tesc_cctvpro

Network latency is almost zero in IP cameras compared to processing latency. 95% of the lag you're seeing is because of the time the camera takes compressing the video, and the time the VMS takes decompressing it for viewing. The two "latencies" are NOT related, despite the same term being used for both. Actually, using "latency" for this issue is probably a misuse of the term anyway, and obviously leads to some confusion - "lag" is much more suitable. We could continue arguing around PTZ cameras, and this and that for ages. The major point is: The nature of IP is such, that Latency as a result of data congestion, is a very real challenge in video ip systems. Congestion on a network will very quickly create latency variations far higher than 250ms, to the extent that not only are PTZ operations affected, but also the integrity of video reconstruction, once too many packets start arriving after the deadline -issues such as synchronization of text and audio with video will also start to suffer adversely. Any and Every Video Network install, if it's professional grade, must address the effective managment, of the potential effect Latency will have on the integrity of video security transmission. Managing latency to ensure video transmission is a complex process. and so the point is: HDcctv requires far less complexity and expertise in this domain- as the need to manage and control latency, resulting from congestion is far reduced. I would be a bit wary of an IP installer that suggests to me that Latency within the context of Video over IP, is an overstated and overused principle! I would like to see the network which suffers almost zero latency - That would mean a high speed network backbone, with such capacity that there is very little risk of video data experiencing queuing delays. In an ideal world yes , the majority of networks could be like that --- my observation in the real world though has been that 90% of networks to which the new breed of IP Installers is plugging IP cameras - are not like that!

Okay, I think I see where the confusion is here: I think you're confusing two different uses of the term. When HDcctv boosters go on about "zero latency", they're not talking about network latency (mainly because there's no network involved anyway). When we talk about latency in CCTV systems (and IP cameras in particular), we're referring to the delay between when something happens and when you see it happen, which is introduced *BY THE IN-CAMERA PROCESSING OF THE VIDEO*. It has nothing to do with network or data latency. I don't see any confusion there: On a HDcctv system when you move the PTZ joystick, the camera moves. Do the same on an IP system and anytime you have more than a 250ms delay (Latency) it becomes very difficult to use. Restricting an examination of the comparison of latency in HDcctv and HDip to a PTZ scenario alone is of little value. Secondly , In order to view a decent image on a HDcctv system, I don't need to apply all the specialist network engineering concepts I would need to apply to get the same quality image on an IP system and I don't even have to start thinking about how much tolerance the corporate LAN offers. All those extra checks and balances on an IP system are necessary, because of latency. There are no such requirements or concerns on an HDcctv system. Of course - as soon as I connect the HDcctv recorder to the LAN, then I need to apply all those concepts any way! Network Latency is indeed an integral part of this comparison: (1) PTZ function on an analog system vs an IP system is a tangible effect of latency vs low latency, (2) The engineering concepts that have to be applied to the challenges of sending video over ip, as a result of network latency, are largely not issues in an analog system. So I don't think I am confusing the terms. Network Latency has tangible effects, that impact a comparison of Hdcctv vs HDip - both directly in terms of upthewire hardware coms, and indirectly in terms of the added complexity it necessitates to ensure video transmission over IP.

I will just mention two basic facts for now --- you probably know this already, and just a necessary elimination - so hope you are not offended. The 192.x.x.x. address is a LAN address, and you need to establish the WAN or Public IP address of your router. Likely that's already being done if you have set up the DynDNS. If the DYNDNS is correctly configured you will get to your router with the the URL that you got from DYNDNS. If you type in that URL and get to your routers logon screen, all is good - if it doesn't get to the routers Logon screen, something is amiss at this layer. Normally, in a default setup you will not be able to access the system from inside the same lan using DYNDNS. use a computer that is not on your lan!

