I Install Home Security Cameras. These are my SCARIEST Stories.

My name is Chet, and I install home security cameras for a living. I grew up around the work because my father started the family business before most people wanted cameras in their homes. Back then, his customers were usually gas stations, small warehouses, and convenience stores. The cameras were heavy, picture was poor, and every wire had to run back to a recorder that used tapes. He would sometimes spend an entire day getting one clear view of a cash register. As a kid, I went with him whenever school was out. At first, my job was to carry boxes and hand him tools. By the time I was 16, I could terminate cable, mount a camera without cracking the siding, and tell the difference between a bad power supply and [music] a damaged connection. My father taught me everything he knew. He showed me how to plan an installation before drilling the first hole, how to keep water from following a cable into a wall, and how to explain a system without making the customer feel stupid. More importantly, he taught me not to guess. If a camera failed, there was a reason. If a recorder lost 10 minutes, there was a reason. You check the power, the cable, the storage drive, the settings, and the network in that order. Somewhere in that chain, you'd find the problem. He's retired now. His knees stopped liking ladders years before he admitted it, and eventually my mother convinced him that supervising me over the phone did not count as retirement. I kept the family business going. The equipment changed from tapes to digital recorders, then to network systems and cameras that could send a clear image to a phone on the other side of the world. The work itself stayed familiar. I still spend most of my days in attics, basements, garages, and backyards trying to give people a clean view of the places that worried them. Most strange camera footage has a simple explanation. A spider near an infrared light can look as large as a person. Rain on a lens can make a face out of nothing. A weak wireless connection can freeze one part of an image while the clock continues to move. Motion alerts are even less reliable. Tree branches, moths, headlights, steam, and changes in sunlight can all make a system think someone is standing where nobody is. I have explained thousands of recordings that frightened homeowners. I've also seen five systems that I could not explain. They were installed years apart in different houses using different brands of equipment. None of the customers knew one another. The only connection was that I worked on every system, and each one showed something that was not physically there. Story one, the smiling man. Albert lived in a one-story house with a small fenced backyard. A concrete patio covered the first 12 ft behind the house, and the rest was grass bordered by wooden fence. There was a storage shed in the left corner, a metal grill near the patio door, and three thin trees along the rear property line. Nothing about the house stood out. It was the kind of place where I could usually decide every camera position before unloading the ladder. He wanted four exterior cameras. One would cover the driveway, one would face the front steps, another would watch the narrow side yard, and the last would cover the backyard. Albert had not been robbed. His rear motion light had started turning on several times each week, usually between midnight and 3:00 in the morning. He'd looked outside more than once and never found a person or an animal. After it happened for the fourth week in a row, he decided he wanted a recording instead of another empty view through the window. I installed the backyard camera beneath the roof overhang several feet to the right of the patio door. It was a fixed network camera with a wide lens powered through the same cable that carried the video. From that position, it showed the whole patio, most of the grass, the shed, and the rear fence. I sealed the mounting holes, dressed the cable against the siding, and went inside to adjust the picture from my service monitor. The living room opened into the kitchen, and the patio door was beside the table. I stood where I could see both the monitor and most of the backyard through the glass. Albert remained near the counter while I adjusted the camera angle. The first picture looked normal. The grill was in the lower left corner, the shed occupied the far right, and the three trees were lined up along the fence. And then I noticed a man beside the middle tree. He stood just inside Albert's fence facing the house. His clothes were ordinary enough that I could not have described them afterward beyond dark pants and a light shirt. He was not leaning against the tree or trying to hide. His hands rested at his sides, and he was smiling directly at the camera. I raised my eyes from the monitor, and I looked through the patio door. The middle tree was there. The patch of grass around it was empty. I could see the entire distance between the house and the fence, and there was nowhere for a person to stand without being visible. When I looked down again, the man was no longer beside the tree. He was several feet closer, standing near a low spot in the lawn. His position had changed, but everything else was identical. His arms still hung at his sides, his feet were together, the smile had not changed. I checked the time in the corner of the picture. The seconds were advancing normally. A breeze moved the leaves above him, and the movement matched what I could see through the grass. A dry leaf crossed the patio on the