Archive for DVR Project

DVR Project – Early Implementation

My old DVR system was a normal PC Tower with a 15″ crt monitor and occupied WAY too much space, it also utilized RCA connectors instead of the more typical BNC connectors. The end result? Every time I had to do any work on the surveillance system I would end up making a mess of things (this is the primary reason I wanted a compact desktop PC for the second version) popping cables off their connectors, mixing wires up. There was A LOT OF CLUTTER.

A few weeks ago I experienced a problem with one of the cameras, intermitant signal loss. An irritating problem to diagnose, and one that caused the usual chaos when I started tampering with things. I got angry and despite the fact that the RAM had not yet arrived for the new DVR I decided it was time to pull the old one out.

A few hours of manhandling hardware in and out, some mucking around with the configuration, more grumbling and more mucking around and finally the entire system comes back online.

The computer specs represent a significant increase in power and storage;

Old System: 1.7Ghz Processor, 256Mg Ram, 40Gig HD
New System: Dual 2.8Ghz processors, 2.5Gig Ram, 500Gig HD

By way of storage the old system recorded 6 video feeds (5 camera feeds at 320×240 @ approx 2.5fps recording only motion and 1 feed from the multiplexer at 320×240 and 5fps recorded 24/7) and would keep the last 4 days of activity before automatically overwriting the oldest activity.

New system records 7 feeds (6 camera feeds at 640×480 @5fps recording motion, motion recorded at 20fps when detected and 1 feed from the multiplexer at 1280×1024 @ 5fps recorded 24/7), settings are still being fine tuned but at present 10 days of activity are kept before oldest events are overwritten.

Of course the switch to 640×480 has introduced some interlacing issues as these are analog cameras, and I may finally have to begin experimenting with IP cameras to get around this.

In the meantime I couldn’t be happier with the system. I still need to clean up the wiring quite a bit, but for now it’s adequate (if ugly).

DVR Project – Asthetics

I suppose it would have been easiest to just go with an HP branded DVR system, but then that would be far too easy to be one of my projects wouldn’t it? I figured as long as I was spending time sitting around waiting for various components to be delivered by the mailman I’d take the time to spruce up the computer just a bit, and give it a “custom” appearance.

Rebranded as a BOS Digital Video Recorder
(Looks kinda official, no?)

» Read more..

DVR Project – The Computer

There are a few critical elements to the computer I was seeking, it had to be fairly recent (2Ghz+/- processor speed), it needed a few Gig of memory (either present or at least capable of increase) and the storage needed to be fairly large (IDE 500Gig HD minimum).

The computer also needed to be capable of working with the Ebay 8 channel DVR card (sadly something I never know until it comes time to plug the damn thing in) and finally the computer needed to be inexpensive.

I may have acquired that computer yesterday. An HP Compaq Desktop, 2.8Ghz processor, 500Meg RAM (one slot used and one remaining open, SDRAM) and 40 Gig HD, along with onboard sound, NIC etc etc etc for $89.99 – A clean $100 after the gummint took their due).

New DVR Computer?

The only problem with the PC appears to be the HP specific half height PCI slots, but thankfully the DVR card is half height, and it was a moments work with tin snips to get the card to fit with the case closed. Some quick exploring in Ubuntu and I confirmed that at least it was seeing 8 video devices (a good start), although all kinds of hardware detection errors kept threatening to rain on my parade.

In the end the card works fine. Up and running within a few hours. Higher resolution and frame rates chewed through 50% of the 40Gig drive in about 3 hours of experimentation (that was just one camera 640×480 at 25fps), upgrades are going to be imperative.

More experimentation with different distributions of Linux to see how much I can decrease the overhead without making the thing any more unfriendly than absolutely required.

Update: Satisfied that the DVR card works, I’m proceeding with upgrading both memory and storage, this requires some bargain shopping for DDR Ram and a 500Gig IDE drive so may take a bit of time. The good news is that experimentation with Linux doesn’t have to wait for the upgrades.

DVR Project – The DVR Card

My current system relies on 4 channel PCI cards each based around a single Conexant BT878 chip. They advertise a rate of 30fps (frames per second) which decreases by 50% for each camera added, resulting in approximately 3.5fps if all four ports are used (these cards can be had on Ebay for $10 a piece, with shipping included).

The card acquired for the new DVR relies upon 8 separate Conexant BT878 chips and claims 240FPS (30 FPS per input), although this will need to be tested (truthfully 10-15fps would be more than adequate).

8 channel DVR PCI Card

Image resolution at 640×480 is doable, although this is a problem for the computers processor not so much for the DVR card.

Increasing both the frame rate and resolution will have a significant impact on the computer (processor, memory and storage), and I’ll have to seek the most bang for my buck in these respects.

As with the existing system all but the multiplexed feed will be recorded on a “movement” detected basis, and while this will reduce (if properly configured) the storage space needed it will increase the demand on the processor and memory.

Obviously I’ll be shopping around looking at computers next.

Digital Video Recorder – Project Outline

The backbone of any surveillance system is the recording package, without it the cameras are useful only as long as a human operator is sitting there and watching them (and even then only if the operator is capable or willing to initiate the appropriate response when something is observed).

For older systems time lapse video tape was used, but this suffered from some serious weaknesses (not the least of which was the physical degradation of the media itself, and the fact that reviewing the tape was a time consuming and cumbersome proposition).

Digital video recording offers longer recording times, no “worn media” issues and when combined with the appropriate software allows operators to scan through “event logs” looking for activity instead of reviewing tape on an almost 1:1 basis).

While Digital Video Recorders for surveillance systems are appreciably less expensive than even a few short years ago they can represent the single largest expense in a “budget” or hobby system.

I’ve decided that I need to increase the capacity of the system in use for my home, and have decided that this will require me assembling a new system rather that upgrading the old one (for a number of reasons, not the least of which is that upgrading the old system would take the recorder offline while it was reconfigured, and this way there’s no pressure to hurry things along.).

Commercial multi-input DVRs are available for between $900-$5000 and offer a dazzling array of options. For the purposes of this project we are going to stick with a fairly basic system comprised of a Personal Computer running Linux, free software (“Free” as in open source, not “free” as in pirated) beefed up RAM, processor and storage space and a DVR card capable of monitoring at least eight cameras).

The goal is to keep the footprint of the entire package as small as possible (so no HUGE tower desktops for this one), and to see if we can’t get a nice slick “custom” appearance into the deal.

We’ll be aiming to bring the entire system in under $400 (CDN) with the parts accumulated over time to spread out the up-front costs.