Security Camera Installation in Texas, Planned Around What the Footage Must Prove

Decide what each camera must prove before choosing a camera

Disappointing camera systems tend to fail for one reason. The cameras were picked from a spec sheet and then pointed at the property, instead of being chosen for a particular job at a particular spot. After a car break-in, the footage shows a shape crossing the driveway, but no face and no plate. EVOTECH IT LLC works in the opposite direction, starting with what each view is supposed to prove.

The security industry has a useful vocabulary for this, drawn from the IEC 62676-4 standard and usually shortened to DORI:

• Detect: know that a person or vehicle is present.
• Observe: see what they are doing and roughly what they look like.
• Recognize: tell whether it is someone you know.
• Identify: capture an image good enough to identify a stranger.

A parking lot overview only has to detect and observe. The camera watching the front door, the register or the side gate has to identify. Those call for different lenses and different mounting positions, and a sound design mixes them on purpose rather than buying one model in bulk.

As for the Texas in this page’s title: our crews install security cameras on site in Greater Houston, covering Katy, Cypress, Houston, Sugar Land, Richmond, Fulshear and the areas in between. Installers are not sent to other Texas metros. EVOTECH has spent over twenty years in low-voltage work, carries a license and insurance for it, and is rated 5.0 by Google reviewers, and we keep our installs within driving distance of the people who will need us for service.

Pixels on target matter more than megapixels on the box

Resolution only means something relative to the width of the scene a camera covers. A 4-megapixel image is about 2,560 pixels across. Spread those pixels over a 60-foot-wide driveway and each foot of scene gets roughly 43 pixels, enough to observe but not to identify a face. Narrow the same camera to a 25-foot doorway and each foot gets about 100 pixels, which is identification quality.

The DORI targets convert approximately as follows. The widths are simple division of horizontal pixels by the target density:

The detection and observation widths are theoretical. In practice, lens angle, night lighting and obstructions limit a camera long before those numbers do. The identification row is the one that shapes real designs.

Two practical consequences follow. First, one wide-angle camera cannot identify people across an entire storefront, so you need a tighter view at the choke point everyone must pass, usually the entry. Second, a varifocal lens, which lets us set the field of view on site, earns its cost at positions where the right width is hard to predict from the ground.

License plates are their own problem. Reading a plate on a moving car takes a narrow view, a fast shutter to stop motion blur, and infrared tuned so the reflective plate does not bloom into a white rectangle at night. A general overview camera almost never manages it, so when plates matter we plan a dedicated camera at the gate or lot entrance where cars slow down.

Gulf Coast glare, humidity and lightning

Houston conditions are hard on cameras in ways a product brochure never mentions.

Backlight and glare

A camera inside a glass storefront, facing a bright parking lot, turns everyone walking in into a dark silhouette against a washed-out background. The remedy is true wide dynamic range, which blends multiple exposures, rather than a digital setting that simply brightens shadows. West-facing cameras hit the same wall at sunset, when the sun sits in the lens, so we note the compass direction of every view during the site walk.

Night and infrared

Built-in infrared illuminators make night images possible, and they also bounce. A camera tucked tight under an eave, or aimed along a nearby wall, reflects its own IR back into the lens and fogs the picture. Infrared draws insects too, and spider webs across the lens are a genuine source of false motion alerts after dark. Better placement, a little distance from the soffit, and at some spots a floodlight or a camera built for color at low light solve most of these problems.

Moisture and storms

Outdoor cameras should carry at least an IP66 weather rating, and junction boxes should be sealed so humid air is not drawn into the housing along the cable path. Every exterior penetration is sealed, and each cable is formed into a downward loop below the entry so rain drips away instead of tracking into the connector. Lightning is frequent enough here that long outdoor runs, especially to poles and detached buildings, deserve surge protection where they enter the building and a proper ground.

Where the video lives, and for how long

Recording follows one of three patterns, and each involves a trade-off.

• On-site NVR. A network video recorder writes footage to hard drives at the property. It keeps recording when the internet drops and carries no monthly storage fee. Its weakness is physical: if a burglar takes the recorder, the evidence leaves too. We install NVRs in a locked or out-of-sight location for exactly that reason.
• Cloud recording. Footage streams straight to a provider’s servers, which protects it from theft on site, but every camera uses your upload bandwidth around the clock, and storage is billed as a subscription.
• Hybrid. Continuous recording on site, with events or selected cameras copied off site. For many businesses this is the practical middle ground.

Storage is plain arithmetic once the inputs are known. A camera streaming at 4 megabits per second produces about 43 gigabytes a day. Sixteen of those cameras recording continuously for 30 days need roughly 21 terabytes. H.265 compression, lower frame rates on overview cameras and motion-triggered recording on quiet views can reduce that figure considerably. We size the drives to the retention period you choose, rather than letting the drives decide it, and we use drives designed for nonstop surveillance writing instead of desktop models.