monitor at the same moment I watched it pass outside. This was not an old recording, and it was not a frozen image. Albert came over when he saw me stop working. Is that one of your workers? he asked. I work alone, I said. Then who is that? I didn't answer because I had no useful answer to give him. I asked Albert to keep his eyes on the yard while I watched the screen. For nearly a minute, nothing changed. The man remained in the low patch of grass. Albert could see the place where he should have been and confirmed that the lawn was empty. I looked through the glass again. There was no one near the trees, the shed, or either section of fence. The gate beside the house was closed. When I returned my attention to the monitor, the man was standing beside the grill. That put him at the edge of the patio, less than 15 ft from us. The grill hid part of his lower body, but his face was clear enough to see that he was still smiling. Chad, he's beside the grill now. Albert said. I see him, I replied. There's nobody on the patio, Albert continued. I opened the camera settings and checked the stream information. The device serial number matched the camera outside. The live bit rate changed as the leaves moved. I switched from the high resolution stream to the lower one used for phone viewing. The picture is softened. But the man remained beside the grill. I restarted the viewing application without restarting the camera. When the image returned, he was still there. For several minutes, Albert watched the yard while I watched the monitor. The figure did not move. I began to wonder whether the change only happened when neither of us looked directly at the screen. So, I set the monitor on the kitchen table and positioned it beside the patio door. That allowed us to see the screen and yard without turning completely away from either one. The arrangement should have removed any chance for another change. Instead, it showed us exactly how limited the camera's role was. I glanced at the cable connection for no more than a second. Albert looked toward me at the same time. When we returned our attention to the picture, the man was standing at the near edge of the patio. There had been no movement across the feed. He didn't walk from the grill. One frame showed him beside it, the next one showed him closer to the house. On the actual patio, there was only an empty chair and a coiled garden hose. I stopped trying to identify a technical fault and I concentrated on ending the feed. The camera cable ran to a small network switch I placed near the recorder. I kept my hand on the connection while I watched the screen. The man stayed at the patio edge. I looked through the glass one last time and I saw nothing. When I looked back at the monitor, most of him was gone from the frame. The top of his head and his eyes were visible at the bottom of the picture. Based on the camera angle, he had to be standing directly beneath it against the rear wall of the house. The smile was still visible, cut in half by the lower edge of the screen. I pulled the camera cable from the switch. The feed went black immediately. Albert stayed inside while I checked the backyard through every window facing it. The patio, lawn, and fence remained empty. I waited several minutes before opening the door, and even then I stayed on the threshold. There was no mark beneath the camera, no movement near the gate, and no sign that anyone had crossed the damp grass. I removed the rear camera without reconnecting it. Albert kept the other three, and I gave him a credit for the one I took away. I tested it later at a workshop using a different cable, switch, recorder, and monitor. It showed an empty room for 6 hours. The next morning, I pointed it toward the parking lot, and it worked normally there, too. Albert never asked me to reinstall it. I checked with him several months later when I serviced the front camera, and he told me the rear motion light had stopped coming on after the camera was removed. I never saw the smiling man again. The wet footprints Lisa's house had a back door that opened into a square kitchen with pale tile flooring. The door was newer than the house and had a deadbolt, a reinforced frame, and a contact sensor connected to a basic alarm. Twice in the same month, Lisa had returned from work and found it open by several inches. The alarm history showed no entry. The frame had no damage, and nothing inside had been disturbed. She wanted one camera mounted in the upper corner of the kitchen. From there, it could cover the back door, most of the floor, and a short hall leading toward the rest of the house. The installation was simple. I ran a cable through the cabinet space above the refrigerator, and I placed the compact dome camera near the ceiling. Before closing the dome, I cleaned the lens and the inside of the clear cover. A fingerprint inside that cover would become a bright blur when the infrared lights came on, so I was careful not to touch it. Lisa stood near the kitchen table while I connected the camera to a small recorder. The first image was clean. I walked through the view to check the motion zones, then open and close the back door several times while Lisa confirmed that the alerts reached her phone. The contact sensor worked every time. There was no looseness in the latch, and the deadbolt extended fully into the frame. I was adjusting the detection area when a dark mark appeared on the tile, just inside the door. It was shaped like the front half of a bare foot. At first, I thought something on the floor had changed the camera's exposure. I looked