next we visited Rauma Oy Oceanics in Tampere, Finland (formerly Rauma Repola) who designed and built the MIR subs, to gather full technical specs and mechanical and electrical drawings. Final design and construction of the camera mounts began at Imax by building a simplified full scale wooden mock up of the sub windows and controls. Aluminum rings simulated the exact size and position of the ports. Extruded aluminum square tubing with semi adjustable fittings made by xxx created a strong but light weight frame to support our heavy camera over the delicate instruments. We designed for a peak loading of 6G's to withstand rough sea launch and recovery, or subsea collisions while keeping the lens within one eighth of an inch of the window, at near focus position.
The entire camera mount pivoted about a point located near the front of the lens, to avoid vignetting while still allowing limited pan and tilt. We designed a semipermanent expanding viewport ring with double O rings and a rain gutter, to divert hull condensation around the lens and window. This supported a small video camera, antireflection lens donut and a blower fan to reduce window condensation. We also designed framework accepting the camera mount for the smaller 120mm starboard viewport, as backup. The complete camera support assembly could be quickly removed by taking out two "t" pins and loosening two screws.
To change lenses or load the camera, the operator loosened one knob and slid the camera rearward on a dovetail, which gave easy access to the movement and magazines. These MIR sub IMAX® camera mounts proved to be quite workable under difficult real-world filming conditions, and strong and secure enough that we could descend into the depths without worrying about the camera coming loose in a collision. The MIR subs in essence became a twenty ton moving 3D dolly for our camera - a smooth and steady filming platform!
Video & Electronics
Although our primary mission was to shoot large format 1570 film, two submersible video systems proved very useful. The first was our "steering video", which consisted of a tiny 1" x 3" Panasonic colour CCD camera with wideangle 3mm lens, recorded on a Sony 8mm VCR, and displayed on small Realistic & Sony colour monitors. Because our large Imax camera completely blocked the viewport normally used by the pilot, this 112 degree video view helped compensate by giving us a close look out the central port. We found this so useful we often left it running for most of the dive.
The second system was a colour CCD video tap built into our Imax camera viewfinder, which recorded exactly what we shot, at a lower quality level. Other electronic equipment consisted of a custom built control panel which directed power to our Imax camera, heater, window fan and all the video equipment, from three 10 amp 24 volt DC sub payload breakers. Our IMAX IW5A camera draws about 11 amps at 28 volts, 20 amps surge, plus 1-2 amps of heater power.
There are several electrical concerns involved when filming from submersibles. First, electrical isolation of the submersible chassis must be maintained. If it isn't isolated, and there is any leakage current or high resistance short to the sub chassis, the hull becomes an electrode and suffers electrolysis in the salt water. Isolation is also important for crew safety, to reduce chances of electrical shorts, fire or catastrophic penetrator failure. Our IW5A camera body is galvanically isolated from both power leads. In addition we isolated the camera mount itself with insulating washers to give additional safety.
It is good practice to have circuit breakers, relays or switches that disconnect both the negative and positive power leads of all equipment, to ensure complete and quick isolation of electrical problems in the event of a ground fault or equipment failure. All VCRS, monitors and control panels had to be mounted in insulating chassis. All cables such as BNC type video with grounded shell connectors must be wrapped with insulating tape so they cannot conduct to the sub.
Another concern is materials. Since the crew is in an enclosed capsule toxic gases should not be produced during equipment operation or failure. Three people must rebreath only 125 cubic feet of reconditioned air during a 14 to 18 hour dive. Hence power lines must have insulation which does not offgas when operating at elevated temperatures. We used Raychem's Spec 55 wire which has insulation rated at 200 degrees C.
A third concern is moisture. Inside the submersible water from the breath of the occupants condenses & collects on the cold hull and eventually drips or runs down. Electronic equipment should be protected with drip guards or water resistant cases. Whenever possible, cables connected to the equipment should be dressed angled downwards to prevent water droplets from running into the equipment. Water inside electronic equipment is never a success story.