Now we are getting somewhere. How often is zero latency required? - BUT , please remember it is not zero latency we pursue, it is nothing more than 250ms. Can you expand on how IP has better system scalability than HDcctv? is that assuming that a LAN already exists? is that assuming that the existing lan has sufficient capacity? Likewise --- with design flexibility? And can we expand on the better system options HDip offers over HDcctv? Zero Latency aside ---- it takes a lot of expertise to ensure effective video over ip, a lot of variables that are just not required in a Hdcctv system - unless we interface it with a LAN. I suspect that most IP installs haven't even begun to address those issues, and are just connected and operated on a prayer~

I don't think he ever said it was an irrelevant term, but in the whole scheme of things, this is CCTV we're talking about, not pro gaming. Whether the response time of a camera is 250ms or 1000ms, once a human factor is involved, then it becomes an insignificant amount as a human response time is likely to be depending on event anything from 5000+ms. I guess I am not getting it ---- human response aside! We are talking about the apparent fact that there is higher latency on an IP system than an HDcctv system. Are you saying that because we are talking about VSS that latency is not significant? Generally the human factor could be counted equal on an Ip system or HDcctv system, so let's discard the human response factor and make an apples versus apples comparison. I gather that what you guys are saying is "Yes there is a difference but it is insignificant?" is that what you are saying? I would be surprised if that's what you are saying .....because The Issue of latency is regarded as significant enough by numerous standards generating bodies and network engineers , in fact so significant that numerous techniques have been developed to try and manage it. I think it is a generally accepted fact that carrying video over ip presents a lot more challenges than most other non-time-critical IP traffic. Network engineers have deemed it necessary- in most professional systems, to deploy some kind of "forward error correction" to ensure that the encoded video stream can be reconstructed even with low probability packet loss and bit error amplification. A video system is not very tolerant of lost packets. Such systems demand high speed networks and data shaping for video, data prioritization and other traffic engineering approaches as well as systems to ensure that the network does not get too close to maximum capacity. Even then there still remains delay, and buffering is often employed but to the detriment of the video system. Once you start to exceed 250ms delays (and believe me that's a minute number) operation of PTZ controls become difficult. Latency variation has a significant impact on synchronization issues, gen locking techniques have to be employed to manage this. Even with all these engineering techniques, the network must have sufficient capacity to carry the entire content with a reasonable expectation of packet loss. Video over IP will not work if the network is congested , as is possible on multi-purpose corporate lans. So we have techniques such as quantity of service, dynamic frame rate or resolution reduction, network admission control, bandwidth reservation, traffic shaping, and traffic prioritizing. Are all these complex network engineering systems and techniques deployed for something which is insignificant? When we use the fact that we are only talking microseconds to imply that such small windows are not very significant, we overlook the fact that even 250ms is actually a very short period of time ----- but significant enough to degrade the operational efficiency of a PTZ operator.

Thank you kindly for taking the time to post the resolution - so often users don't come back with a resolution after many users have tried to assist them. These sorts of questions and resolutions tend to add to everybody's repertoire - thanks again!

The difference in latency is just one aspect of the HDip vs HDcctv comparison. Are you saying that latency then is a contrived and irrelevant term? That a ping of 400ms vs 1000ms for instance is insignificant and has no impact on performance when it comes to video transmission? Is latency only of significance to the IT guys when they transmit data? Is latency unimportant when synchronizing video and audio, or synchronizing other aspects of video security? I can't really provide a definitive answer around the latency issue --- so it does reduce the variables a little if someone comes along and says - "nah latency is insignificant in the broader scheme of things" Should we just then remove latency as a comparative factor and focus on the rest of the issues? I have had situations where controlling a PTZ camera has been extremely frustrating over a network, on which the numbers where not even 100 or so ms perhaps that has shaped my perception. Perhaps that kind of lag is just part of the process , but doesn't really cause any harm other than annoyance?

Hi & Welcome --- I am sure somebody will help: What is the hardware - you are using. Brands, Models, etc.....and what type of modem/router do you have?

Help required with the POS function of the netvision dvr. Our integrators have seemed to grasp the concept of TCP/IP POS quite well ---- but they are all battling to get this to work via serial. Has anybody achieved multiple POS integration using serial coms on the netvision DVR? What problems where experienced and how where they resolved? Thanks

I do , but I would like to get more feedback from others first as to how they have managed to view the videovista formats? before sending you to a professional program!

