Soundy

Yeah, got a special request for something like this... still gotta find out if it actually needs to be recorded or if it's just for display, but assuming the former... anyone got any ideas?

Nice, thanks! Waiting to hear back more specifics from the client... so far it doesn't sound like this actually needs to be recorded, just displayed live, so I'm considering a basic HD camcorder of some sort as well, but overall, I'd prefer to have the OPTION to record... because you just KNOW if we don't have that, someone WILL ask for it down the road.

I have, and it works well... if properly implemented. Doing RAID in software is a poor place to start anyway. MythTV is not the same load as a DVR. For one thing, it's not reading *and* writing heavy data streams simultaneously. You're also typically only accessing one, maybe two files at a time (if you're recording one thing and watching another). Remember that a DVR doesn't just have to write to the drive, it also has to keep the files indexed, cataloged in its search database, and keep things cleaned up (deleting old files, etc.). That generally means a lot of constant R/W operations *all over* the volume. And in most cases, that means doing it with the drive(s) heavily fragmented (a few systems pre-allocate large files and then keeps its video in those, but most write multiple smaller files). So no, you can't really compare the performance of a HTPC with that of a surveillance DVR.

I guess the rules are different for Linux. Not really... http://www.prepressure.com/library/technology/raid "(RAID5)Advantages Read data transactions are very fast while write data transaction are somewhat slower (due to the parity that has to be calculated)." RAID levels and their relative advantages/disadvantages have nothing at all to do with OS, especially when you're using a hardware controller. You can, actually... as mentioned previously, some drives don't have sufficient performance or reliability for DVR usage. Rory has mentioned many times the issues he's had with "green" drives on GeoVision systems (stuttering playback, dropped frames, all rectified by switching, even to standard desktop drives). I've found "green" drives generally don't have the longevity of server- or enterprise-grade drives when used in DVRs and RAID arrays, either - remember, this isn't your basic home media server that's seeing maybe a few hours' sporadic access a day; this is something that's seeing sustained read/write for days at a time. Most array manufacturers (QNAP, Synology and Enhance Tech are the ones I've used) list recommended drives for their systems, and none of them recommend green drives for anything but SOHO use. Edit: And yes, RAID10 will give you some benefits as well... at a cost of space. Everything is a trade-off.

Have a request from a customer, they want an ID camera feed to a TV delayed by about 10s. Camera will be a CW504, so just standard NTSC video, nothing too fancy. Any suggestions? I give it about a week before too many people are weirded out by it and they ask us to disable the delay, but what the hey, it's their money. A TV or monitor with this capability built-in MIGHT be acceptable as well.

Read operations are faster on RAID5, but not write operations - don't think that would be a big advantage in this case. RAID0 would be faster, but I really don't like using it for video data - seen too many DVRs lose *all* their data due to a single failed disk.

That tends to be true of the cheap low-end stuff... mid- to upper-end gear, however, has so many components from other places, and often sees various stages of assembly in different places, that it can't really be pegged as being "from" any one country, other than maybe where the company's head office is located. Heck, one of the most flimsy, poorly-assembled cameras I've ever seen - the aforementioned AV3155DN - *claims* to be fully built in the US... see my opinions here: viewtopic.php?f=19&t=22353 Most segments of the electronics industries aren't much different than the auto industry in this regard. As they say in the movies, "American components, Russian components... ALL MADE IN TAIWAN!"

No they aren't. (Almost) all domes should allow you to adjust the tilt of the camera module within its yoke, and rotate the yoke within the dome. Three-axis models add the ability to rotate the camera module within the yoke. This allows you to adjust the angle that the camera looks at the scene from. Not such a big deal with a ceiling-mounted dome, but VERY practical if you're wall-mounting. However, NOT all domes add this function!

Basic optical lens theory, nothing (directly) to do with restricting the amount of light - look at the old-school "pinhole" camera, which will have near-infinite DOF. More good reading here: http://en.wikipedia.org/wiki/Depth_of_field#Effect_of_lens_aperture Correct; the camera needs to have DC iris drive as well. This can help DOF in the day by closing down the iris, but at night, most will open the iris all the way up anyway.

You wouldn't really see a difference with large amounts of RAM, and besides, you'd need 64-bit Windows to even see more then 3.5GB of it (you could have 16GB in there and 32-bit Windows will only see up to 3.5GB). As thewireguys says, more RAM wouldn't really help at this point; it's more a matter of whether there's enough that the system isn't swapping to disk excessively.

Is this a single-drive system (OS and data on one drive)? How much RAM is in it? Possible if the OS and data are on the same drive, and there's not enough RAM, that pagefile read/write is hampering things.

Depends on the drive as well... "green" drives aren't generally as well-suited for this sort of purpose.

Yeah, this really shouldn't require ANY end-user intervention, ever.

