Member Discussion

Will Dahua HDCVI Destroy The Prices Of HD Camera Market?

What do you think about Dahua HDCVI 720p Cams for under 50$.

Will they destroy the prices for HD Cameras in 2015?

What should be the reason for a customer to buy an IP Cam for few 100$?

[IPVM Note: background, see: Testing Dahua HDCVI, Dahua HDCVI 2.0 Tested, Hikvision HDTVI VS Dahua HDCVI]

Will they destroy the prices for HD Cameras in 2015?

It's not just the HD* analog variants, even 720P and 1080P IP cameras can be found for under $100 these days. When you figure that wiring costs and installation labor are mostly a constant, and exceed the price of a simple camera, almost any price under $100 is effectively the same, or effectively $0.

The 3MP and under camera game is pretty saturated, uninteresting, and not particularly profitable for the dealer. Manufacturers are still making pretty good margins in that segment, as long as they are not greedy.

Now that we have cheap moderate resolution (I don't call anything 5MP and under "high" res anymore, that's the standard resolution), a consistent connectivity model (ONVIF) and prices at a point where end-users don't have to limit their options as much on how many cameras they can afford I think we'll start to see the "smart phone" era of cameras. Just like basic cell phones raced to the bottom and manufacturers pivoted to making "smart" devices, we should start to see the same thing with cameras. Usable video analytics, "store and forward" edge recording, onboard basic LPR, stuff like that. Customers will/should start to expect more than just a good image from their camera, they should start to expect the camera to be more of an "active" participant in their corporate security.

On a semi-related note, I think some of the big names are starting to hate ONVIF. Previously these random overseas companies had a hard time competing because there were no drivers in the mainstream VMS's for their products. It was a chicken and egg problem, VMS's wouldn't support those products with no user base, and most users rightly wouldn't purchase something that came with a huge headache tax support-wise. Now with ONVIF essentially being the de-facto camera protocol, there is a unified way to discover, connect to, and handle basic settings for cameras.

Users can choose a camera based on cost/features and be less concerned about VMS support for it. This is good for the Dahua's of the world, and maybe bad for the Axis's of the world. And potentially really bad for the caught-in-the-middle big brands like Canon or Samsung. Their brand alone isn't strong enough in this market to carry a premium price and compete head to head with Axis, but their corporate overhead prevents them from being price competitive with Dahua. Their best move (IMO) is to dodge left and build a really feature-enhanced (usable features, not gimmicky shit) camera that is priced in line with the comparable Axis model. Auto-focus, adaptive IR, advanced motion detection, motorized zoom, SD card recording, etc. are all standard expectations, not features. All that stuff is expected in a $250 camera (the $100 cameras are usually fixed focal length cams with relatively wide (2.8-5mm) lenses.

"It's not just the HD* analog variants, even 720P and 1080P IP cameras can be found for under $100 these days."

Though those are the same manufacturers (i.e., Dahua, Hikvision, their OEMs) selling Analog HD and IP HD at such low prices :)

As a point of reference, for similar feature sets, analog HD to IP HD is roughly 2x - e.g., $60 for analog vs $120 IP HD. One can still argue how much of a difference this really is....

"When you figure that wiring costs and installation labor are mostly a constant"

Is that really? With analog, the process is less technical and I can send out a less skilled person, which means lower cost, in terms of hourly rate, total time and risk of something going wrong (firmware incompatibility, device discovery problem, etc., they happen).

There's lots of reasons not to use analog HD but I am not sure if lack of a cost advantage is one of them.

Though those are the same manufacturers

Right, what I was trying to say is that it's not Analog HD variants that will cause price pressure, it's that overall EVERYTHING is cheaper. Analog HD may be even cheaper than "equivalent" IP cameras, but ALL this stuff coming out China is dirt cheap, and acceptable quality.

Is that really?

