IP Camera 2009 Mid Year Review

Published Jul 27, 2009 00:00 AM
PUBLIC - This article does not require an IPVM subscription. Feel free to share.

Megapixel was the most dominant trend in IP cameras over the first half of 2009. So dominant was megapixel, that not only did most manufacturers release new megapixel cameras but a new non-IP megapixel standard was proposed. Not surprisingly, industry people voted megapixel as the overwhelming choice for having the most impact on the video surveillance business.

Megapixel Product Explosion

In the last 6 months, megapixel has transitioned from a product that only a dozen or so manufacturers offered to being available in almost every video surveillance booth on a trade show floor (this was literally the case at SecuTech Taipei this April).

Non-IP Megapixel

In the mix of this, the recent announcement of the HDcctv Alliance, a trade group advocating non-IP megapixel standards has created quite a stir. Many (most?) of the powerful IP manufacturers are dismissing this as too little, too late. On the other hand, some traditional analog manufacturers have started to rally around it. For an overview of this, see our HDcctv introduction and our HDcctv competitive analysis (both featuring extensive community debate).

IP Camera Testing

We launched our IP camera testing program. The first of its kind, they are independent tests that provides extensive training and analysis of the product's performance. [See our IP Camera Testing Guidelines for background information.]

IP Camera test results released include:

Additionally, we are currently in the middle of testing Cisco's HD camera, IQinVision, Mobotix, Sanyo and VideoIQ.

The Next 12 Months and IP Camera Standards

Upcoming 720p PTZ under $1,000 USD

Darim has an interesting megapixel PTZ coming out soon (see brochure for pictures/details). Using piezoelectric motors, the camera is significantly less expensive than normal PTZs. The MSRP of the Darim camera, model SPT-100, will be under $1,000 USD. 

There are a number of important limitations of this product compared to normal PTZs. (1) There's only a digital zoom and (2) the pan/tilt speed is much slower than speeddomes - about 1/6 the speed.

With these specifications, the camera is suited more for indoor use (because of the limited zoom range) and where live tracking of suspects is not critical. However, it could be an interesting niche product where some re-positioning is needed but the cost and features of a traditional PTZ are over-kill.

Axis on SaaS

Axis talk on SaaS at ISC West [link no longer available] was recently released. You can listen to the audio as well as download the slides. It's about 45 minutes and you can skip the first ten as it's a general introduction.

SaaS/Managed Video is not only a hot trend but Axis is a major proponent of it. This webinar is worth watching.

Webcast of Pelco  PTZs

Pelco has released two very useful webcasts on their product lines - one on the Hybrid DX8100 DVR [link no longer available] and the other on Spectra IV 2.0 [link no longer available]. Both provide significant information but are short (maybe 10 minutes or so each) and can be watched on demand.

NVR Support Details for ArecontVision

ArecontVision has released a matrix listing their NVR partners [link no longer available], the models each supports and whether it has been verified by Arecont. There are over 40 NVR partners listed, representing a large portion of the market including both newer IP video software providers and larger incumbents including AD, Bosch, March, NICE Verint, Vicon, etc.

FLIR on Thermal Cameras

In the recent Security Products magazine, FLIR published two articles - a Q&A and a survey of applications using FLIR cameras. The most note-worthy claims FLIR makes in these articles: (1) costs are being driven down to high volume commercial level, (2) FLIR is aiming for "IR everywhere", (3) thermal improves video analytic performance and (4) thermal provides crisp images in total darkness.

FLIR's marketing has been focusing on lower price, broader use for some time now. Currently, even at about $3,000, it's still quite expensive for security use (especially compared to fixed cameras with an IR illuminator (costing about 1/3rd that price). It will be interesting to see how far thermal camera pricing can drop. Also, a frequent complaint about thermal is that you lose the details of people in a thermal image. While thermal may provide 'crisp images', many specifiers still prefer the greater details provided by D/N cameras with IR illuminators.

