Disqus - Latest Comments for NotoriousBRK

Re: ADT Customer Acquisition Economics

NotoriousBRK — Wed, 23 Jan 2013 16:06:24 -0000

Jonathan - It's been a while since I was really involved enough in burg panels to care about the details, but some quick googling definitely turns up references to ADT's panels being unmonitorable by other stations. Whether the particulars are complete proprietary protocols or other nuances, I'm not sure...

http://www.youralarm.com/ip...

It might just be that the firmware is locked and is the panel is not reprogrammable via standard technician tools, but once the panel dials out it uses a standard CID format.

Re: WDR Shootout 2013

NotoriousBRK — Wed, 23 Jan 2013 11:46:55 -0000

Any details on how the tested cameras performed in terms of adjusting to a scene when the overall exposure changed greatly, like a person walking through the doorway?

Re: ADT Customer Acquisition Economics

NotoriousBRK — Tue, 22 Jan 2013 13:18:23 -0000

Couple of minor points/questions...

1) You use the example of cable TV and being able to cancel at any time. This is generally true for CATV, but most satellite TV providers do lock you into a 2 year contract. Presumably to offset the same basic installation and equipment costs.

2) Regarding the perpetual equipment lease and inability to switch monitoring providers, it was my understanding (at least years ago) that the panels ADT installed were essentially proprietary and could not be monitored by most 3rd parties. So, a customer could end up owning their panel, but that wouldn't really change things anyway.

Good analysis overall though.

Re: Digital Zoom Tutorial

NotoriousBRK — Fri, 18 Jan 2013 18:53:51 -0000

John - you did a good job with the digital zoom, but you forgot to initiate "detail enhance" and "image match"... Then you would have clearly seen the reflection of the book the killer was reading in the woman's glasses.

Re: Look How Horrible Avigilon's Quality Is

NotoriousBRK — Thu, 17 Jan 2013 10:15:47 -0000

I like the fact the John belts out these articles when claims appear over hyped.

Re: Testing Thermal Performance When Snowing

NotoriousBRK — Wed, 16 Jan 2013 12:01:56 -0000

It would be really interesting to see this test re-done side-by-side with a decent camera (720p?, yes I know it's not a direct resolution comparison) in day and night settings (night with an IR illuminator of course). I think the benefits of the thermal camera would be more apparent in that case.

Re: Arecont's Wacky Pixel Per Dollar Metric

NotoriousBRK — Fri, 21 Dec 2012 15:38:18 -0000

On the surface this is "good" marketing... Many people might look at a 10MP camera and think "I can't afford that". But, that camera can (in theory) use the same housing, much of the same electronics, and so on as a 1MP camera. You're really only paying additional for the price of the sensor+ISP, and a lens upgrade.

But, the cost of the camera is only part of the story here. That 10MP camera is also going to need a LOT more storage. I think it would be safe to assume that someone who wants a 10MP camera wants a 10MP recording. The storage cost is going to scale much more linearly with the increase in resolution. So, the camera hardware might only cost a "little bit" more, but the overall cost of that channel when you factor in the storage aspect of it could be a LOT higher.

Re: Liars at Arecont Vision

NotoriousBRK — Wed, 28 Nov 2012 14:12:50 -0000

But only the big circle is labelled as "actual image". There is nothing in the overview shot that indicates it is an "actual image" or even a CCTV image at all. You coud say that the overview shot is just showing a typical business, and the "actual image" is what the camera is actually capturing as it sits in that environment.

Believe me, I totally see your point, but you're arguing on a technicality here. The "actual image" did in fact come from an actual Arecont camera, however that camera's FOV looks nothing like the overview shot of the bakery...

Conceptually, I agree with you. The ad encourages readers to draw unrealistic conclusions about how the products would work in real-life. I don't think that Arecont "lied", but I do think that the scenario could have been illustrated in a more realistic manner (which would have also been rather boring from an advertisement perspective).

Re: Liars at Arecont Vision

NotoriousBRK — Wed, 28 Nov 2012 12:38:24 -0000

Well, in all honesty we don't know if the "actual image" is actual or not (besides the fact that is truly "an image" and so therefore must be "actual").

Thinking about what a typical customer would be looking for in the scenario depicted in the ad, they'd probably want reasonably good coverage of as much of the floor as possible with a concentration of detail on the cash handling areas.

