Jump to content
Security Installer Community

Police Push On Poor Quality Systems


Chorlton

Recommended Posts

Very interesting comments chaps :yes:

I've lost count of how many arguments I've had in the past with industry 'experts' about one compression method over another.

Being a very simple soul at heart, the bottom line for me is not what compression is actually built in to the recorder, but rather the primary quality of the recorded images; in other words, it shouldn't matter whether the DVR is based on M-JPEG, MPEG-4, Wavelet, H264 or any other for that matter, as long as the recorded images are providing sufficient clarity to be "fit for purpose".

Where that somewhat simplistic theory collapses totally, is when we look at the way many cameras are set up, often producing images which have little if any evidential value, and so the burden of necessity falls on a Forensic Video Analyst to try and extract something useful from the recording, which is of course where the problems creep in particularly with various lossy compression techniques.

I always find it interesting to consider the requirements for a DVR depending on the individuals perspective; for example, a far eastern manufacturer may want it to come in at the lowest possible price, which perhaps means using a compression with low licensing costs, that produces small file sizes and consequently only requires smaller volume HDD's, in other words, a cheap recorder to produce.

Given that it may record four times longer than a rivals looks great for an uninformed customer, but unless all the various technical functions are correctly weighed up on merit, the chances are that a less than incisive judgement will ultimately influence the final purchasing decision.

The installer generally wants quality, reliability, and hopefully affordability ;) ... and a happy customer (getting paid on time would also be rather nice)

The end user wants somefing called CCVT wot records, and hopefully isn't going to break the bank. Unfortunately the majority have little if any genuine interest in the subject, and if the truth be told, would probably prefer not to have to pay for any type of security at all, given half a chance.

The criminal justice community of course, just want a perfect picture of the suspect ... period.

So how do you make sense of all these different perspectives? Well as piggy in the middle, I have to say it just isn't easy :banghead:

Until we have some proper standards for equipment and image requirements, the best we can do is probably, the best we can do.

Personally I would have preferred a JPEG, Wavelet or MPEG-2 based recorder given the choice, but then being realistic about the challenges of designing effective CCTV systems, I've long believed that optimising the images is probably the most important single factor even before we get to the subtleties of various record options.

Incidentally, for various reasons I'm not actually going to comment on this manufacturers offering, but for your interest, there is actually a new player in the States that's recently introduced a record solution that .... degrades the image quality over time, to reduce the amount of storage capacity (i.e. cost) required. Nuff said.

Funny old world isn't it :whistle:

Link to comment
Share on other sites

The unfortunate thing as already mentioned is that customers often want the cheapest solution not nessarily the right one for the job, i.e a 160Gb hdd instead of say a 250GB and they always want to squeeze as many days recording out of it as possible much like the "96hrs on a 3hr" tape in the days before DVR's were as available as they are today.

I mean lets look at the numbers here - a standard CIF picture (352 x 288) now multiply that by 25 (realtime frame rate) and then by number of cameras on the system say 4 that is 352x288x100 which is 10,137,600 individual pixels of information per second, this doesn't take into account any colour information which could double / triple / quadruple that number depending on how many colours are stored by the system, that's roughly 10megabytes of information and that's just 1 second. Now if you could put that on the Hard drive inside it at that rate you'd run out of space in roughly 4h 42m using the 160Gb drive above, ignoring colour. Mind you how often would you need to run 4 cameras @ 25fps unless your recording something very fast. So drop that to say 6fps and you've quadrupled the time or you could add in colour to this now and maybe get double.

So Doktor Jon I hear what your saying, if they could get the compression right it's fine. I've dealt with JPEG images that got compressed too much, it resembled block art almost. Each compression technique has it's "balance" point where what you gain in space outweighs what your losing in quality i.e sure you've got loads of recording space but the images recorded aren't anything like the "live" image.

I do like the ideas some are coming up with to beat the space problem since hard drives are becoming bigger they increase the maximum size you can fit and even offer "external" units that can increase capacity to silly figures, I think the most I've seen is 4TB which is a lot but of course those systems aren't cheap so back to square 1.