June 26th

iCanTek Company Overview

iCanTek is a Korean manufacturer of DVRs and IP cameras. They are better known through some of their OEM partners including Securitas Direct [link no longer available] (in Europe),  CCTVStar (in the US) and a large North American manufacturer that I was asked not to disclose.

They provide their own VMS - NVR Pro - which is free for their cameras. Additionally, they are supported by a number of VMS providers including Milestone, OnSSI, digifort and ipvision.

The product lineup seems solid but not many features or functionalities stand out as being unique or different. One product that iCanTek emphasized is the iCanView 290optimized for alarm monitoring applications.

NoblePeak Company Update

NoblePeak manufacturers cameras for low-light / no light viewing. Winning the 2008 ISC "Best in Show", the company received significant attention last year.

NoblePeak reports that they expect cameras to be available in production in 6 months (towards the end of 2009). 

NoblePeak is pursuing an OEM business model where they provide their camera cores [link no longer available]for inclusion in 3rd party cameras. While public pricing has not been finalized, they are targeting a price point similar to lower end FLIR thermal cameras. Relative to thermal cameras, NoblePeak claims that they can provide clearer, more detailed images at similar low light/no light levels.

No Subscription Alternatives to Managed Video

This week, DLink released 'plug n play' IP cameras as part of its MyDLink offering (see the product page for the MyDLink wired cube camera [link no longer available]). Essentially, these cameras can be viewed remotely without configuring port forwarding or setting up DDNS (similar at a high level to Starvedia's cameras that we reviewed recently). 

From a business perspective, the growth of these 'plug n play' cameras increases the alternatives to managed video / subscription services. To the extent that small businesses can use IP cameras without complex setup and monthly cost, this could cut into the potential growth of managed video. Of course, these 'plug n play' cameras do not support 'plug n play' recording so that still will be an important limitation.

Mobotix Panoramic Camera Demo

Mobotix continues to promote their second generation panoramic camera with a new video demo of the Q24. It's referred to as a 'review' but it's from Mobotix's Installer magazine so it's unlikely to be objective. Nonetheless, if you are interested to see what the unit looks like and some sample shots, the demo is worth a look.

Also, Mobotix has released their user manual for the Q24 [it's a 40 MB download].

June 19th

Can ONVIF lose?

In conversations with over a dozen manufacturers, the consistent theme I hear is a preference for ONVIF. Interestingly, the reason is always about ONVIF's fame/market share. Second most common point is Cisco joining ONVIF. Regardless of why this happened, the negative PR impact on PSIA from this is clear. 

The only two reasons I can see for ONVIF to fail at this point is (a) there are technical problems (i.e., implementing ONVIF's SOAP format is too much load for lower end cameras or (b) ONVIF delays. PSIA has lost momentum and unless significant numbers of PSIA members release production products soon, it's hard to see why PSIA survives.

Depending on how much traction HDcctv group gets, this may be a factor in spurring the ONVIF and PSIA to merge to strengthen their position against a common rival.

June 6th

Axis Timetable for ONVIF Support

Axis told me that they plan to release ONVIF support to their camera firmware in Q3, 2009. This will be in addition to their existing VAPIX protocol.

If Axis does release ONVIF support in Q3 and it works well, this would have a major impact in standards development. Given international preference for ONVIF/Axis, a timely release would likely mean ONVIF wins.

Axis also reports that they will not add support for ONVIF to Camera Station, their VMS software. Axis' decision here is important because if they did add ONVIF support, Camera Station would become much more of a competitor to 3rd party VMS providers.

In other ONVIF news, Bosch reports that ONVIF 2.0 may include specifications for accessing recorded video. If and when this happens, this could be valuable (1) for supporting on-board storage in cameras and (2) to support DVRs (which I think would help significantly in allowing customers to migrate from DVRs.

Training Videos from IQinVision

IQinVision has released a series of videos on their 4 series dome including an  4 overview video, a review of what's included, and a description of the camera components.

These videos are useful for the community to learn about the technology and product offering. It's a low cost way to train and market products.

May 29th

Will P-Iris Improve Megapixel Image Quality?