directly at the tile. It was clean. The mark remained on the monitor. A second print appeared ahead of it. This one showing a heel and five toes. There was no fade or distortion before it formed. One moment [music] the tile was blank. The next moment a wet footprint was resting on it. "Are those footprints?" Lisa asked. "They're on the picture, but they're not on your floor," I I We stood on opposite sides of the kitchen and watched. A third footprint appeared, then a fourth. They formed a walking pattern from the back door through the center of the room. The prints were darker around the edges and slightly reflective in the camera image. They looked fresh enough that I expected to smell wet soil or feel moisture in the air. But the room remained dry. I moved closer to the marks while watching myself on the monitor. My boots crossed directly over the first two prints in the picture. Nothing transferred to the soles. And the prints did not smear. I knelt and touched the tile where the third one appeared. The floor felt cool and smooth with no water on my fingertips. On the monitor, my hand covered part of the wet mark, proving the camera was showing the current room and not playing a separate recording. Another print appeared behind my hand. I stood and moved away from the trail. The prints continued toward the kitchen table. There was no body above them, no interruption in the background, and no bend in the light that suggested a transparent person. The camera simply showed wet feet pressing against a dry floor. I switched to another monitor. The prints were there. I replaced the short network cables between the recorder and the camera. They remained. I bypassed the recorder and connected the camera directly to my service unit so that no stored footage or software processing sat between the lens and the screen. And the same trail appeared immediately. By then, the first prints had begun to look less glossy as if they were drying. New ones continued forming. They curved around the table, passed within a foot of Lisa, and turned toward the camera. Lisa moved into the hall. I stayed beside the counter because the service monitor and camera connections were there. Neither of us stepped into the line of the prints again. Chad, it's coming this way. Lisa said. Stay near the hall, I told her. The trail stopped beneath the camera. Because of the lens's angle, the final two marks were partly hidden by the lower edge of the cabinet. I could still tell that the invisible walker had approached the wall directly below the dome. Nothing appeared on the rest of the image. The kitchen stayed quiet, and the back door remained locked. The alarm sensor stayed closed. And then moisture formed across the center of the picture. It began as several round spots, the kind I would expect if someone had breathed against cold glass. More spots appeared around them until the center of the view looked damp. The camera was mounted near the ceiling in a room with normal temperature and humidity. Condensation could not form that quickly. And if it had, I should have seen it on the outside of the dome. I climbed one step of my ladder and examined the cover. From below, the plastic looked dry. On the monitor, water was spreading across it. A white mark appeared in the moisture. I could make out the heel of a hand, then the shape of a palm and five fingers pressed flat against the inside of the dome. The mark covered most of the picture. It was not a print left during installation. I'd cleaned the cover less than 20 minutes earlier, and the hand was far larger than any smudge I could have missed. Water gathered along the lowest edge of the image and ran in small lines toward the bottom. I disconnected the camera before touching it. The screen went black. When I removed the clear cover, several drops of water fell onto my ladder. The outside was dry. The inside was wet enough that water had collected around the rubber seal, and a complete handprint was visible against the plastic. There was no opening large enough for a hand to reach inside. The dome had been locked to its base, and I'd been in the kitchen from the moment I closed it. Lisa had never gone near the ladder. I checked the camera body for a leaking pipe or cable path, but the wall and cabinet space were dry. The cable came down from an interior ceiling with no plumbing above that corner. I put the wet cover and camera into separate plastic bags and removed the recorder. Lisa chose to keep the alarm sensor and add a second mechanical lock to the back door, but she did not want another camera. Before leaving, I wiped the places where the footprints had appeared on the monitor, even though the floor had never become wet. The image was gone, and there was nothing physical to clean. Well, I kept the camera disconnected for several days. When I opened the bag, the handprint had dried into a faint mineral outline. I cleaned it again and tested the camera without the dome. The image remained normal. I later installed that camera in a detached garage after replacing its cover, and it worked for years without another strange picture. Lisa actually called me once, about 6 months after the installation. She needed a new battery for the alarm panel. And while I was there, I asked about the back door. It had not opened by itself again. There'd been no wet marks in the kitchen and no problems with the sensor. Whatever had crossed that floor had stopped appearing when there was no camera left to show it. The door that wasn't there. Now, Jean's basement was unfinished but organized. The furnace and water heater stood