Intruder / CCTV / Access Control Technical Support Personal

Subscriber to the "K.I.S.S" principle, that's Keep It Simple Stupid, are you?

Link to comment
Share on other sites

A lot of these points will be answered in the suggested 'Quality of system' standard that BSIA are leaning towards atm.

Each system being graded on evidential quality will encourage a higher standard of install as well (Note that as Jon says this does not neccesarily mean more cost - just better practice).

On a seperate subject how about Dirac or Agw as a compression format? Dirac was good enough for the BBC @ the olympics recently...

'J

btn_myprofile_160x33.png


 

Link to comment
Share on other sites

...

I went on to ask about H264 and his response was not encouraging( by the silence). Apparently the lawyers like JPEG due to the reason that a new frame is stored everytime and that complex algorythums are not used to "assume" pixels......

What do laywers know, even jpegs use 'complex algorithms' to assume the missing pixel data, thats the point of compression schemes, they compress the data, and assume the missing data from algorithms. Unless its a lossless jpeg algorith, which is not something I have seen on any dvr. It may be out there, but not common.

JPEG suffers from artifacts, due to the compression, and a lot of dvrs tend to overcompress the image.

Not only that but most modern cameras have their own complex algorithms inside for digital signal processing, so even before its got to the dvr the image is not the same as what came through the lens.

Lets face it, no compression algo, or DSP is going to make a length of footage show something that did not happen. Its not going to make a innocent person look like they pulled a gun and robbed a bank.

All they will do at worst is leave the image looking pretty c##p if its a bad algo, or over compressed. But the things that occured in the footage, did occur, jpeg or mpg, or whatever.

Link to comment
Share on other sites

Yes, geo for example can watermark their mpegs.

Never seen the point of watermarking though, as surely if someone is actually going to bother faking footage,they either reverse engineer the watermark, and recalculate new accurate watermarks for the fake footage, or they could just play the fake footage back into the dvr to let it make real watermarks..

Either way, the watermark issue would be easier than faking hundreds or thousands of frames to fit someone up....

Link to comment
Share on other sites

  • 3 weeks later...

Sorry for not conforming to the general concensous but why is h.264 being bashed in a general sense? For those of you that has watched a Blue-Ray or HD-DVD, a DVD-T broadcast or IPTV have most likely been watching h.264 (a profile of it) encoded/decode video stream!

To say that all JPEG/MJPEG DVRs produce acceptable picture quality would be ridiculous but no more ridiculous than saying all h.264 DVRs produce unacceptable picture quality. At the same time if someone stated that they have yet to see an acceptable picture from an H.264 DVR then I would completely understand but would say "wait, it will happen! in my experience it has happened!" All it needs the for H.264 to be applied in a true and correct manner.

Please remember that even the highest specification home PC would struggle to encode a real time H.264 video stream and how a "cheap" version may misrepresent the technology.

BTW, I agree with kenplace first post, if H.264 (MPEG4 Part 10) is unaccetpable to a court of law even with watermarking then why would any compression method, like MPEG4 Part 2 which is also known as H.262. Compression methods are unbiosed and without ulterior motives, unlike the shop keeper downloading the footage (not that I have met one capable of doing so :rolleyes: )

Link to comment
Share on other sites

My reason for disliking mpeg based machines is because none that i have seen offer a high enough bitrate to provide a good qaulity image if mopst of the scene is changing.

I agree that you cant say all jpegs are good etc and h264 is caperble of excellent quality IF the bitrate is high enough..

securitywarehouse Security Supplies from Security Warehouse

Trade Members please contact us for your TSI vetted trade discount.

Link to comment
Share on other sites

Ah but of course that all goes out the window when the customer changes their mind and wants double the record time without paying for the additional hard drive space..

Intruder / CCTV / Access Control Technical Support Personal

Subscriber to the "K.I.S.S" principle, that's Keep It Simple Stupid, are you?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.