Recently, Axis announced an upcoming product enhancement that aims to improve megapixel image quality. Called P-Iris (for Precise Iris), it is an alternative to DC auto-iris lenses. P-Iris is a combination of a new lens and software. The lens, provided by Kowa [link no longer available], provides finer grain motor control to more precisely position the iris. The software, provided by Axis, allows the Axis camera to gain feedback on the positioning of the lens and to adjust more precisely. Today, most megapixel cameras use manual iris lenses. The P-Iris lens will be available in the Summer-Fall time frame. At that point, we will be better able to assess actual imagining improvements.

Axis made a point that they are not patenting any of the technology. They are encouraging third parties to adopt and use similar technology. It will be interesting to see how this develops.

Comments on ACTi Strategy

A new press release from ACTi demonstrates 3 interesting characteristics about ACTi's strategy and market approach. The press release is about IFSEC but it covers their general market focus:

  • ACTi acts like a western company: They talk about how amazing they are and the blockbuster products they are delivering. While this is (unfortunately) typical US behavior, this is surprising compared to other Taiwanese companies. The typical Taiwanese approach is understatement and focus on not offending potential OEM partners. ACTi wants to compete head to head against companies globally.
  • ACTi's goal is to drop prices while matching Western companies' product quality. They say as much -- "lowering the price while delivering even better video quality and all essential camera functions."
  • They are emphasizing new products that improve video quality like the TCM-5311, a 1.3MP camera that offers Sony's new CCD to improve low light image quality.

There are certainly plenty of risks involved in this strategy but with their lower cost structure, they have a legitimate shot of causing more disruption.

Megapixel Cameras from 3xLogic

Continuing the trend of new manufacturers offering megapixel cameras, US vendor3xLogic just announced the release of a series of 1.3 and 2MP domes and box cameras (see datasheets on MPEG-4 2MP and 1.3MP domes). The price points are competitive with ACTi/Arecont level pricing.

I think 3xLogic is a company worth watching (I profiled 3XLogic a few weeks back). They are offering pure hybrid DVRs at low prices with strong support for megapixel. That's an attractive combination in today's market.

May 22nd

3 Key differences found in Megapixel H.264 camera tests

So far, I see 3 important differentiators among the cameras I am currently testing:

  • Effective or 'true' resolution: The stated pixel count (1.3MP, 2MP, etc) is proving to be a misleading metric to judge cameras. More pixels often do not mean better image quality. This is a critical element as better image quality (the ability to see smaller details) is the key driver in the use of megapixel.
  • Bandwidth consumption variation: I see wide variances in how much bandwidth is consumed, both depending on the scene and, more interestingly, between manufacturers. Some manufacturers are using 2Mb/s for a given scene while other need 6-8Mb/s. This could be a major factor in choosing cameras as the differences in storage could be quite expensive.
  • Handling low or changing bandwidth levels: Some cameras have significant problems with insufficient bandwidth. The image will break up almost entirely under these conditions. It is essential that surveillance cameras are robust to lower bandwidth. While it's obviously understood and accepted that with less bandwidth, either frame rate or quality will suffer somewhat, the connection and the basic image must remain.

These are certainly not the only 3 issues - just issues that generally receive less attention. Additionally, I am still in the process of investigating CPU utilization.D Finally, low light/WDR remain important issues for megapixel as they do for all surveillance cameras.

Why IP cameras force analytics to the edge

In a recent Pelco webinar, Pelco offered strong technical points about why analytics have to go to the camera.  Performing analytics requires decoded video frames. With analog cameras, the encoder card in a DVR can provide decoded video frames instantly and with little overhead. With IP cameras, if you use MJPEG, a similar result occurs. However, with interframe codecs such as MPEG-4 or H.264, you have to decode the video. Pelco explained that MPEG-4 and H.264 significantly this reduced the number of cameras that a server could analyze.

the future of video surveillance belongs to these CODECs because of the bandwidth/storage savings. As the migration to IP evolves, this will increasingly penalize and cause problems for providers of centralized analytics.

Is SVC the next big CODEC?