near the stairs, metal shelves lined one wall, and plastic storage bins were stacked along another. The rear foundation wall was bare poured concrete except for a small window near the ceiling. Jean had lived there for 12 years, knew every crack because he had sealed and painted the concrete himself. His alarm had started reporting motion in the basement while he was away from home. The sensor was old, so neither of us considered the alerts mysterious at first. Temperature changes, insects inside the house, and failing component can all trigger an aging motion detector. Jean asked me to replace it and add two cameras so we could check the basement without driving home. I mounted one camera near the stairs and aimed it across the furnace area. The second went in the opposite corner facing the shelves and rear wall. Both were fixed cameras with overlapping views. If someone entered through the basement window or came down the stairs, at least one camera would show them clearly. The camera near the stairs looked normal. The second camera showed a door in the foundation wall. It was narrow, made of dark wood, and positioned between two sections of shelving. The top of the frame sat below the basement window. There was a round metal knob on the right side. Nothing about the door looked unusual until I raised my eyes the monitor. The actual wall between the shelves was solid concrete. "That door isn't there." Gene said. "I know." I replied. We moved the shelves far enough apart to inspect the wall. There was no filled opening, no wooden framing, and no change in the concrete texture. Gene's paint covered the entire surface in one continuous layer. I measured the distance from the corner to the window and compared it with the camera view. The door occupied the exact stretch of wall we were examining. I removed the camera and held it in my hand. As I changed its angle, the door shifted naturally within the picture, just as a physical object would. When I pointed the lens toward the floor, it disappeared. When I raised it toward the wall, it returned. This was not an image burned into the sensor. I carried the first camera from the furnace area and aimed it at the same wall. The door appeared on that feed, too. The cameras were different models made 2 years apart. They used separate cables and separate ports on the recorder. A fault common to both would have affected the whole image, not added the same detailed object to one precise location. Well, Gene and I spent nearly an hour testing the wall and equipment. I turned off the overhead lights and used the camera's infrared mode. The door remained, pale now, but clearly defined. I placed a work light against the concrete. Its light fell across the wooden surface on the monitor, including the raised grain and the small highlight on the knob. In the room, the lamp shone against painted concrete. The door stayed closed throughout the visit. Jean decided to leave both cameras in place because that was the only way to learn whether the motion alerts and the image were connected. I configured the recorder to save continuously rather than only when movement was detected. I also set the new motion sensor to log every activation without sending a loud alarm. The first night produced 6 hours of an empty basement and a closed wooden door. At 2:43 in the morning, the new sensor recorded movement. Nothing visible crossed either camera. The door did not open, but the knob was turned downward in the recording. By daylight, it had returned to its original position. On the second night, the sensor activated at 2:41. The recording showed the door open by several inches. The gap beyond it was completely dark even under infrared light. No person appeared. There was no visible movement between the frame where it was closed and the frame where it stood open. Jean had not entered the basement and the other camera confirmed that the stairs and furnace area remained empty. The next afternoon, I returned and checked the wall again. It was still solid. I placed strips of blue tape on the floor at the exact points where the door frame appeared on the monitor. The tape ended at bare concrete. I also marked the position of the shelves so we would know if anything shifted overnight. The third recording showed the door halfway open. This time the infrared lights reached through it. Beyond the doorway was another basement. Chad, does that look like my basement? Gene asked. It it it has the same layout. I replied. At first it looked like an adjoining room. And then I recognized the furnace on the left and the same small window near the ceiling. The shelves were in the same positions and blue tape marked the floor beside the open door. It was Gene's basement repeated beyond the wall, but the second version looked neglected. The furnace casing was stained, several ceiling panels hung loose, and the storage bins had collapsed inward. The painted concrete was darkened in uneven patches. There was another wooden door in the rear wall of that basement. It stood partly open as well. The camera could not resolve much beyond the second doorway, but enough light passed through to show another set of shelves and another small window. The arrangement repeated again. Each room seemed to contain a doorway into an older, more damaged copy of the same basement. Later that day, Gene came downstairs with me and we compared every object in the feed with the room around us. The first repeated basement contained the same number of storage bins, but their lids were missing. A folding chair Gene kept beside the furnace appeared in the same