Two settings deserve more attention than they get. The recorder’s clock should synchronize with a reliable time source so timestamps hold up when footage is reviewed later, and retention should be set on purpose rather than left at the factory default of recording until the disk fills and then overwriting.

Upgrading from an aging analog or DVR system

Many properties already have cameras from an earlier generation: analog models on coax cable, feeding a DVR in a closet. Replacing them does not always mean starting over.

• HD-over-coax. Cameras and recorders built on the HD-TVI, AHD or HD-CVI formats can use existing coax at far higher resolution than the original analog system, as long as the cable tests clean.
• Ethernet over coax. An adapter at each end carries network data, and with some models power, across a coax run, so IP cameras can reuse the old path.
• Fresh Cat6. Where coax is weathered, spliced or routed through paths that no longer make sense, new cable is usually the better long-term decision and opens up the full range of IP cameras.

We test existing coax before recommending reuse. Old installations often hide splices above ceiling tiles, water inside outdoor connectors, and one power supply daisy-chained to several cameras, and each of those changes whether reuse is sensible. If any footage on the old DVR still matters, export it before the recorder is disconnected.

A camera is a computer with a lens, so secure it like one

IP cameras run an operating system, store passwords and talk to the internet. Installed carelessly, they become the weakest device on your network. Our baseline on every install:

• Cameras run on wired power over Ethernet, not WiFi, which keeps video traffic off the wireless network and makes recording independent of WiFi problems.
• Default passwords are replaced, and each person who reviews footage gets an individual account with suitable permissions.
• On business networks, cameras sit on their own VLAN, isolated from workstations and point-of-sale systems.
• Automatic port forwarding through UPnP is disabled, and remote viewing runs through the manufacturer’s secured app or a VPN rather than an open port.
• Firmware is brought current at install, and we show you where to check for updates afterward.

The switch must also deliver enough power. An 802.3af port sources at most 15.4 watts and an 802.3at port at most 30, while pan-tilt-zoom cameras and models with heaters may need the higher 802.3bt classes. We add up the power budget before choosing the switch, so no camera reboots every evening when its infrared switches on.

Organizations with federal contracts should know that Section 889 of the 2019 National Defense Authorization Act restricts video surveillance equipment from certain manufacturers. If that applies to you, mention it at the estimate so the equipment list is compliant from day one.

Where cameras belong on a house and on a business

Homes

The most common residential mistake is mounting every camera at the highest corner of the roofline. From up there, a camera looks steeply down on hats and the tops of heads. For the front door, a camera at roughly eight to ten feet, looking at a shallow angle toward where visitors stand, captures faces, and a video doorbell adds the close-up view. Driveways, side yards, gates and the back patio each get a view planned around whether we need to detect movement there or identify a person.

Businesses

Commercial layouts usually combine inward-facing cameras at every entry, one over each register or counter, overviews of the sales floor or warehouse, dock doors, and exterior views of the lot. Cameras at entries face into the building toward people arriving, so faces are lit by daylight instead of appearing as silhouettes. Where card readers control doors, cameras at those doors can be tied to door events, a subject our Texas access control installation page covers in detail.

Privacy lines we keep

We do not aim cameras into a neighbor’s windows or yard beyond what a view unavoidably includes, and we never install cameras in restrooms or changing areas. Microphones stay off by default. Recording conversations falls under wiretap laws that treat sound differently from images, so audio should be a deliberate choice you make after checking your obligations, not a factory setting nobody noticed.

Installation day, in the order it happens

1. Mark and confirm. We mark each mounting point with you before drilling and restate what each camera is meant to capture.
2. Run the cable. Cable travels through attics, walls and soffits to a central location, in conduit wherever it is exposed outdoors. Our network cabling page explains how those runs are supported and terminated.
3. Mount and seal. Cameras go on junction boxes or mounts rated for the surface, and every exterior hole is sealed.
4. Aim against a live picture. Each view is adjusted while watching the image in real time, then locked down.
5. Configure the recorder. Schedules, motion zones, retention, time sync and user accounts are set up.
6. Check the night picture. Where possible we confirm low-light performance after dusk or by simulating darkness, because a clean daytime view can hide infrared problems.
7. Teach playback. Before we leave, you find an event, scrub through it and export a clip yourself.

Getting a clip out for police or an insurer

Footage only helps if you can get it off the recorder. After an incident, export the clip promptly, because continuous recording will eventually overwrite it. Most recorders offer two formats: a native file that plays in the manufacturer’s player and preserves timestamps and watermarking, and a standard MP4 that plays on any computer. Saving both is a good habit. Write down the date and time range, the cameras involved and who performed the export. We walk every customer through this at handover, since the day you need a clip is the wrong day to learn how.

Related services

Frequently asked questions