I think the money shot is, well, the money shot. And if it came from an actual camera, then they are using real data in the portion of the ad that is most critical, and the overview scene is the part with more "creative license". The more deceptive part of the ad would be that to get the "actual image", you'd need a mostly overhead shot, which would greatly impair your ability to get a good overview shot (based on the ad, I'm assuming a standard single-imager camera here) of the rest of the store as shown in the ad.

Calling them a liar without knowing the true origin of the actual image is, IMO, a bit over-reaching. A more reasonable critique of the expectations set by this ad would be that their cameras somehow magically allow you to get good overhead shots *and* good profile shots in an indoor setup like this. A customer who purchased the camera and had it setup to give solid images of the cashier/customer cash hand-offs would probably be disappointed if he expected the wide-view image to look similar to what you see there. This would also tie in with setting a false expectation that for a small store like this one camera could "do it all".

But, calling them outright liars over this ad needs more proof on our part, IMO.

Re: Liars at Arecont Vision

NotoriousBRK — Wed, 28 Nov 2012 07:45:24 -0000

I'm not sure all the hate is really justified here. What's unclear is if their camera is capable of producing the "actual image" shown, and if this is *actually* an image from one of their cameras, or just stock art (and when I say I'm unclear, I'm not saying I doubt them, just that I don't have any data).

The ad is *clearly* a "serving suggestion" scenario. Just looking at the wide-angle shot and the zoomed in shot, you can see that you're looking at two completely different camera angles.

I think a reasonable person would take away from that ad that you can use some given Arecont camera that would give you a good "wide" area view *and* the ability to read currency denominations with the same camera. Looking at the FOV, and assuming those are 18" floor tiles, the camera is covering an area about 20' x30' (roughly). Let's say you want 100ppf to read currency clearly (and that's probably being a little bit generous), that would mean you'd need a camera with about 2000x3000 pixels, or about 6MP to cover that store general floor area roughly as shown and have a reasonable chance of being able to zoom in and read the numbers on a bank note.

Given Arecont's camera range, that doesn't seem unreasonable, other than the fact the ad is for their Megadome series, which tops out at 5MP (according to the ad).

I do agree that we need more truth in advertising in the security industry, but I think that every company attempts to showcase their products in slightly extreme or overly-ideal scenarios. If I were going to pick on that ad, I wouldn't doubt the resolution aspect of it, I'd doubt that the wonderfully exposed, beautifully saturated "overview" shot came from a CCTV camera and not a high-end DSLR.

Is this ad taking some liberties? Yeah, probably, and I'll bet you could spend all day critiquing ads in SDM or any other magazine. Can Arecont deliver on the "promise" of this ad as shown? Yeah, probably, depending on the exact layout and interior of the site. Certainly not with a 1 or 2MP camera, but with a higher-res one most likely.

Personally, my guess is that the "actual image" is in fact an actual image, but the overview scene is taken with a DSLR as a representative example. And, much like lux ratings, without the technical details behind it, it's really hard to take too much away from this data.

Re: Axis Doubles Edge Storage Size Supported

NotoriousBRK — Thu, 15 Nov 2012 11:45:51 -0000

It varies by card, but you can find datasheets from most of the manufacturers on this.

All other things being equal, larger cards will last significantly longer, since they are being cycled less often. Some of the lifespan items can also be affected by exactly how the camera manages the storage (writing of data).

In many cases though, I think a decent reliable card would have a lifespan measured in years today. And through it is inconvenient to change SD cards on a camera in many cases, having to go around once every 3 years to swap cards and clean/tweak lenses and so forth probably isn't an insurmountable thing for the benefits you get...

Re: Honeywell - The Most Open VMS on the Market?

NotoriousBRK — Wed, 07 Nov 2012 10:14:09 -0000

Given the options in the VMS market today, I don't think that talking about the mechanisms you support for pulling a basic video stream should be in any way at the top of the list of things you promote.

Standard RTSP support is just a *requirement*, as is ONVIF and/or PSIA support. In a modern VMS pitch these are just base-level user expectations. Same thing for supporting MJPG, megapixel cameras, mixed 4:3 and 16:9 camera aspect ratios.

Differentiators would be how you handle the things that are NOT simplistic functions. What about edge-storage support? Can you handle inputs from cameras (either alarm I/O, or analytics style triggers), how is the VMS storage managed? How do you pull up live camera views or retrieve stored video?

As you said, direct integrations, and things beyond what video formats you support are the defining features worth promoting.

Re: CCTV Camera Trees / Clusters

NotoriousBRK — Fri, 02 Nov 2012 15:18:33 -0000

A 360 panoramic (of either variety) is mostly impractical in those scenarios, the mounting pole would create a significant blind spot in a good portion of the view.