location with rust covering its legs. Even the work light I'd left on a shelf was present beyond the door. Although its shade was bent and its cord had been cut. Nothing in those rooms moved. There were no people or shapes waiting in the repeated spaces. The camera showed enough detail to make each basement feel physical, but the real wall was less than 8 in thick. On the other side was soil. We left the system running for one more night. At 2:40, the motion sensor activated and stayed active for 11 minutes. The first door opened fully. The matching doors beyond it were open, too, creating a straight view through several versions of the basement. Each one was darker and more damaged than the one before. The final visible room had no working light, but the camera's infrared reached far enough to show that its shelves had fallen across the floor. Near the end of the recording, the doors began closing. There was no hand on any knob and no movement in the rooms. The farthest visible door was simply closed in one frame. Then, the next one was closed. The change continued toward the camera, one room at a time, until only the original wooden door remained open in Jean's wall. And that door closed last. I disconnected both cameras the next morning. When I connected either one to a test monitor facing another part of the basement, the picture looked normal. When I aimed them back toward the rear wall, the door returned. I removed the cameras, motion sensor, and recorder from the house. The saved footage continued to show the door, but after the equipment left the property, neither camera displayed it again. Jean moved several months later. He told the buyer that the basement sensor had been unreliable and chose not to install another one. The people in night vision. Diane lived alone in a large two-story house. She traveled several times a year and wanted cameras in the common areas so she could check for leaks, open doors, or anyone entering while she was away. She specifically did not want cameras in bedrooms or bathrooms. We agreed on six positions: the front entrance, kitchen, dining room, den, rear hallway, and upstairs hall. The system used small indoor cameras that produced a color picture during the day and switched to infrared after dark. I installed and tested them over most of an afternoon. The house had wide rooms, pale walls, and enough windows that the daytime images looked almost too bright. Every camera covered its intended area without showing private spaces. By the time I finished connecting Diane's phone, the sun was low. The den camera switched to night mode first because that room received the least daylight. Its color drained away and the picture changed to the familiar black and white image produced by infrared light. A person was sitting in the armchair beside the bookcase. I'd been in the den less than a minute earlier. The chair was empty. On the monitor, the person sat upright with both feet on the floor and both hands resting near the ends of the chair arms. I couldn't identify clothing, age, or facial features from the night image. The figure had the proportions of an ordinary adult and occupied the chair with enough weight that the cushion appeared slightly lower beneath it. I looked through the open den doorway. The chair was empty. When I changed the camera back to daytime mode through the settings, the figure disappeared. The chair returned to its normal color and the cushion looked level. I enabled infrared again. The same person was sitting in the same position. I assumed the camera was interpreting an object near the chair as a human outline. Infrared reflection can flatten depth and connect unrelated shapes. I moved a floor lamp, removed a folded blanket from the chair, and closed the curtains. And none of those changes affected the figure. It remained seated whenever night mode was on, and vanished whenever I switched back to color. Diane joined me at the monitor after I repeated the test several times. "How many people do you see?" Diane asked. "Five." "And every room is empty?" she asked. While I'd been concentrating on the den, four more figures had appeared. One stood beside the kitchen counter. Another sat at the dining room table. A third occupied the far end of the rear hallway. The fourth stood upstairs near Diane's bedroom door. The entrance camera still showed an empty foyer. Every physical room was empty. Diane and I walked through the common areas together while I carried the monitor. The camera feeds showed us moving normally. When I passed through the den, my image crossed in front of the seated figure, and there was no distortion or contact. I stood directly where its knees should have been, and I felt only the edge of the empty chair against my legs. The same happened in the kitchen. The figure beside the counter remained visible while Diane stood within inches of it. On the feed, they occupied overlapping space, but the image did not flicker. Diane could not feel a person beside her, and I saw nothing there without the monitor. I tested one camera for my service, Kent, a different brand that was not connected to Diane's recorder. I placed it in the dining room and powered it directly from a portable battery. In color mode, the room was empty. In infrared, the person at the table appeared from a second angle. The shape matched the first camera closely enough to show that both devices were recording the same seated position. The figures did not respond to us. They did not turn toward the cameras or change posture when we entered the room. They looked less like intruders than people