Well, there's a whole larger philosophical discussion around this, but I think that even low-end techs need to have basic IP skills these days, and if the integrator somewhat standardizes on cameras and VMS things are very "scriptable" for the install. Some lesser qualified techs may not exactly understand the finer details of what they're really doing, but they can follow a set of steps and get an expected outcome. If things DON'T go as planned, yeah, the IP system is probably harder for that tech to troubleshoot in some cases.

A basic camera install then is essentially:

1) Pull some wire - more less same materials cost (are these AHD cameras powered over coax??) and same labor. Maybe slight time variance for punching down cables vs. crimping BNCs?

2) Mount/hang/wire camera - more or less the same process for each camera

Assume somewhere around 2.25 hours on average for "a camera", billed at an "average" rate of $75/hr (taking some huge interpretations of "average" here) and you've got about $168 for labor, plus some materials ($30/camera?), or right around $200/camera for labor/cable.

This means your all-in cost for a cheap AHD camera is ~$260, and your all-in cost for a cheap IP camera is ~$320, or a little less than a 25% up charge. If the cameras are all outdoors (more difficult/expensive installs) the variance might be less. If they're all indoors in a small shop attached to drop ceilings, the variance might be more.

After that part, configuring a VMS/IP system is probably more labor intensive than the relatively simple setup of an NVR, but the Razberri-style devices and similar "appliances" from some VMS companies are neutralizing a lot of that (albeit at a price premium compared to a simple DVR).

For the most part, I still think the AHD stuff is primarily limited to (or, geared to) simpler systems. Where the requirement for higher resolutions, wider variety of form factors, and advanced features are less of a selling point.

AHD then would do well in the DIY market, and the "gas station/nail salon/Donut Shop" sector. IP would do well in the "corporate" sector, large retail, etc. market. If the customer is super cost-conscious, the dealer might propose an AHD system in 2015. If the customer is performance-oriented, the dealer would propose an IP system and VMS, but will be more forced to justify the $800 fancy IP cam compared to the 3MP Hikvison dome on Amazon for $130

With analog, the process is less technical and I can send out a less skilled person, which means lower cost, in terms of hourly rate, total time and risk of something going wrong (firmware incompatibility, device discovery problem, etc., they happen).

In my experience, most integrators have pretty much the same skill level technician on all jobs. Sometimes they'll have one or two more highly skilled that will go in and do all the programming on a project (alot of times these guys are called "engineers") and act as internal trainer/tech support to the other technicians. Because of this, I think most integrators bid labor at a single standard rate as opposed to a skill level tiered rate.

Come to think of it, that might be an interesting survey. How many used a single rate versus a skill level tiered rate for labor estimates.

Analog HD (Dahua HDCVI, Hikvision / HDTVI, AHD, etc.) are definitely one of the biggest sales growth drivers for 2015 (see: Security Sales Forecast 2015). It will certainly put price pressure on IP competitors.

That said, it is not going to go that far as to 'destroy' pricing.

HDCVI / Analog HD's constraints:

  • Not enough manufacturers selling it, i.e. big brands are mostly not offering it now.
  • Manufacturers that do sell it often have no or poor support (e.g., Dahua direct in the US).
  • Analog HD cameras are pretty basic right now and few have the advanced features high-end buyers want (varifocal, autofocus, WDR, super low light, more form factors, etc., etc.).
  • Analog HD manufacturers have limited marketing budgets meaning a lot of buyers won't even know it exists.

All of those will mitigate price drops, especially from bigger brands. Also, big brands are generally cautious / conservative about cutting prices as it goes against their corporate philosophy. They typically will need to be hurt for a year or more before they seriously drop prices.

As for your final question, why buy more expensive IP cameras? As mentioned above, advanced features, more form factors, bigger brands, (often) better support plus higher MP, panoramics, etc.

One feature that I have seen from Dahua lately that is a sure sign they are going to be a serious threat to IP is the addition of "smart features" like face detection in their DVRs. If they continue to develop and offer value added features for CVI, I could see it becoming a much bigger threat to IP, other than obviously the very bottom end of the market. CVI and other analog HD variants will most likely kill the SD analog market, once and for all.