When asked what is next after H.264, Pelco listed SVC as a leading choice. SVC, or Scalable Video Codec, can dynamically change the resolution and bit-rate of video streams. This is extremely important for security applications as so many viewers are in remote locations. It's is especially relevant to megapixel cameras that stream at very high bit rates (relative to WAN bandwidth) and can easily cause problems for remote viewing.  For more information on SVC, see an overview on SVC posted on the site last year. It's going to take some time for SVC to go mainstream (as it requires more processing power but it will address the increasing pain as megapixel camera use expands).

In the meantime, there are a number of techniques video surveillance vendors are providing. One popular technique is for the NVR/VMS system to transcode video before it streams it to clients, enabling the bit rate to be dynamically adjusted for remote viewers (see 3xLogic's Aztec CODEC and Salient's Dynamic Resolution Scaling [link no longer available]). Another alternative is Avigilon's use of JPEG2000 which provides scalable video but lacks the bandwidth efficiency gains of H.264.

Baseline vs High Profile for H.264

A final interesting note in the Pelco seminar was Pelco's advocacy of High Profile H.264. H.264 provides a variety of profiles and levels [link no longer available] (H.264 is more like a family of specifications rather than a single specific one). Manufacturers may select from these profiles. Most IP camera companies are choosing baseline profile - the lowest of the options. Pelco says they have selected High. It is reported that high provides greater bandwidth and storage efficiency at the expense of increased processing power.

How much of a difference this makes will be interesting to see. I do, though, expect Pelco and a few other vendors to emphasize this difference in the marketplace. I am looking forward to testing the Pelco Sarix camera to investigate this aspect.

Genetec vs Mobotix on H.264's Value

A few months ago, a Mobotix partner posted a video criticizing the use of H.264 for centralized video recording. 

Recently, Genetec has started its own blog [link no longer available]. In one of their first posts, Genetec examines the value of H.264 [link no longer available] and responds to these criticisms.

May 8th

New Panoramic Camera from Mobotix

Mobotix is announcing to dealers a new version of their panoramic dome series - Q24. Mobotix states that the 3MP camera offers 8fps at 3MP, double the frame rate of the old Q22 as well as enhanced panoramic correction and a built in 4GB SD card. They claim very high market demand and are offering an MSRP of $798 EUR ($1070 USD) for the 3MP version.

With Grandeye/OnCam refocusing on a niche market and very few alternatives (Immervision lenses), Mobotix has a fairly strong niche here.

Mobotix in Search of Acquistions

VB Research also reported that Mobotix is seeking acquisitions. Mobotix's report cash of $13.7M USD. They are looking for complimentary technology or to expand product development with VoIP being a particular area of interest. Mobotix noted that only 10% of their 2008 annual revenue of $48 M USD is from the US. This is not surprising given how famous Mobotix is in Europe and Asia but relatively unknown in the US. Mobotix is looking to ramp up US sales.

ACTi maps out H.264 products

By the end of 2009, the expansion of H.264 across the IP camera world will be fairly complete. In this pdf document, ACTi literally provides a map of their various new and upcoming H.264 offerings.

Access Control From IQinVision

IQinVision has announced a small scale access control software offering to be used with their Megapixel cameras. It works with their PRO series of cameras and allows users to open doors, activate alarms, etc from a user's PC. This solution is not meant to replace access control systems in traditional applications. Note: I am obtaining a copy of this software to test.

This type of solution appears popular with megapixel cameras. Compare to Mobotix'sExtIO and CamIO

May 2nd

Current Level of Megapixel Commercial Availability

It's become clear that 1.3 MP, H.264, 10 - 12 fps cameras are now everywhere. At SecuTech, basically every camera manufacturer, even the laggards had such cameras available.

Also, 2MP, H.264 cameras are destined to be widely available in 6 months or at latest by ISC West 2010. Already, at SecuTech, there were at least 8 manufacturers announcing 2MP H.264, in addition to the half dozen global companies that had previously announced last month at ISC West.

What's driving this is the ready availability of DSPs and reference designs. What will be interesting to see is what the next jump in resolution is after 2MP. It may take some time as the 3MP and up providers all seem to have developed their own ASICs or FPGAs.