Hmmm, pasted the wrong link. WTF? Here it is: http://www.avigilon.com/products/cameras-video/pro-ip/ I believe that lens would be about 1/5 the total cost of this lens/camera system... Something else to consider with a parking lot is that you're generally going to want your camera at a fairly high vantage point to get maximum area coverage... at that point, you're far less likely to get plates or faces just because of the angle. Add the fact that a moving car will be moving across the frame, making any motion blur caused by a slow shutter even worse. Depth of field is going to be a trade-off as well, because you'll get that with a smaller lens aperture... which then reduces the light. To compensate, you either need to boost the signal electronically - creating more noise - or reduce the shutter - creating motion blur. What I would suggest (and most here will as well) is to use a lower-resolution, better-low-light-gathering camera or cameras for parking lot overview, and then have a dedicated camera for capturing plates, mounted at a low level on a "choke point" (such as a gate or driveway). This does not necessarily need to be a megapixel camera, if the choke area is tight enough, and the more straight-on you are to the direction of movement (so the plate is moving in-line with the camera instead of across the frame), the less you have to worry about motion blur. There are cameras specially designed for license plate capture that would do well here.

They want a live delay from their ID camera to their customer-awareness monitor... record-delay isn't really an option (at least not via the DVR). I suppose a consumer PVR would work for this, but if it ever lost power, the delay would be gone - someone would have to pause it, then resume it manually.

Of course, if you really want to go all-out with high res, low light, and uber-sharp optics, there's always... http://www.avigilon.com/products/cameras-video/pro-ip/ 11MP and 16MP on a full-frame 35mm sensor... This bad boy will give you a nice wide angle and let in LOTS of light: http://www.usa.canon.com/cusa/consumer/products/cameras/ef_lens_lineup/ef_24mm_f_1_4l_ii_usm

One thing you have to remember (that I believe has been mentioned) is that the higher you go in resolution, the worse your low-light performance. This is a simple matter of physics - for any given size of sensor, the more pixels you have, the smaller each pixel becomes, and the less light it can collect. So the next question becomes whether you REALLY need 5MP. As with the lens issue, this is why we ask so many questions - because there are SO many different requirements and options out there, and despite what the big-box systems would like you to believe, there really isn't a one-size-fits-all solution. There are lots of factors involved in framerate besides just the camera, not the least of which are the capabilities of the NVR and network. Compression type and level will also play a factor, not just because of bandwidth required for the stream, but the amount of in-camera processing required (a higher level of compression MAY, depending on the codec and the scene and the amount of movement and various other factors, might have reduced framerate due to higher processing demands.... for example). And again, the next question becomes, do you really NEED that high a framerate? Sure it's NICE... but is it NECESSARY? Even for a parking lot, outdoor and exposed to direct weather, IP63 should be sufficient (see: http://en.wikipedia.org/wiki/IP_Code#Liquids.2C_second_digit). You're really limiting your options with over-ambitious requirements here...

Whoa, found their price list... yeah, $1000-$2000 for the video delay units. Ouch.

Welcome aboard!

That's pretty vague - "reformat" to what? What resolution DOES it support, relative to the cameras? Losses may not be significant. Either way, if you don't already have the NAS, try looking at a different model... or even a different brand. QNAP have some nice NAS/NVR units as well - www.qnap.com or www.qnapsecurity.com. They're in the same price range. Some sacrifice may be necessary, because honestly, if you're doing all your recording as FTP or SMB streams, you have no video management, and using the footage later will be come extremely unwieldy.

Uhhh... sorry, typo on my part: "6.3mm is the widest it goes".

Thanks! Wow, the AMV units are SPENDY. The Allen box might be more suitable, any idea what the MSRP is on the thing?

Really... those camera's are capable of streaming to an FTP server. So even if they are not supported by the NAS (which they are) it should be possible as long as the NAS supports the connection (FTP/HTTP, SMB/NFS). They should... but that seems overly complex. Just one potential issue I see with using ftp or smb, is that each brand of camera will write in its own format - not even file format so much as NAMING format and storage folder structure. For example, one might use a year\month\day\time.avi structure, another might go year-month-day\time.mjp format. Trying to coordinate those for playback would be a PITA. Heck, I've done it with a single IQ camera to an FTP server, and searching the video afterward is nothing but a hassle. Use the built-in NVR functionality that's a feature of the Synology - it will be easier to configure and overall more effective. It also gives you the ability to view and search all cameras, live or recorded, in a single browser interface. It may not sound like a big deal up front, but the first time you watch a subject move from one camera to the next on a split-screen display, or search for specific events, you'll realize how convenient it will be.

Cool. I'll have to check with the guys providing the TV if it will do component input...

Hrrrm... 6.3mm is the widest it goes, that may not be enough. Any suggestions as to some way to take an MP stream and output it to HDMI? I guess I could do it with a Vigil HD Viewer...