What does the "face detection" do? (I haven't seen/used it).

I've only read about it so far, but they claim to identify and index people based on facial recognition. Bear with me if I broke any covenants by misusing the wrong vernacular here. I'm still new to facial detection/identification/recognition. I know there are differences, but the lines are still blurred for me.

If they are claiming facial-rec indexing, they're full of shit.

I was thinking it might be something like simple face detection to scale recording (a form of semi-advanced motion recording).

I think your flat out denial of their claim is a knee jerk reaction. Maybe we should wait until it is tested before we jump to such hasty judgements.

Maybe you work for Hik? ;)

Jon, your link confirms his claim.

Dahua is offering face detection (i.e., find and index a face), not facial-rec / recognition (i.e., match against a database of suspects).

Maybe I'm splitting hairs here, but the way I read the press release, they are indexing faces and allow playback based on a given face detected. To me, that is a pretty cool feature and I stand by my original statements.

"iHCVR, based on Dahua’s in-house High Definition Composite Video Interface (HDCVI) technology, new generation of smart face detection which realizes 1ch face detection, instant playback and face index playback. iHCVR series products not only support face detection but also can realize face image snapshot and face image 20 seconds playback assisted by software."

To me, that is a pretty cool feature and I stand by my original statements.

Companies have struggled for *years* to do reliable face recognition. And we're talking about a scenario where the "database" face (the master image stored for comparison) is generally a high-res clean "mugshot" sort of image, taken at an angle to get maximum detail and an undistorted image.

Somehow, practically overnight, a company that specializes in making simple, cheap cameras and recoders has leapfrogged over all technology and created a $500 DVR that can reliably search faces captured at relatively low resolution and odd angles, and compare images quickly, and with an acceptable accuracy level?

This is the sort of thing where (IMO) you need to be able to call "bullshit" on something. Dahua did not just solve the mainstream facial recognition problem, and then bundle it in as an freebie feature in a product that arguably targets a low-end market.

This is where understanding the fundamentals of things comes in very handy. So that you don't misinterpret, or misbelieve, press releases like this. Otherwise you end up selling/promising something to your customer that you have no chance of delivering. (I offer this as helpful advice, I don't mean for it to sound harsh). The Speco and Tri-Ed reps are probably clamoring for your contact info right now...

What Dahua *appears* to be doing here is recognizing when a face is present in a video segment. The logic being that most of what we care about in security is stuff involving people. Shadows and lighting changes and stray dogs don't cause many problems. This is +1 beyond advanced motion detection, they've taken a fairly proven thing (face *detection*) which has been in consumer point-n-shoot cameras for nearly a decade, and rolled it into a security system. You can search for any video that contains faces, and you can set the DVR to upload 20 second clips of faces to remote storage (presumably on the premise that the face detected might be a thief who takes a hammer to your tin-can DVR and destroys the stored video). Dahua has solved a couple of problems (trying to deal with usable motion detection, and remote recording) in a practical way here. It's kind of poor-mans analytics, and is the sort of thing that really should have come about 5 years ago, but anyway...

Also, you'll almost never see the Chinese companies be first movers in anything. Their basic business mentality is essentially frugality and cost-cutting. They look to see what other companies are having success with, then they try to copy with a reduced-cost varient. They will typically shy away from unproven, unpopular or support-intensive stuff. You'll see facial *recognition* from Dahua, et al, 5-8 years after it's a mainstream offering from the "big names", all the kinks have been worked out, and is the kind of thing that customers expect by default, or major deals are won and lost over. This is why (IMO) you still don't even see any kind of real analytics from these companies, the global market there is still too small and the technology is not simple enough yet for Dahua to suddenly start offering something as simple as tripwire detection in their cameras.