In the Netherlands 90% of the DVR's are connected to LAN/WAN. I mean light sensitivty. The HDcctv camera's does have bigger chips and therefore should be more sensitive for light the the smaller MP IP camera's. Unfortunatly the HDcctv aren't more sensitive. Also the comparison I made showed a more grainy image for the HDcctv camera's then the MP IP camera's. So you mean HDcctv is the new standard for those 10% of DVR's that won't be connected to the network? Isn't that a little broad - I mean I have seen HDcctv cameras with 1/2" and 1/3" sensors some with CMOS and Some with CCD, and I have seen HDip With the same. For instance ---- I am battling to understand how a HDcctv camera with a 1/3" Pixim CMOS sensor ends up with less light sensitivity than an Hdip camera with a 1/3" Pixim Cmos sensor -- i can't make the link between why that should be so.....are you saying it is so? That's interesting!

Thanks, nice to see some common points emerging - which always provides tangible areas to focus on.

Hope this is the correct forum. I am interested in understanding the major - dealer originated problems, that CCTV installers and integrators face. What has driven you to find another supplier in the past? Thanks in advance.

Are there still installers using standalone DVR's -- completely detached from IP and not offering remote access? Are they a significant group - if any still exist?? Could you expand a little on sensitivity? Do you mean light sensitivity? Are you using similar sized sensors in both, similar type sensors? This is something I haven't come across before - HDcctv cameras are less sensitive that HDip cameras, and would really like a little more information about that?

What software's have you already tried? VideoVista professional...?

hi. so thats what you are asking questions like that How does the unit compare on user friendliness: is the unit easy to setup in comparison? How do you rate the CMS/Network software in comparison Hmmm..... does it seem that unclear? Aren't these normally the types of questions one would include in a quality survey form, but I would also like to include vendor service issues in the questionnaire as well. How does our stock level/availability compare to other vendors? Stuff like that --- so that I can send the Form to my customers and ask them to assess our performance.

Hi All, I am looking to compile a quality questionnaire for our integrators, revolving around our range of TVT DVR's - which currently form the mainstay of our analog range. There are basically two chains of thought around quality: The product must exceed the standard of most every other product - which I think is a little aggressive, and the product should compare well with the standard of similar products - I prefer this second definition. Specifically the survey will be aimed at integrators and not end users. So I am looking for a little help as to comparative items to include in such a survey/questionnaire. For example: How does this product compare to similar products price wise? How does the perceptual resolution of the product compare to similar products? I would be very interested in any additional suggestions as to points of comparison one could include in a dvr quality comparitive survey - from an integrators perspective, dealing with a Vendor. So the items would include issues such as support and backup comparisons.....thanks in advance. I am sure the completed product would prove useful to all vendors of dvr products!

Yep, often times we also reverted to 24V when power over UTP was concerned. Sometimes just pure 24V cams and others using 24V/12V step down! Nice tip there numb-nutz! Thanks...

Are we talking about HDcctv or HDip. Are we measuring "one way" or "round trip" latency? How about "minimal latency" - the minimum potential latency of every link in the network, added together? Are we purely measuring or perceiving true latency, or the combined effect of network latency, queuing and processing delays? Are the cameras being deployed where precision or immediate intervention is required? What codec is being used to decode, and display the HD video? On average we are talking about 5 microseconds per kilometer over a fiber connection. Generally that should not present a problem , but how about viewing an HD camera from the other side of the world, or from thousands of kilometers away? So what transmission media is involved in your case - cat5, fiber? and what distances are involved? What operating system are you using? and is the PC dedicated to VSS, or multi-tasking? OS are not real time systems but work on queues and priority, the more executables that are being executed , the higher potential processing latency on the system - without even starting to transmit via a network? I believe that the timing between hardware and commands on a windows PC is less than exact, so how significant can a delay of 24/1024 transitions be in a precision system --- if you need precision? Such as synchronizing POS text overlay to the recorders clock to get accurate time/text overlay for example. Basically ----- there are many factors that might contribute to perceived latency on a HD camera and many of them do not originate with the fact that a camera is HD. Correct Codecs, real time low level OS, etc and numerous other configs can significantly impact upon the degree of total latency returned by a system.

The system that ak requested got me thinking about frame rates. Is 30fps ALWAYS a requirement. Is 15fps ALWAYS ineffective? Of course those issues are very relevant when those specs are called for - but what percentage of installs, truly demand 30fps, and then what percentage of cameras within any install should be 30fps? If the design objective called for monitoring, without recognition, identification, forensics, etc --- is HD or 30fps a real requirement? What percentage of cameras in the average install are designated Forensic/Identification/Evidential Recording as opposed to Monitor, detect, or recognition for instance?