You could do 2 180 cameras off either side of the pole, possibly, reducing your camera count to 2.

PTZs are certainly an option, depending on the purpose. If you're trying to get continuous area coverage (which, IMO, is common in these parking lot environments) then the PTZ also has a lot of "blind spot" in essence.

One of the biggest drivers I've seen for this setup (and I've designed several like this myself) is simply budget. It's cheaper easier and faster to setup and many cameras as possible in one location. FOV compromises are negligible, as large areas like this are rife with tradeoffs working within the average budget. I also hear arguments both ways about the psychological impact of having the environment seem like it's under prison surveillance.

Re: Killing Hosted Video

NotoriousBRK — Thu, 01 Nov 2012 19:29:46 -0000

Duncan -
Most flash memory manufacturers have datasheets that tell you the number of bit-flips the media can handle over it's lifetime.

The technology has gotten better (more durable and self-correcting) over the years, and the lifespan for a given application is almost directly correlated to the size of the media. Larger drives fill up slower, so each bit is over written fewer times over a given time period.

As flash memory prices have dropped, it's cost effective now to use a 64GB chip where 4 years ago only 8GB might have been feasible. That would also extend your expected lifetime of the media by a factor of 8.

Given that many of these applications are being limited to fairly low-res cameras (720p and less) in many cases I'd expect a largish media (32GB or more) to outlive the useful life of the camera, or maybe require one swap (not necessarily low-burden, but few small retail customers have cameras that you couldn't reach with a ladder).

SD media as primary storage would still concern me for a critical hi-res camera (1080p or more), but wouldn't be my first concern in a deployment of cheaper cameras in a lower-concern retail type setting.

Re: Killing Hosted Video

NotoriousBRK — Thu, 01 Nov 2012 16:15:04 -0000

That would actually be an interesting service for a VMS company to provide to their users... A "data vault" for high-value events. Within the UI an operator can select video clips to save and store, add annotations, etc. The service would do some watermark verification to authenticate the video and provide a means for the customer to enable access to 3rd parties (police, etc.). It would create 1-time use URL's, trackable URL's, and maybe even subtle watermarks or manipulations to the video files to track leaks in the event something ends up on YouTube. Data could be burned to DVD and submitted for court evidence for an additional fee.

Re: Killing Hosted Video

NotoriousBRK — Thu, 01 Nov 2012 16:11:31 -0000

Can you provide specifics on what their hosted video solution is doing? EG: resolution, frame-rate, amount of storage, approximate billing cost?

Re: Killing Hosted Video

NotoriousBRK — Thu, 01 Nov 2012 14:16:22 -0000

I've heard of some of these setups also, but they always seem to be associated with deep-pocketed national brands that appear to think they can put this out as a loss-leader now and somehow make it up when economies of scale kick in at some point.

The problem is that it just never gets to a point of profitability or practicality.

Even in this model it would still make more sense to put a basic NVR on-site that can be managed and monitored remotely. From the end-users perspective the storage is still managed and handled by some other entity, and it also wouldn't preclude them from getting remote access to their system.

I would personally bet against this long-term. Small retail shops are a large, but generally stingy, customer base.

Re: Killing Hosted Video

NotoriousBRK — Thu, 01 Nov 2012 12:04:13 -0000

Cloud storage for the surveillance industry is following the same curve and trend as for the general IT industry.

Off-site storage is not cost-effective as an NVR replacement of any sorts. Even a modest CCTV system can accumulate multiple gigabytes of data per day, and 7-30 day retention times are not uncommon. If you look at the IT side of this and see what has been cost-effective you see that the dollars per gigabyte ratio is pretty high, especially in cases where you want some kind of reliability with the service. Paying $10-$20/mo for an account that holds a few dozen GB of data is typical, and even these services are priced this way because with IT data it is generally low transaction rates. Look at that from a video perspective, and you're easily paying 10x for cloud storage would you would for a basic DVR (ex: 100GB Dropbox plan is $10/mo., a cheap/basic DVR with 1TB of storage is $250).

Costs are not trending towards anything that would make "the cloud" a replacement for large-scale local storage cost effective anytime in the predictable future, especially when we look at this for megapixel cameras instead of D1 gear.

Where cloud storage does make sense is in the surveillance version of the Dropbox model: files that have been identified in some manner as being "interesting" or "of high value" are moved to a remote storage location. You have to get even more precise than just "motion" based recording, and there are a number of practical ways to do this.