The water outside the sub at ocean bottom is only 1-2 degrees C above freezing, so the sub hull, crew and equipment get very cold after many hours of submergence, unless heaters are used. Equipment supposedly rated for these temperatures does not always operate reliably near freezing, so cold tests of all submersible equipment are advisable before installation. For example, one video monitor failed during an early cold test at Imax.
Our top priority in all equipment design for Titanic filming was safety. Fire, implosion, collision and entrapment are key safety concerns in deep ocean submersible work. We had to balance cinematic and ergonomic issues with safety factors to ensure that the crew could film adequately without compromising life support. Avoiding fires is largely a matter of proper electronic design, with redundant circuit breakers, proper insulation and fusing. With 70 amps of 120 volts dc of lighting power running through our hull penetrators, inadequate fusing could conceivably lead to penetrator meltdown in a short circuit situation, which would turn a leaky penetrator into a nice water-knife at 6000 psi.
Avoiding implosions is largely a matter of minimizing implodable volumes like lamps, using oil filled electronic housings and keeping cameras and lamps safely away from the plastic viewports, with shock resistant mounts. The roughest motion typically occurs during launch and recovery of the subs in high seas. We depended heavily on our weather lady to predict the variable moods of the North Atlantic.
By far the most dangerous hazard in filming the Titanic was entrapment. The pilot has such an oblique and mechanically restricted field of view that constant dialogue was necessary between the camera assistant on the right side and the pilot on the left, neither of whom could see what the other person was seeing. The extended lighting booms at each side were a constant concern, as well as the entire overhead and rear portion of the subs, which could not be seen at all. The Titanic wreckage has many overhanging hazards such as cables and broken railings, and the fickle ocean currents have a habit of sweeping the sub into difficult situations.
We always dove in pairs, with both subs, so that there would be some hope of disengaging each other from unseen obstacles, but on two of my dives, we got momentarily caught on our lighting boom from obstacles above after the other sub had headed up with low batteries. The latter time, we had to break our starboard lighting boom to escape from a hole in the side of the Titanic. This is when manipulator arms come in handy. Our other two dive teams encountered similar problems. In future, it would be safer to have more outside video cameras spread around so we could more safely maneuver in difficult situations that could not be seen from our viewports.
TAG site Filming Tests
Test dives are invaluable learning experiences prior to filming production footage from submersibles. On May 30th and June 3rd, 1991 we dove to 3800 meters at the Trans Atlantic Geothermal or TAG site in the mid-Atlantic rift, to test all camera systems except the HMI lights, which were still being built. We wanted to try all lens and camera mount configurations, and work out the logistics of working efficiently with 340 pounds of Imax equipment and film inside the MIR sub. We quickly learned several valuable lessons:
First, was the extreme difficulty of finding, let alone positioning a second submersible at the bottom of the ocean, due to the difficulties of navigation, currents and communication, and being lost in clouds of silt disturbed by the sub thrusters. Lighting setups and tracking shots took hours to coordinate.
Second, it was extremely difficult and time consuming to achieve pre-plannned objectives, and many desired shots of specific TAG features were not achieved - it is far more time efficient to shoot targets of opportunity from a single sub when water conditions are clear, and to face the current if possible, so that silt raised by the submersible blows away.
Third, it is unwise to depend on focusing through the reflex camera viewfinder alone, in such dim, low contrast light situations with wide angle lenses. Almost 80 % of the test shots were slightly out of focus. The extremely shallow depth of field forced by shooting wide open without the HMI lights didn't help. We decided much higher focus accuracy could be achieved by using a camera assistant to pull focus by estimating apparent distance through the starboard viewport and setting the lens appropriately. This system worked well for the actual Titanic shooting - 80% of that was in good focus.
Fourth, use the widest lens you can for the shot, to minimize objectionable backscatter, lighting falloff, and sub motion and to maximize depth of field, exposure and angles of view already limited by the flat port. We learned that even a twenty ton submersible could move faster than desirable and create strobing on the huge IMAX screen. What looked perfectly acceptable in video and previous Vistavision proved to be too violent a motion in 1570.
Fifth, all attempts at undercranking the camera during tracking shots to gain more light by running at 6 or 12 frames per second were quite unsuccessful because of excessive, jerky motion. Even at 24 frames per second, the sub pilot must just inch along at about one eighth to one half of a knot maximum to avoid strobing in our wide field format.
Viewing this TAG site footage at Ontario Place in Toronto prior to filming Titanic proved to be invaluable feedback. It made us all suitably paranoid about achieving adequate depth of field and critical focus. It also proved how badly we needed the extra lighting of the HMI's, installed later in Bermuda.
Cinematic Challenges
Our last filming dive to Titanic was on July 15th, 1991. We had captured about 39,000 feet of Titanic footage between three sub crews from our seven Imax dual sub dives. With the hindsight gained from viewing the submersible footage, we asked what the other members of our sub film crews felt was their biggest technical challenge in filming the Titanic from the MIR submersibles.
For Director Stephen Low it was to use the widest angle lens to give the longest shot with the least water possible between the subject and the camera, and to side and back light the subject to avoid backscatter from particulate matter.
For cameraman Paul Mockler and assistant Per Inge Schei, focus was the biggest challenge, particularly when filming artifacts in the debris field with longer lenses. They could not touch down with the sub, or billows of sediment would arise, so they had to hover and maintain focus while drifting in currents up to one knot or so.
For experienced submersible cameraman Ralph White, lighting a sufficient area of the wreck using both subs and setting up complicated long tracking shots of the other sub within the limited twenty minute burn time of the HMI lights was the largest challenge. Alternating sets of HMI lights alleviated this difficulty somewhat, but the limited three minute run time of our 1570 film magazines, and time-consuming reloads in the cramped cold sub, were also problems for our dive crews.
All of us found it difficult to establish a sense of distance and scale, particularly with the gargantuan proportions of certain objects on the Titanic. This was fairly easy with familiar objects like windows and port holes, but the size and distance of engine cylinder heads, anchors, masts and propellers were very difficult to judge. The Titanic truly is titanic!
In hindsight, all of us regret not having sufficient dive time to set up more wide angle tracking shots of the other sub maneuvering around the wreck, which are beautiful but extremely difficult to coordinate. Even with our bright lights and fast film, exposures at any distance are still limited, so perhaps slight undercranking with very slow sub motion or better lighting positioning with pan/tilt capabilities would help in the future.
In summary, there are some limitations to filming from submersibles and also some big advantages.
The biggest disadvantage is the flat port - the 30 to 40% loss of field of view with our widest lenses, compared to a fisheye lens with proper dome port. Another disadvantage is the danger and discomfort of long dives in cold and cramped submersibles. Difficulty in focusing is a third limitation.
The biggest advantage to filming from submersibles is the ability to change film, which is impossible with deep sea underwater housings or ROV's. We carried up to nine rolls of film per dive, for twenty seven minutes of Imax. Another big advantage is lots of power for lighting - the large 100 Kilowatt-hour battery capacity of the Mir subs could drive our HMI's for hours of filming, and dives up to 18 hours duration allow ample bottom time. The submersible itself is a pretty good camera mount, in terms of having considerable inertia which makes for smooth camera motion.
Submersibles open up new areas of the deep sea that are impossible to film with divers or ROV's. By using two MIR subs and off-axis HMI lighting, for the first time we achieved a satisfying amount of light in the deep sea. We believe we achieved the best motion picture documentary footage recorded of the Titanic to date. We hope you enjoy seeing Titanica - Stephen Low's movie.
(Originally presented at Cinesphere Ontario Place for SMPTE Toronto on Nov 8/92 by Gord Harris)
Please click here to read other papers.
Copyright © 2003, Gord Harris, go-R&D Consulting. IMAX®, IMAX® Dome, IMAX® 3D, IMAX® 3DDome, IMAX®
Magic Carpet, IMAX® Simulator Ride IMAX® HD and The IMAX® Experience are trademarks of IMAX Corporation, Mississauga, Canada.