How would they know? WHY would they know? There are two ways you'd even NOTICE latency: one, controlling a PTZ and having the movements lag behind your actions; two, sitting and watching the same scene you have on camera and seeing a visible delay. The former is (outside of a few specialized areas, like casinos) generally considered to not be a serious issue below a couple hundred milliseconds, and little more than an annoyance up to one second... the latter is just silly. In active surveillance, "latency" in the actual response time of the observer (ie. the time it takes for the guard to realize there's something going on, get off his ass, pick up his radio, and call in a response) is going to be probably dozens of times that of the delivery system, making even a second or two delay in the video itself irrelevant. In passive surveillance, probably 90% of the time, you're playing back recorded video, equating to a "latency" of minutes to months... again, even a second or two delay in the video transmission becomes wholly irrelevant. Hi, your input appreciated. Thus far it would seem, based on the responses, that latency is a pretty important issue to some. So having the opposite views arise is also important. I think when one discusses latency on it's own, it is less effective - in the context of this discussion. The discussion seems to lend itself to the value differentiation between Latency within an HDip system compared to Latency within an HDcctv system. I didn't quite get your last statement though. People usually make a conscious decision to deploy "Active" or "Passive". When Passive surveillance is "chosen" it should be from the understanding that the specific target or issue of concern does not demand an immediate response, or intervention. Passive surveillance under these conditions is still considered an effective management solution - exactly because the issues and concerns can be viewed at a later opportunity. When Active surveillance is "chosen" it usually costs 10x more than Passive - why? Because the issue of concern is deemed to be more significant, or critical. The very decision to deploy active surveillance implies an intent for immediate, live intervention. An unmanned Active PTZ camera is imo far less effective than a passive camera at detecting a specific event - so I am figuring an Active solution where a motion detection or other form of alert from beams, analytics, pir's requires immediate intervention by the control room operator. The scenario with the guard that you mentioned, would seem to me to describe either a poorly defined Active install, or a passive install where an immediate intervention was required. I am not saying that the delay is significant ---- I am asking for input, and opinions as to what the perceived value gap between a HDip and HDcctv system is - based upon the latency issue?

Hmmm it seems my post about Sony Vegas and Adobe After Effects didn't submit. But anyway --- wireguy has it spot on.......zooming won't have much effect if your source input doesn't have any extra pixels. Basically if you have a 380TVL source and you are trying to zoom it, or increase the image size to 4cif - (D1 Technically not correct on a DVR) it isn't going to help you any! Think of resolution as pixel saturation per image area. so you only have a given number of pixels, at 380TVL you are only going to saturate a given size of image to create a decent quality , likewise with any resolution ---- it's not much use trying to stretch the image to a bigger area , if the input source doesn't have sufficent pixels ---- as you stretch the image you are going to actually start seeing pixel blocks.

Interesting - my experience has been that we don't ever use single inline baluns, and probably will never send power over UTP again. We rather use a setup where we have distribution blocks of 4 or 8 channel baluns and co-axial drops from the balun to camera. Inline baluns are a nightmare during fault finding and repairs - especially if height or access is a problem. I think RJ45 connectors on balun may be a bit more desirable than the cheap screw type of clip on connectors . The power drop over UTP is astounding , and having to use duplicate pairs, kind of detracts from the appeal of UTP. Thats just the way we do it though, and maybe time have changed ---- maybe the baluns are not so finicky any more?

Chinese manufactured! IR creates heat, heat and cold variances creates condensation. The IP66 rating refers to external moisture entering the casing, i imagine. Condensation forming at the interface between hot and cold environments is a natural phenomena - maybe IR is not the way to go!

Hi EarlIT, I think you are liking the wider field of view. But yes I hear what you are saying --- and specifically HD in a 720p format or UXGA format is going to give you more coverage/forensic detail resolution per width of area. That is going to translate to less cabling infrastructure for pixel depth coverage compared to analog. But really , for a 1m x1m area , those $150 CNB cameras are very likely going to give you sufficent pixel/area coverage to provide identification detail surveillance and give you a damn good picture at that - to the extent tha spending an extra $100 might not be justified by some --- just because they like the image better!