I have yet to hear of a scalable model for the cloud-as-NVR-replacement approach, but I do see many companies offering essentially "off site backup" of key video files in different approaches that seem to be doing well.

Re: IR Indoor vs Outdoors

NotoriousBRK — Tue, 30 Oct 2012 18:55:55 -0000

This would generally be true of any light source, visible or not. I think what makes it seem somewhat unique to IR is that typical lights do not have a "distance" rating as IR illuminators commonly do.

Re: Arecont Magic Marketing

NotoriousBRK — Tue, 16 Oct 2012 16:46:56 -0000

So then if someone developed a solar-powered 3G 50MP camera you could cover almost any site with a single camera and almost no install costs...

Re: Arecont Magic Marketing

NotoriousBRK — Mon, 15 Oct 2012 14:43:11 -0000

I don't think even the stadium example holds up. For this to be practical, it would have to be 32 (.7) cameras all installed in the same location, and looking at a total FOV of no more than about 100degrees.

I know you can put a wide or super-wide angle lens on 10MP camera, but then you're introducing a lot of distortion to the image making it a non-uniform comparison.

A better way to present this from a marketing angle would be to say something like:
"In order to get the same pixel detail as our single 10MP camera, you'd have to install 37 SD cameras and arrange them carefully across the FOV. This would obviously be impractical and cumbersome, making the 10MP camera an ideal choice for scenarios that need higher detail and want to maintain a small infrastructure footprint."

This basically turns it around, or to put it another way if the 10MP camera disappeared and there were no more megapixel IP cameras, you'd have to use 37 SD cameras to match it. But, there is no practical scenario where you are ever going to replace maybe more than 4 fixed cameras with a single higher-resolution camera (that being a scenario where you had already created a sort of manual panorama view with multiple cameras). Suggesting that a high megapixel camera has any sort of day-to-day practicality of being substituted for a gob od SD cameras is wishful thinking at best.

Re: PPF vs Small Variances in Lighting

NotoriousBRK — Tue, 09 Oct 2012 19:54:39 -0000

I was speaking in very general terms. But you can go even further down that rathole if you like... You'd have to take into consideration the sensor, the resolution of the lens itself, compression mechanisms (as you noted), shutter speed relative to speed of moving objects, and a few other things.

Pixels per foot is a good way to describe very rough approximations of what you should or should not commonly expect. But if there is a spec for something like "faces at 100' suitable for identification", then your design would have to take all of these factors into consideration. If the customer simply wants "the best possible detail for a budget of X", you can use ppf as a way to describe potential trade-off between various cameras with "all other factors being equal".

Re: PPF vs Small Variances in Lighting

NotoriousBRK — Tue, 09 Oct 2012 18:06:26 -0000

When I've used ppf specs, it's always mentioned as a *minimum*, and is often used as a reference in cases where a customer expected to get "faces and plates" from a camera with something absurdly low, like 10ppf, at the target. So, the response is usually something like "you'd need to cover about 1/5th of what you're currently covering in order to get the detail you expect". Most of the time I find that end-users (and many installers) haven't thought about pixel density at all, and just assume that a high MP camera will yield fine details across the FOV.

Lighting is obviously important, and the relationship between light and ppf requirements is somewhat inverse, as you have less lighting you will typically need higher ppf density to see the same level of detail in a given object.

Re: How Far Can a PTZ See? (2012)

NotoriousBRK — Tue, 25 Sep 2012 14:21:37 -0000

I think I'm going to propose adding a new feature to our VMS: Sub Pixel Application Zooming (SPAZ). Basically you can zoom down to the point than an individual pixel is rendered across the screen with the sub-components of the pixel shown as their individual RGB values. This would of course best any existing digital zoom functions, and it could probably be patented as well.

Look for it soon.

Re: How Far Can a PTZ See? (2012)

NotoriousBRK — Tue, 25 Sep 2012 08:25:39 -0000

Great post. You should also cover the other most common question I hear in regards to "zoom", which relates to VMS's and HD cameras. People often ask about things like "how much digital zoom does it have", based on the assumption that one VMS could potentially "zoom" in on a given FOV better than another.

It is sometimes befuddling how little people understand about basic FOV math, assuming that software can make up for a lack of resolution. And in some cases I feel that trying to truly educate on the basics makes it seem like I'm hiding something or making excuses for not being able to offer a gimmick that fits their bad assumption.