Let me clear one thing up, I am not selling this feature. I wouldn't even mention to a customer without trying it myself, like I have done for years. I would possibly sell something if IPVM was very convinced of its abilities without using it first, but even then, I would have to have a compelling reason to do so.

So, my point was *IF* Dahua could pull this type of feature off and maybe other useful features that are usually only found in a real VMS, then they could claim more market share away from the IP market.

Face whatever has never made or broken any surveillance product, so safe to say Dahua HDCVI will live or drive regardless of what they do on the face front.

[Correction: face has broken many things but never made them...]

That is face *detection*, something which is relatively easy to do. They are claiming to spot faces in the image, and use that for what is essentially bookmarking and image upload.

That is difference than facial recognition, but *is* an example of what I was talking about in my earlier comments of the kinds of features users will expect in mid to high end IP cameras (for AHD, it makes more sense to have it in the DVR).

Overall, this is still of limited use, as cameras placed for any kind of facial detection/recognition/detail are often different than those placed for general security overview. In a typical camera FOV faces will be obscurred by hats/headwear, or just from people not looking/walking in the direction of the camera. Still, detecting the detectable faces (and bookmarking them) is better than doing nothing at all.

Facial *recognition* in a DVR attached to a basic 720p or 1080p camera with a typical 3 or 4mm lens would be much much more difficult.

"Face detection", in one form or another, has been around for a while. Modern digital still cameras often incorporate it for optimal focusing. It appears that Dahua has just taken that technology and applied it to bookmarks.

Now a real advancement would be Ugly Face Detection. There's an app for that.

"Camera Ugly Face Detector is a brand new Ugly/Beautiful Android scanner that allows you to know if you or your family/friends are beautiful or not.

It is very simple to use, just place your face or anyone face in the camera area and click start. Your mobile phone will analyze your face and will give you a result.
It is very funny!

And remember this is only for entertainment purposes."

At $130 for 4K*, Dahua is squeezing the market from both the top and bottom...

Yielding over 900 PPP (PixelsPerPenny), is this a record?

*thru Alibaba, Qty 1, Min qty 1 (Maybe Misprint? Others say $200-$400, still pretty low)

Welcome the United States of buying cheap shit. You will always have those US customers so focused on price that no matter what you have that May or may not be better will always come to price. As for myself and I would like to think that for other customers who will still want high quality products that there will always be a market for high end products that carry a larger price tag. Not that my statements have much to do with the exact subject but I do believe there will be some impact of HDCVI but not so much in the near future. Just my two cents.

One aspect of the newer analog HD that can keep customers away is the fact that it is still a proprietary interface. Some folks are put off by this and want something more standard.

Currently they do not have a HDCVI compression card that we can purchase to build out a hybrid solution. This means we will need to purchase DVR/HVR only from these manufacurers. This limits the choices for a larger installation that is a mix of IP and HDCVI.

Hopefully, these manufacturers will make the cards available to us like they did for the standard 4CIF/D1 cards.

Cards, hell. Before we would even consider buying into Analog HD, there would need to be high quality ONVIF compliant encoders capable of delivering multiple streams from each device. DVRs may be a stopgap measure for some applications but true enterprise-quality encoders with network and power redundancy are an absolute must.

All this discussion abouth HD-CVI cameras dropping the bottom out of the market is only half the story. From what I've seen, the DVR is the weakest link in the closed system. And, until Dahua opens up their technology to other makers, it is by far the least sophisticated piece of gear that I've worked with in some time. As a stand-alone DVR, it offers a clumsy interface and flimsy construction. HIKvision's HD-TVI is at least planned to be shared/licensed among a variety of makers.

As an encoder, an HD-CVI DVR definitely does the trick if only used as a bridge to convert & carry the HD-CVI signals from the 1080p or 720p cheap cameras into a sophisticated VMS. I have a few systems now in operation using Video Insight as the backend, and they perform way better than I had dared hope.

