printer friendly version

Letters to the editor

Dear Lynne,

I am not an expert on video systems, but having worked a little with video over IP I was struck by the assertions made in the March/April edition by Vision Catcher regarding the bandwidth requirement of IP video. The author claimed that a 100 Mb/s Ethernet network could only carry 8 frames per second of video, which from my own experience I knew to be nonsense, having just done some tests with Axis IP cameras on an old 10 Mb/s network at 30 frames/s!

A moment's analysis revealed the error: the writer had dropped a factor of 100. His calculations should then have shown that the network could support 800 frames per second - a somewhat different proposition. In fact, I would estimate that 200 to 400 frames per second would be a more realistic estimate, because users of IP video do not generally want to accept the rather poor picture quality that 5 KB/frame gives you, and would more likely be using 10-20 KB/frame. The constraints of analog cameras and the rather expensive, proprietary hard disk systems used in DVRs make 5 KB the common figure in that world, but IP video uses much cheaper, standard hardware (especially disk drives), so you do not need to make such severe cost/quality trade-offs.

Also, of course, it is rare for an organisation to have only one segment in its LAN: given that switched hubs are the norm these days you are actually looking at a limit of 200-400 frames per second per LAN segment, not as a company-wide limit.

Dr Chris Crozier

Apology for my mistake

Dear Lynne

I apologise unreservedly for my inaccuracy and in no way did I intend to mislead or misrepresent the detail of my article; I did not check my figures after writing the article due to time constraints of Securex, a poor excuse, perhaps.

If you look at my calculations and factor for the dropped 100 then my calculations are correct. Thank you very much for picking this up, it is reassuring that articles are read and understood. It is Vision Catcher's aim to educate within the security market rather than misdirect. In essence there are issues on bandwidth that need to be investigated in more detail prior to sweeping statements being made.

The costs of differing systems needs to be explored and the facts in the marketplace do not always favour one choice over the other. The key point is to be aware of all the facts before you purchase and not be led by the hype towards total IP solutions without understanding both sides of the story.

This was the intended direction of my article and comments made by Dr Crozier highlight my point exactly; ask the questions and do your homework before purchasing.

Niall Beazley

Director

Vision Catcher

Ed's note

Hi-Tech Security Solutions welcomes feedback from its readers on the content it publishes, particularly when it comes to errors of fact or omission. We try hard to ensure that what we publish is of value to readers, by keeping a beady-eye on vendors' claims, stripping out hyperbole and marketing hype, and ensuring that credible content continues to be the mainstay of the publication. It is somewhat astonishing that after 10 years of education in the pages of Hi-Tech Security Solutions that we still find such vigorous debate on matters digital. It is clear, however, that with IP becoming such an important component of security systems, it is likely that there will be more debate on more contentious issues than ever before. Vendors should not, however, lose sight of one simple fact: End users have security problems to solve.

Keep it simple, solve their problems, gain a client.

Share this article:

Further reading: