Things We Learned at SecTech Camera Shootout
Every time you bolt a bunch of CCTV cameras to a text jig and start playing with the lights, there’s something new to see and we learned plenty at this year’s event. Most importantly we discovered the best cameras are even better than they were last year.
SECTECH Camera Shootout was enthralling and challenging in multiple ways. Despite a day of pre-commissioning in Sydney there was still plenty of work required in every city to stay on top of things – in fact wrangling the big group of engineers and technicians was the biggest operational challenge of the event and organisers didn’t get on top of that until Brisbane – this meant a few management related hiccups along the way. Next year’s pre-commissioning is slated to go for days – we’ll start the tour with an operational system that is locked in place with only minor tweaks allowed.
The nature of SecTech’s Camera Shootout is that in order to get a sense of what’s best you need to be in the room and to that end, we don’t list which cameras we thought were best, though there were a number we consistently liked, including a newcomer we’d never seen before. The shootout is subjective in any case, with considerably different test spaces offering variations of reflectance and ambient light. Arguably Melbourne was the toughest for light – we could only get down to 0.7 lux at the face and there was ambient IR in the room (albeit only a small amount with a wide angle of view around 25m from the target) that centre management refused to turn off.
The PTZ group in Sydney attracted a great deal of interest – the cameras tested were the best in the world and the huge venue gave us the opportunity to get a clear sense of which looked best. It’s fair to say that in this test there was a clear-cut optical performance winner on the day. There’s a dynamism to PTZ tests that’s well worth re-visiting. We should point out here that Bosch’s extremely capable MIC PTZ didn’t appear in the test group because of IP address contention – we’ve reviewed that camera separately in SEN.
Every year organisers battle to get uniformity across the contenders. This is partly because each group of engineers has its own ideas of ‘best’. For instance, they may use a wider angle of view to retain maximum aperture in low light (with a subsequent impact on WDR performance). Each year there are pleas for organisers to stipulate stricter conformity – the same resolution, the same bitrate, the same frame rate, the same lenses and angles of view. That’s never easy to do in performance-based tests like these.
This year, organisers divided the groups not by resolution but by form factor, given form factor is often the governing consideration. There were also cost variations but they were not as considerable as in the past. The disparity in lens prices was less this year, too. Certainly, all contenders are entitled to install quality lenses on full body cameras, even if these are just high end Fujinon lenses, not specialised manufactures. We saw again that narrower angles of view return higher quality imagery.
Something else we saw this year was a couple of manufacturers locked in colour, which gave us a clear sense of how long they could retain colour but we didn’t get an idea of how well they’d have done in night mode. This was a shame, according to some visitors. We certainly noticed the value of night mode – monochrome gives court admissible images in challenging situations, no doubt.
While all the variations tend to suggest an apples-to-pears comparison, SecTech Camera Shootout is first and foremost a test of which camera does best in the scene conditions all the cameras are pointed at, regardless of specification and camera settings, with viewers able to decide for themselves on the basis of image quality (sharpness and noise levels), bitrate and motion blur who is robbing Peter to pay Paul. Throw aircraft landing lights into the equation from the rear and things get very messy indeed – next year we’ll be adding strobes.
The idea is that attendees think about performance, see variables, get a sense of the limitations of the technology and run headlong into the laws of physics (and their sub-laws of signal processing). This was most noticeable in Brisbane, where in a compact and very dark room, we found that while the latest generation of IP CCTV cameras are very good, they are cannot (Sony’s VB770 notwithstanding), see in near total darkness.
For a number of years, surveillance camera choice has been predicated on low cost, even if silly angles of view on 1080p domes have robbed image streams of what little sharpness they had left after burrowing through a couple of millimetres of low quality smoked plexiglass. Camera demos like SecTech’s Camera Shootout don’t only start a conversation about image quality, they allow vistors to see whether or not the emperor is wearing any clothes.
What’s at the heart of this discussion is raw camera performance in challenging conditions – gloomy or night time applications under sub-10 lux where so many incidents take place and where blur-inducing movement – even walking pace – as well as headlight flare and poor camera positioning, can render CCTV systems not useless but certainly very much less capable than security managers need to be.
We saw some interesting developments at SecTech. Perhaps the biggest one for me was the realisation that improvements in compression are liberating ISO – lots of ISO – and that amplification is certain to be one of the key areas of development in the future. It’s not surprising that ISO is such a big deal in DSLRs – the difference in quality between early digital cameras with sub-1000 levels of ISO and the latest cameras with up to 20,000 is enormous. In low light applications, you either spool up ISO and slow down the shutter and if there’s movement in a scene, the latter option is the end of face recognition and the death of recognisable license plates.
Camera performance in the presence of IR when in night mode was almost uniformly excellent this year with many cameras showing great contrast with the Vivotek 48W array we were using on the test jig. In the presence of IR, shutter speeds are normalised, too, and with a couple of cameras we got almost perfect delineation of Ronnie’s chest bars.
Something we noticed was the differences between the monitors we used – one being 1080p and the other 4K. It goes without saying that monitors introduce variables into performance and in the eyes of some people, this was significant enough that they believed the higher contrast display of the 1080p monitor trumped the high resolution of the 4K monitor. The 2 monitors – the 4K is the newer model – also rendered colour differently.
SecTech Roadshow got great support from the Hills and Genetec teams this year, often under serious duress from ravening hordes of technicians, intent on making one very last cross-my-heart-and-hope-to-die tweak to an obscure slider. The Hills team in Melbourne copped the worst of this, with talk of code being written under the jig during setup and general fiddling leading to delays that cost a vital dummy run. Many thanks to all the engineers involved in keeping the wheels turning – it was much appreciated.
Thanks too, to the techs from manufacturers and distributors who re-built the test jig in 5 cities – it’s a big job and it’s a team effort – well done to everyone who took part. Of course, the biggest thanks of all go out to the installers, consultants and end users who came along to see the show. We’re looking forward to showing even bigger and better things next year! ♦