Jump to content
Security Installer Community

X-vision Xp16sw100e


xviz

Recommended Posts

i will accept you know more about photography than me lol

so assuming 400 iso what is roughly the equivilent megapixel count?

I have used 16MP Colour cameras with varying types of canon 35mm, 50mm f1.2 and f1.4 lenses. The resolution is that of what your talking about with a dslr. I get 4872 x 3248. I was just workin on footage last night and on a site that 150m x 80m wide, one camera covered the whole site. I could read reg plates to almost the end, and that was the recorded footage, not the live. Its absolutely frightening. I have been looking at this stuff for quite a while because when I got into this business I was disgusted with the quality of footage out there.

Id rather sell one High Def sytem than fifty low def c##p ones, yes I will be poorer but I will also have the best footage in the country!!!

Eoghan O'Flaherty

Fusion Networks

Southern Ireland

info@fusionnetworks.ie

VOICE VIDEO DATA NETWORKS

Link to comment
Share on other sites

i will accept you know more about photography than me lol

so assuming 400 iso what is roughly the equivilent megapixel count?

For 400ISO the tests I'm talking about only had B&W film. However, for ISO 200, they had colour Fuji Sensia compared with a digital image, and reckoned somewhere between 4 and 6 megapixels on the then current pro dslr cameras would have matched that image. They worked it out for medium and large format film. Of course, it was hard to compare for MF and LF because the lenses would be different etc.

The end result was was that even ISO 50 Velvia and ISO 25 Tech Pan (B&W) at 35mm format could only resolve around 16 megapixel equivalent. Given the film hasn't changed much at all, and the digitals have improved hand over fist since 2002, this could even possibly be considered a little lower now.

Either way, a 21 megapixel Canon 1DsIII is more than likely going to out-resolve film using the same lens. Your pixels are getting smaller than the film grain of even the tightest grained films, and noise is almost non existent at low ISO speeds on these modern dslrs.

Bit old now, but still interesting if you want a look: http://www.clarkvision.com/imagedetail/fil....digital.1.html

Link to comment
Share on other sites

interesting stuff. I was always under the impression film still had the edge... it appears not then.

out of interest whats the image size or raw and low compression jpeg on that 21 mp cannon?

I'd say, very tight grain films in medium or large format have the edge over current medium format digitals, by a small margin. However it also depends how you print it. Given most printing is digital now, film has to undergo the double loss of scanning then printing, whereas digital only has to undergo printing. If you do traditional printing, you can perhaps gain more for film, but there's a limit to how large you can go without resorting to digital printing, and given the resolutions, you need to go very large before you start to notice where these cameras/films are showing their limits.

I think most of the people who say that film has the edge aren't scientists at heart. They're artists, and they love the grain and the tonality of film. Fair enough, and everyone to their own. If it was purely about accuracy, and science, painting would have died out with the advent of photography.

RAW files on the Canon 1DsIII are 25 megs and jpegs at full resolution are around 6-7 megs - not sure how low you can get them as you whack up the compression though. They're 5616 x 3744 pixels. You don't notice the size difference so much on the jpegs because they're only 8 bit, but the raw files are massive. I noticed a bigger than expected (I knew it'd be larger with more pixels) jump in raw file sizes from a previous 12 bit sensor. 14 bits really multiplies the amount of data stored for each pixel, and you get a lot more recoverability at either end of the histogram, without the image looking "stretched", and banding appearing in smooth toned areas like blue skies. This was quite an issue with 12 bit sensors, and a huge issue if post processing 8 bit jpegs rather than shooting raw and converting the final image to a jpeg.

Link to comment
Share on other sites

Problem with 7 meg jpegs is at say 10fps that 70 megs and pretty close to 100 meg ethernet limits, just for one camera. Gigabit would help, but im wondering if the future is 1 ethernet port per cam. Also we will start to hit pci bas bandwidth limits.

securitywarehouse Security Supplies from Security Warehouse

Trade Members please contact us for your TSI vetted trade discount.

Link to comment
Share on other sites

RAW files on the Canon 1DsIII are 25 megs and jpegs at full resolution (5616 x 3744 pixels)@ 14bit

Now doing a raw calc on those figures 5616x3744x14 bit => 294,368,256 bits (36,796,032 bytes) or about 35Mb that's a LOT of freaking data and that only rreally gives 4 bits for red / green / blue + 2 for correction.

Now if that translates into 8-bits per color then the numbers just get NUTS at 60Mb, we'd defiantly need Gigabit ethernet for transmitting that as raw for any speed and some darn good on the fly compression to write it to disc...