The other important aspect to watch in the megapixel space is how manufacturers start optimizing low-light performance and WDR to improve the image quality of their 1.3 and 2MP cameras.

April 24th

ACTi Partners with PSA

ACTi will now be distributed through the PSA, a large co-operative of 200 American [link no longer available]integrators (total annual sales 1.7 Billion USD). 

This is likely to further help ACTi's expansion in the US at a key time where traditional security integrators are looking to move to IP.

ACTi releases a CCD 1.3 Megapixel Camera

With megapixel becoming more commonplace, manufacturers are starting to focus on improving image quality and nigh-time performance (for instance, Panasonic won a NPS award for its CCD megapixel camera). Now, ACTi has released a 1.3 MP camera using a Sony CCD. Pricing is not publicly available yet.

April 17th

Advances in Megapixel Form Factor

In the beginning of 2008, there were very few form factor options for megapixel cameras. Almost all of them were "box" units (except for Mobotix). Form factor options are important for environmental and aesthetic issues.

Since then, we have seen a lot of advances - Axis now has multiple megapixel dome options (including the new 720p dome). IQinVision has 3 lines of advanced form factor options (including the new 4 series). Basler has entered the surveillance market withtiny megapixel cameras [link no longer available] (even smaller than Arecont's).

This is another sign of the maturing of the megapixel market.

Cisco Releasing Documentation on their HD Cameras

The Cisco HD cameras were announced 7 months ago at ASIS. Recently, they released their 84 page User's Guide.

Also, a price check on the 4300 model on Google shopping reveals the current online price is about $1100, $300 less than it was a few months ago. Compare to the Axis Q1755 whose online price is about $1500.

April 10th

Pelco IP Resources: Pelco has released a series of new resources on their IP products including a webcast on Endura, a webcast on Sarix and an article on Endura. The webcasts are both particularly detailed so if you are looking to learn more on the product lines, I recommend.

Mar 27th

Examining Basler's Megapixel Cameras

I spoke with Basler this week. They are a machine vision company that is increasing its presence in the security industry with a line of megapixel cameras [link no longer available]. They claim better image quality using CCD and H.264 on all cameras up to 2MP (1600 x 1200). The MSRP on their 2MP camera (BIP-1600c [link no longer available]) is $1100 USD (H.264 with maximum  of 3 fps). That's fairly inexpensive and if you want H.264 and are ok with low frame rates, this could be an interesting option. Basler is supported by a fair number of VMS vendors including Milestone, LuxRiot and Aimetis.

Examining Pelco's Partnership with TechData

Two weeks ago, Pelco announced that TechData would be their exclusive IT distributor in the US. I spoke with Pelco. They said that TechData will be only be selling Pelco's IP product lines - no analog cameras or traditional DVRs. They also said that TechData would not be selling to end users nor would this change Pelco's overall channel strategy.

Mar 20th

Axis mini-encoder details and pricing 

Last month, Axis announced a small form factor, single channel encoder. They have added a data sheet [link no longer available] on it. Also, online pricing is now available - about $300 USD. That's a pretty reasonable price for a H.264 30 fps 4CIF encoder and there are not a lot of alternatives for that form factor. Interestingly they are also offering it as a kit for convert surveillance (online price about $400 USD). Since it takes analog inputs, you certainly could choose from a wide variety of spy cameras on the market.

Mar 6th

Pelco's Sarix Megapixel Cameras

Pelco is about to make a major marketing push around their new IP/megapixel line. This week, we offered a detailed breakdown of Sarix and contrast it to existing market leaders. The line is targeted at the high end of the market and has some impressive features but will face significant challenges in displacing market leaders like Axis.