who believed they belonged in those places. As the remaining daylight faded, the entrance camera switched into night mode. A sixth figure appeared in the foyer. It stood with its back near the front door, facing into the house. The system now showed one person on each feed. I changed all six cameras to color mode. Each figure vanished. Because the rooms were growing dark, the color pictures became grainy, but they remained empty. I turned on more lights and restored infrared one camera at a time. The figures returned in the same positions. Diane wanted the android cameras removed. I agreed, but before shutting the system down, I saved a short section of video from each feed. I'd loosened the den camera but had not disconnected it [music] when the monitor signaled motion upstairs. The hall camera still hung in place. Its view showed no obvious movement, but the figure near the bedroom door was gone. I checked the other feeds. The kitchen figure had disappeared. The den chair was empty on the monitor, and nobody remained in the dining room or rear hall. The entrance figure vanished last. For several seconds, every camera showed an empty room. Chet, where did they go? Diane asked. And then all six figures appeared on the upstairs hall feed. They stood close together along the wall, facing Diane's bedroom door. The narrow hall should not have had enough space for them, yet none of the bodies overlapped. They were arranged in two short rows, filling the area between the stairs and the bedroom. Their postures were calm. No one reached for the knob. No head turned toward the camera. The other five feeds remained empty. I reviewed the previous minute, and I found no recordings of the figures leaving their rooms or entering the hall. Between two consecutive frames, they were simply gone from one location and present in another. Diane stayed downstairs, while I disconnected the hall camera to the recorder. Its picture disappeared. I removed the remaining units without checking night mode again. The temporary camera from my service kit went back into its case, and the six new [music] cameras went into their original boxes. None had been damaged, and everyone worked normally when I tested them elsewhere. Diane replaced the indoor system with three exterior cameras and additional door sensors. The outdoor cameras used infrared, too, but they never showed anyone on the property. She moved her bedroom downstairs into a room near the kitchen. The upstairs remained furnished, clean, and mostly closed. I serviced her exterior system twice over the next few years. On my first visit, Janet told me her old bedroom door remained shut. She had never seen the people without a camera and had never heard them moving through the house. She preferred it that way. She could live with empty rooms. She did not want another device telling her they were occupied. Channel 7 Janet's system was already installed when she called me. Six analog cameras ran through the walls to an eight-channel recorder in a cabinet beside her television. Two cameras covered the front and rear doors, two watched the sides of the house, one covered the garage, and one looked across the entry hall. The recorder was older, but serviceable, and every cable was labeled neatly. Janet said the problem had started three nights earlier. The monitor normally rotated through the six connected cameras. Shortly after midnight, it had displayed a seventh channel. She had assumed an old camera was still connected somewhere, but the extra picture showed parts of the house from locations where she knew no camera had been installed. Janet arranged a late appointment because the extra channel had always appeared after midnight. I arrived at half past 10:00 and inspected the recorder before turning it on. Ports 1 through 6 had cables. Ports 7 and 8 were empty. The system was not connected to the internet, and the recorder did not support wireless cameras. Any image assigned to channel 7 had to enter through the empty connector on the back. For the first hour and a half, the monitor showed only the six normal feeds. I tested every camera, reviewed the recorder settings, and confirmed that no duplicate channel or demonstration footage had been enabled. A few minutes after midnight, a seventh box appeared on the screen. "Where's that camera?" Janet asked. "There isn't a seventh camera." I said. Channel 7 showed a narrow unfinished space. Wooden studs divided the view, and several pipes ran vertically through the center. Yellow insulation filled part of the background. A metal strap crossed one pipe near the bottom of the picture. Janet recognized the arrangement from photographs taken during a kitchen renovation. The camera appeared to be inside the wall behind her cabinets. There was no room for it. That wall contained water lines, wiring, and insulation with drywall on one side and cabinets on the other. Even a camera small enough to fit would have needed power and a cable. No wire entered the empty seventh port. When I disconnected the entry hall camera, channel one went dark. Channel 7 remained. I continued through the other connections until all six physical camera cables rested beside the recorder. The monitor showed six blank channels and the same live view inside the wall. The image had details that made pre-recorded footage unlikely. When Janet turned on the kitchen faucet, one of the pipes trembled slightly. I touched the wall near the sink, and a small amount of dust fell through the picture. The camera was showing that exact wall at the exact moment from a position that could not hold a