Intruder / CCTV / Access Control Technical Support Personal

Subscriber to the "K.I.S.S" principle, that's Keep It Simple Stupid, are you?

Link to comment
Share on other sites

Problem with 7 meg jpegs is at say 10fps that 70 megs and pretty close to 100 meg ethernet limits, just for one camera. Gigabit would help, but im wondering if the future is 1 ethernet port per cam. Also we will start to hit pci bas bandwidth limits.

That's way over 100Mb ethernet - it's only 12.5 megabytes per second! 70 megabytes per second is 560 megabits. That's even outpacing USB and firewire I think. You'd only get one 10fps stream on a gigabit connection...possibly two if you came and went on fps/compression.

Just thinking out loud, 12.5 megabytes per second for 100Mb ethernet. Say 10fps is acceptable for most applications, that's a maximum of 1.25 megabytes per frame. Given a reasonable compression, that's probably around 7-8 megapixel, and compressing like hell, up to 10, or perhaps 12 at the very outside?

For gigabit, you're talking about 125 megabytes per second, which is 12.5 megabytes per frame (again 10fps). Depends on compression, but it'll be somewhere around the size of what many medium format digital backs give out - around 30-40 megapixel. But that's assuming 1 ethernet port per camera on the dvr with current technology.

That's going to make dvrs very costly because afaik they currently use standard computer motherboards etc. If you're talking about having to have multiple gigabit network ports, then you're going to need a lot of expansion slots for those cards. I don't think out-pacing PCI bandwidth limits is going to be of concern though because the gigabit ethernet limit will kick in first. By the time they start going for terabit ethernet, or whatever networking system comes in to replace gigabit, they'll have had to look at the whole architecture of computers to find a whole new ballpark of speed.

One other thing is getting the information off the sensor. That 1DsIII can manage 5 fps, with a physical shutter. The only dslr I know of capable of 10fps is the Canon 1DIII which is popular with sports shooters. It is only 10 megapixel though.

Link to comment
Share on other sites

yeah sorry i meant the 7 meg was close lol

That's going to make dvrs very costly because afaik they currently use standard computer motherboards etc. If you're talking about having to have multiple gigabit network ports, then you're going to need a lot of expansion slots for those cards. I don't think out-pacing PCI bandwidth limits is going to be of concern though because the gigabit ethernet limit will kick in first. By the time they start going for terabit ethernet, or whatever networking system comes in to replace gigabit, they'll have had to look at the whole architecture of computers to find a whole new ballpark of speed.

The proper dvs, use custom boards not pc boards. Some of the pelco units do but not many... at the moment

Agreed re the ethernet bandwidth but if you used multi cards that would/could swap the bus.

regarding the cameras, from a pro point of view you wouldnt use a dslr anyway, but they are good for proof of concepts. Wait for the inq's to come out.

James

securitywarehouse Security Supplies from Security Warehouse

Trade Members please contact us for your TSI vetted trade discount.

Link to comment
Share on other sites

yeah sorry i meant the 7 meg was close lol

The proper dvs, use custom boards not pc boards. Some of the pelco units do but not many... at the moment

Agreed re the ethernet bandwidth but if you used multi cards that would/could swap the bus.

regarding the cameras, from a pro point of view you wouldnt use a dslr anyway, but they are good for proof of concepts. Wait for the inq's to come out.

James

Might be custom boards, but I suspect they'll not be building something that complex from scratch. I'm guessing they'll probably integrate circuit designs from other manufacturers on to their boards, and as such, while it'll be much more focussed towards throughput for video, it'll probably have many of the same issues as far as bandwidth is concerned. At the end of the day, computer technology is driving speeds up more than the comparitively small dvr market.

One other factor that will cause a vast increase in price over current megapixel cctv cameras is the fact that you'll get tremendous heat build up in the sensors running that fast continuously, so probably will have to produce much larger sensors or lower voltages (making noise an issue with the signal/noise ratio getting poorer) to get round this. Low light performance will drive away from risking noise, so I can only see prices skyrocket as sensor size increases dramatically.

Other directions might start to happen though. You might get pixel counts levelling off, but computer processing taking off. For example, imagine an area monitored by 3 or 4 cameras from all sides. There's no reason why a computer couldn't process this into 3D data and provide a fully 3D experience where you could fly round the suspect, and based on previous information the computer might intelligently be able to gather from frames it has already seen, it could build up a very much improved picture of a suspect's details, allowing you to fly in and "look him square in the face" to the resolution afforded by the best pictures that were gathered of him from all angles.

Neil

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.