roygbiv

Kind of ironic that a guy from sales accuses me of marketing bull since I work in the purely technical area of DSP engineering - currently not applied to cctv but may do in future. Anyway, does not look like any contribution from me would be welcome here, so you win Thomas, I am out of here.

Was looking into this myself recently but could find no definitive answer. Problem with that report is the date. LEDs have increased in output dramatically since then. It looks like the technology is improving so fast that the safety committees can't keep up.

Soundy. Do you have heaters in the domes or does that cause more problems than it solves?

Yes. You just rename it to NVR. Now you're getting technical.

One solution – assuming you have the network capacity, connect cameras and mics to a video server in each room, connect servers to the network and record over the network on remote computer with recording software. no need for a DVR!

The .dvr extension is just a container like .avi. The video is encoded in a proprietary format using a codec specific to the model of DVR. You generally use a player application or client software to view the video on a PC. The software is usually included on a CD supplied with the DVR. You may be able to find the player app on the web.

How much extra recording time you get by dropping the framerate depends greatly on the compression codec used. With mjpeg the relationship is near linear ie halving framerate doubles recording time but with mpeg4 and especially H.264 halving the framerate may only increase recording time by 20%. H.264 codecs are usually dynamic. In other words it will take more recording space when recording a lot of movement (I mean continuous recording, not motion only recording). This can have a bigger effect than changing the framerate. Also worth noting is that 'noise' cannot be compressed so cameras with good noise reduction will give much smaller files with H.264.

You should be able to save the expense of upgrading your PC hardware by using hardware compression cards. I am only a techie so don't know much about brand names etc but you should be able to find suitable cards that will do what you want if you google SMICT compression. They are sold under a variety of different names but are all the same hardware and software from a company in Taiwan. Have seen them used widely in stores across Europe, good quality video and good networking and yes you can view all channels on screen at once (unlike similar cards from China). SMICT compression is a form of H.264 and offers good framerate on remote view and low storage requirements. Look for the cards that do Real Time on all channels, the cheaper ones will do part of the compression in software and will put a strain on your hardware. I don't remember seeing an 8 channel SMICT card but I have seen fairly modest PCs running 4 X 4 channels at full frame rate without a problem. Sorry I can't be more specific about brands and models, I really need a memory upgrade .

Not so much with MJPEG but with MPEG4/H.264 yeah it can become a huge burden to decompress for on screen viewing. I also have an MPEG4 camera on that system but I have it set to only decode keyframes. The local display looks like maybe 5 fps but the recordings are much closer to 25fps and the CPU load was dropped quite a bit. Granted this is a VGA camera not megapixel. I'll make you a deal on a little valley if you wanna play with one. Collin Hopefully you are not basing this on experience of the Arecont H.264 ( or I may weep!) You may well have found older H.264 samples a real hog. The scalability extensions for H.264 were only finalised in 2007. When implemented there is no need to decompress then re encode for display on screen. I can play multiple H.264 streams at D1 25fps on an old 2GHz Celeron. On a Dual Core it barely registers cpu usage.

Good reading on that site, thanks for the link. Apologies for going off topic but I see no one has really answered your concerns over decoding H.264. This is really not a big issue. H.264 is certainly processor intensive to encode but decoding is comparatively easy, as it is for most standards/codecs. However, one of the major advantages of H.264 is scalability. The scalability extensions (SVC extensions) built in to the standard allow the video stream to be scaled without re-encoding. This allows a high resolution stream to play on a low resolution device, the decoder simply rips the appropriate resolution from the high res stream. I know it works as I use an H.264 DVR that records every channel at 25 fps, 704 X 576. Playing back a dozen streams at once on a Dual core PC, without scaling, uses 10% CPU. This is close to the equivalent of playing a 5 megapixel stream at 25fps on a monitor with a 5 megapixel resolution. With scaling, so all the player windows fit on my monitor, barely registers on the CPU meter. Hope this answers your concerns.

Good to hear that healthy competition should soon drive prices down. Axis have had it all there own way for too long. They must have made a fortune out of those little plastic jobs with cheapo camera-phone sensors.

Valid comments, unfortunately you appear to have confused frequency with wavelength. IR is longer wavelength (higher nm) therefore lower frequency than visible light.