The near-term HD-CVI opportunity as I see it is two-fold: First is as an upgrade path to bring hard-to-re-cable cameras into a larger IP system where a traditional analog-to-IP encoder could be used -- for about the same cost, the coax-attached cameras can be stepped up from 650tvl max to the equivalent of 2MP.

The second scenaria is the conversion of a typical 4- or 8-camera analog 'package' that a trunk-slammer security dealer might have sold a small retailer or homeowner. There, the DVR operation is no big deal, but the image quality is improved 10x. From what I assume of IPVM's readership, this market is not interesting.

Michael, you can use the analog HD DVRs as encoders, see: HD Analog DVRs With VMS Software Tested

Yes, that's how I'm using them.

But, the Dahua DVR-as-encoder route brings a higher cost to the system than I would expect of a bespoke encoder. I fear the DVR also has more bits & pieces that could go bad, raising the risk of follow-on service. I'm looking forward to sample encoders coming from HIK, as they promise to support both HD-TVI and SD analog cameras. This flexibility would be nice to have available.

Both Dahua and Hikvision support DVR-as-encoder. And Hikvision's current DVR (which functions as a multiple port encoder) supports both TVI and SD cameras. And Dahua's newest CVI DVRs support CVI and SD analog cameras. We've tested both - see: Hikvision HDTVI VS Dahua HDCVI.

As for prices, those analog HD DVR (from both sides are quite inexpensive) - $500 or less for 16 channels. By not using the hard drives on the HD DVRs (treating them like encoders) you remove the greatest risk component.

That said, if you want hot swappable encoder blades with redundant power supplies from Axis, you may be waiting a long long time.

A waste of electronics, if you ask me. DVR-as-encoder equals:

As has been mentioned previously, they can be also loaded with disks to be used in a crude but serviceable VMS/storage failover scheme.

John, Dahua have announced they are releasing a range of HD-CVI encoders. Our software engineer advises us integration to these encoders via the Dahua SDK to a VMS for recording is far more effective and reliable than adding a HD-CVI DVR as an ONVIF device. What is your view on the pros and cons. We are not aware of Dahua recommending and offering any support for their HD-CVI DVRs being added to a VMS as an encoder for recording.

If and when Dahua and Hikvision are shipping real analog HD encoders, certainly that would be preferable to integrating with DVRs.

Btw, Dahua lists this as a 16 channel 1080p HD encoder. It doesn't say it is CVI but I assume it is. Also, only sales listing so far is on Alibaba.

John, yes the NVS1604HDC-A is a 16 channel H-CVI encoder which can also be fitted with a local HDD for redunancy recording if there is a network/VMS failure. Dahua are also releasing 4 and 8 channel models. We have samples on order for testing. Prices will initially be higher than the comparable embedded DVRs until sales volume increases. Customers wanting to upgrade from SD analogue to HD utilising existing coax will be able to consider an embedded DVR or VMS/PC recording architecture subject to their functionality and budget requirements. Mainstream CCTV users with coax infrastructure will enjoy a much lower cost and simpler migration path to HD which will accelerate the uptake. Once you view for like comparable HD-CVI and HD IP cameras from the same VMS, most people struggle to pick the difference.

John, yes the NVS1604HDC-A is a 16 channel H-CVI encoder which can also be fitted with a local HDD for redunancy recording if there is a network/VMS failure.

HDMI/VGA outs also. Forgive my skepticism but I'm not grasping the essential difference between the encoder with DVR capabilities and the DVR with encoding capabilities.

Do you know if the encoders going to work as ONVIF devices as well?

We don't have any information yet confirming ONVIF support, but I'd expect it does. However, our VMS supports these encoders via the Dahua private protocol which provides a more extensive integration with better outcomes compared to an ONVIF integration of Dahua hardware. We are expecting these encoders to have different firmware to the DVRs which will be optimised for an encoder/VMS recording application which will be based upon feedback to Dahua from partner VMS manufacturers.