camera. I removed the recorder's storage drive and restarted the system without it. Channel 7 returned as soon as the monitor displayed the grid. I covered the empty seventh input with a metal terminator and then disabled the channel through the menu. The box disappeared for several seconds. It returned on its own, again labeled seven. But the view was different. This time, the camera appeared to be in Janet's attic. The roof boards formed a low triangle above loose insulation. A cardboard storage box near the hatch gave us the direction of the view. The camera seemed to be positioned near the far end of the attic, pointed back toward the center. Something occupied the space beside a roof support. Chet, that's my attic. I know. Keep the hatch closed, I replied. The image was clear around it. I could count the printed lines on the storage box and see individual fibers in the insulation. The thing beside the support would not resolve in the same way. I did not describe it to Janet. She could see the monitor and neither of us would have gained anything from deciding what it resembled. But it looked like something out of a nightmare. We stayed downstairs. The attic hatch was visible from the hall and it remained closed. The picture held for several minutes. The roof support, box, and insulation never shifted. That nightmare thing the attic remained in the same place. I checked the recorder, temperature, and voltage, even though no ordinary defect could create a live image of another room. Both readings were normal. Channel 7 vanished while I was examining the power supply. A moment later, it returned with another view. We were looking out through the slats of the kitchen return air grill. The camera appeared to be inside the wall, no more than a few inches behind the vent. Part of the kitchen table was visible through the middle openings. The perspective was low and slightly tilted, as if the lens rested against the bottom of the duct. I stood at the edge of the real kitchen and compared the feed with the grill. It matched. The angle showed the same chair leg, cabinet edge, and section of floor. I removed the grill while keeping my body out of the camera's view. Behind it was a shallow duct that turned downward. I used a flashlight to examine as much as I could see. There was no camera, wire, or opening into the wall cavity. On the monitor, the grill was still in place. That was the first time Channel 7 stopped matching the current room. The physical grill rested in my hand, but the feed continued looking through its slats. The kitchen beyond it appeared exactly as it had before I removed the cover. And that nightmare thing from the attic was now visible beyond the grill. It occupied most of the space close to the hidden camera. The slats divided the image into narrow sections, and none of those sections provided a complete outline. I could only tell that the space behind the vent was no longer empty and that whatever filled it was positioned near the lens. Janet remained in the adjoining room where she could see both me and the camera. She did not approach the kitchen. I turned off the recorder. Channel 7 stayed on the screen. I disconnected the recorder's power cable and pulled the video cable from the monitor. The six blank channel boxes disappeared along with the clock and menu bar. The picture from behind the grill remained by itself. Without the recorder connected, the monitor had no source for any image. The thing near the lens did not move. The image simply changed between one frame and the next. It had been several inches beyond the grill and now it occupied the entire view. After several seconds, the screen went dark. I packed the recorder separately from the cameras. I did not reconnect anything in that house. Janet kept the door sensors and alarm panel, which used different wiring, but asked me to remove all six cameras. I checked the attic in daylight the following morning with another person present. There was no camera near the roof support and no sign that anything disturbed the insulation. At my workshop, I connected Janet's recorder to a test camera and monitor. The six original inputs worked. Ports 7 and 8 remained blank. I let the system run through three nights and never saw the extra channel. I checked the storage drive separately. It contained recordings from the six ordinary cameras, but there was no file for channel 7. Well, as you can imagine, Janet sold the house the next year. She kept the alarm, but never installed another camera there. And she never saw another nightmare creature. I stopped accepting interior camera jobs after that installation. I still work on exterior cameras, door sensors, alarms, and access systems if you need somebody. But I will not place a live lens inside a customer's living space. Not after that. My father believes every camera problem has a cause. I still believe that, too. What changed is my understanding of where the cause might be. A damaged wire, a weak power supply, and a bad recorder are all things I can find. I can replace them and prove the problem is gone. I cannot repair a camera that shows a door inside a concrete wall. I cannot dry footprints that exist only on a monitor. Or clear people from rooms that look empty in daylight. Most of all, I cannot explain a video source that keeps producing a picture after every camera and cable has been removed. And I cannot explain that thing in the attic. I still check every feed before leaving a job. I make sure the doors are closed, the yard is empty, and the picture [music] always matches the room in front of me. >> [music]