New developments in IP camera standards debate

TechSec featured a debate on IP camera standards - ONVIF, PSIA, etc. Sam has a long writeup and the discussion on standards is at the very end of his post. What's most shocking to me on this discussion is that the claimed material differences between ONVIF and the PSIA is a difference in communication format. (The REST/SOAP debate is a fairly low level technical issue and if you care to understand it better, see this writeup on REST v SOAP.) Most developers these days prefer REST but the reality is it is not a significant material difference. If this is all that is really holding up resolving ONVIF v PSIA, then there is little reason not to resolve this immediately. If they don't, it's a sign that this is more about vendor positioning than it is a genuine difference in vision or functionality (which is what informed observers have believed from the beginning). Sam now has a transcript of the entire debate available.

Immervision's panoramic lens product positioning

From a recent briefing, I found Immervision's positioning to be interesting. Immervision makes a 360 degree lens for using with existing SD and 1.3 MP lens cameras (demo on their home page). Immervision does not claim they replace multiple cameras. Rather, it's an improved camera that can provide basic action identification over a large area. It won't show the details on someone's tattoo but it will clearly show how someone moved through a large area. Not revolutionary, of course, but likely to work well in this application. Key issues are cost ($749 MSRP for the lens) and 3rd party VMS compatability (still limited).

IQinVision's New Edge Recording Solution

With an upcoming release, Certain IQinVision cameras can record locally without the need for network bandwidth or a third party management system. Their Event Recording System solution is offered as a kit with storage and software license bundled together. It will be a modest premium (cost for the storage and software license) over existing cameras.  It is mean for small camera deployments. While it has some similarities to Mobotix's de-centralized architecture, it is not designed nor intended to be a general replacement for 3rd party VMS (as Mobotix is). 
Designing Megapixel for Schools

Theia, the wide angle, megapixel lens manufacturer has an interesting diagram about how a school designed a video surveillance system using their lenses and megapixel cameras.  They used 6 cameras to cover an entire courtyard. It's a good real world case study in line with my recommendations on how to win with megapixel cameras.

Mobotix New Marketing Materials

Mobotix has a number of new marketing materials out, including: a revised company brochure and product portfolio. There's a marketing slogan that I think is particularly strong for IP advocacy: "It's not analog or IP but LowRes or HiRes. HiRes is only possible by using IP cameras." Unlike Axis, Mobotix can claim this because they are focused on megapixel where Axis has substantial product offerings and revenue from Axis. Nonetheless, I think it's this type of focus that will provide a boost in the mindshare for IP. It's probably something IP providers in general should focus more marketing attention.

February 20th

Axis on IP Camera Migration: This week, Axis released a presentation on migrating to IP cameras [link no longer available]. The presentation is basically a treatment of encoders. On the one hand, for most IP video surveillance software offerings, encoders are the only practical choice. However, migration plans would certainly benefit from hybrid recorders and the drive to put SSD storage in encoders is certainly a step in that direction.

AgentVi on Sony: AgentVi announced that Sony's IP cameras will now support AgentVi analytics. AgentVi's approach to analytics is different from well known providers like ObjectVideo and ioimage. AgentVi splits the processing doing some of the work on the camera and the rest on a server. Because of this, AgentVi can work on a variety of common IP cameras including many Axis cameras and now Sony. By contrast, ObjectVideo's analytics are constrained to cameras with higher end chips, thereby reducing the cameras that can run their analytics (e.g., OV is not supported on Axis nor Sony). Of course, ioimage only runs on their own cameras and encoders. 

February 6th

Axis Updated Product Guide

Axis has updated their product guide to include the new megapixel and H.264 cameras. It does a nice job showing the range of high quality offerings they have. Unfortunately, it also shows how lacking their product set is to budget products that are gaining ground due to the economy.

January 16th

Axis on HD

Axis released a number of detailed document on their new HD camera - the Q1755, including a whitepaper promoting HD [link no longer available]. They are emphasizing that their camera meets the HDTV standard while other 2MP cameras do not. I do not see the point here. Maybe it makes end users feel better ('true HD') but the business benefit is not clear (I asked, their answer did not help). Also, at $1495 USD MSRP, it's expensive (relative to other 2MP cameras). It has a 10x optical zoom which they claims "automatically zooms in when there is activity in the scene" but I am still trying to figure out the common applications where this really helps. See our review of the Axis Q1755.